Coactivation of MEP-biosynthetic genes and accumulation of abietane diterpenes in Salvia sclarea by heterologous expression of WRKY and MYC2 transcription factors

Plant abietane diterpenoids (e.g. aethiopinone, 1- oxoaethiopinone, salvipisone and ferruginol), synthesized in the roots of several Salvia spp, have antibacterial, antifungal, sedative and anti-proliferative properties. Recently we have reported that content of these compounds in S. sclarea hairy roots is strongly depending on transcriptional regulation of genes belonging to the plastidial MEP-dependent terpenoid pathway, from which they mostly derive. To boost the synthesis of this interesting class of compounds, heterologous AtWRKY18, AtWRKY40, and AtMYC2 TFs were overexpressed in S. sclarea hairy roots and proved to regulate in a coordinated manner the expression of several genes encoding enzymes of the MEP-dependent pathway, especially DXS, DXR, GGPPS and CPPS. The content of total abietane diterpenes was enhanced in all overexpressing lines, although in a variable manner due to a negative pleiotropic effect on HR growth. Interestingly, in the best performing HR lines overexpressing the AtWRKY40 TF induced a significant 4-fold increase in the final yield of aethiopinone, for which we have reported an interesting anti-proliferative activity against resistant melanoma cells. The present results are also informative and instrumental to enhance the synthesis of abietane diterpenes derived from the plastidial MEP-derived terpenoid pathway in other Salvia species.

Recently, we have used hairy root (HR) technologies in combination with metabolic engineering strategies, and provided one of the first evidences that the production of tryciclic abietane diterpenes can be enhanced by 3-4 times by overexpressing in S. sclarea hairy roots the heterologous genes DXS or DXR from Arabidopsis thaliana 3 . We have also demonstrated that the expression of the genes encoding these two enzymes, acting up-stream of geranyl-geranyl disphosphate (GGPP), the common precursor of most of the MEP (2-C-methyl-D-erythritol 4-phosphate) pathway-derived terpenes (Fig. 1), are transcriptionally regulated by methyl jasmonate (MJ), along with other genes acting upstream and downstream of GGPP 5 . Significant correlations were found between the content of total abietane diterpenes and the level of expression of the DXS, DXR and GGPPS genes as well as of the CPPS gene (copalyl disphosphate synthase), the first enzyme involved in the synthesis of copalyl disphospate, precursor of several plant abietane diterpenes 6 . Altogether this set of data indicates clearly the existence of a MJ-dependent gene regulation of the MEP-derived terpene biosynthetic route.
In this work, we aimed at enhancing the content of abietane diterpenes production in S. sclarea HRs by overexpressing transcription factors (TFs), whose expression is known to be MJ-inducible, and that might be potentially involved in the regulation of the MEP-pathway genes. Recent findings have elucidated the role of some TFs in the biosynthesis of plant secondary metabolites, including terpenes, as reported for the sesquiterpene artemisinin in Artemisia annua (AaWRKY1, AaERF1, AaERF2, AaORA1 and AabZIP1) [7][8][9][10] . Similar evidences have been also obtained for other TFs, such as OsTGAP1, which controls the synthesis of phytocassanes, a class of diterpene phytoalexins in rice 11 ; GaWRKY1, regulating gossypol biosynthesis in cotton 12 and TaWRKY1 controlling the regulation of 10-deacetylbaccatin III-10-O-acetyl transferase involved in taxol production in Taxus chinensis 13 .
Interestingly, several TFs regulating secondary metabolism are MJ-responsive 14 , among which, MYC2 is one of the best characterised. It belongs to the basic helix-loop-helix (bHLH), which positively or negatively regulates secondary metabolism during JA signalling in a species-specific manner. The basic region of the MYC2 protein, containing 15-20 mostly basic amino acids, is involved in binding the conserved G-box (5′-CACGTG-3′) found in the promoters of MJ-responsive genes, by forming homo-and/or heterodimers with other members of the MYC family, such as MYC3 and MYC4 15 .
Another emerging class of TFs involved in the regulation of plant secondary metabolites is the large WRKY family, a plant specific class of TFs, firstly discovered for their role in abiotic and biotic plant response 16 . Accumulating evidence suggests that certain WRKYs regulate the production of valuable natural products by modulating transcriptionally the biosynthetic genes involved in the synthesis of phenylpropanoids, alkaloids and terpenes 17 . Although some WRKY proteins may contain additional domains, they all have a common highly conserved DNA-binding domain (DBD) 18 of ∼60 amino acids characterised by the conserved heptad WRKYGQK amino acid motif at their N-terminus, which may vary in number, and a zinc ion chelating finger structure at their C-terminus. WRKY TFs bind their cognate (T)TGAC)C/T) W-box cis-elements in the promoter region of target genes 19 .
Owing to the limited available information on S. sclarea genome and transcriptome, we used Arabidopsis thaliana as a genetic source of potential TFs involved in MJ-dependent regulation of the biosynthetic genes of the MEP-pathway, relying on the fact that the structural organization of TFs is largely conserved in the plant kingdom.
The WRKY TF family has been well characterized for their involvement in SA signalling and plant defense in Arabidopsis, but their role remains less clear for jasmonate signalling, especially for their contribution in the regulation of secondary metabolism 20 . Among several members belonging to the large WRKY TF families we focused our attention on the AtWRKY18, AtWRKY40 TFs as either publically available microarray datasets and published resulted have demonstrated that they are involved in the methyl jasmonate signalling pathway 21 , where they have been suggested to operate also with overlapping roles 22 . In addition, a recent genome wide binding analysis of WRKY18 and 40 detected WRKY binding sites in several genes encoding functions related to MJ signalling [22][23][24][25][26][27] .
More robust evidences are available for the involvement of AtMYC2 as the best characterized and most multifunctional TFs, acting as a regulatory hub within the JA signalling pathway 15 . Furthermore, a genome-wide search has revealed that 25% of early JA-responsive genes (i.e. genes responding to JA treatment within 30 min of JA application) contain the G-box cis-acting sequence, providing additional support for the potential importance of this sequence in MYC2-regulated expression of JA-responsive genes 28 .
Herein, we confirmed that in A. thaliana the expression of these three TFs is tightly controlled by MJ elicitation and precedes the activation of the biosynthetic genes of the MEP-pathway from which abietane diterpenes derive. In addition, by a preliminary promoter scanning analysis, several W-and G-cis-elements were identified in the promoters of the biosynthetic genes encoding enzymes of the MEP-pathway. Thereafter, we proved that the ectopic expression of AtWRKY18, AtWRKY40 and AtMYC2 activates in a coordinated manner the transcription of several genes of the MEP-pathway in S. sclarea HRs, which, in turn, accumulated significantly higher contents of abietane compounds.
These data suggest a concerted gene regulation mediated by MJ, possibly through activations of TFs. To identify putative cis-regulatory DNA elements that might concert the MJ dependent activation of the MEP-pathway biosynthetic genes in A. thaliana a region of 1000 bp upstream the transcription start site point of AtDXS, AtDXR, AtCMS, AtCMK, AtMCS, AtHDS, AtHDR, AtGPPS genes was scanned by MatInspector software. W-box (TTGAT/C) and G-box (CACGT) binding sites of WRKY and MYC transcription factors, respectively, were found in the promoters of the analysed genes prevalently in multiple copies, conversely G-box were less abundant and not detected in CMK and GGPPS promoters (Fig. 3a). In particular, by using the online tool FIMO (MEME suite) for genome-wide promoter scanning analysis, a significant enrichment of the W-box TTGAT/C motifs was detected in the promoters of the MEP-pathway genes (Table S2) and a slight frequency increase of few G-boxes (Table S3). Interestingly, cis-acting elements involved in the MJ-responsiveness as well as W-boxes and G-boxes were found also in the promoter of AtWRKY18, AtWRKY40, and AtMYC2 gene (Fig. 3b).
Altogether these data predict a potential involvement of MYC2 and members of WRKY TFs in the transcriptional regulation of biosynthetic genes of the MEP-derived terpene pathway in this species. To gain more information about the expression profile of these TFs in response to MJ elicitation, we measured their transcript levels in two-week-old A. thaliana seedlings exposed to 150 μM MJ for 24 h or an equal volume of DMSO (mock control). The expression of AtWRKY18, AtWRKY40, and AtMYC2 was strongly induced by MJ as early as 30 minutes of treatment (Fig. 3c, d, e, respectively). Remarkably, AtWRK40 showed the highest up-regulation (fold change > 80, 1 h post treatment). AtWRKY18 was also strongly up-regulated (fold change >15, 30′ post treatment). The MJ treatment induced also a significant overexpression of AtMYC2 (fold change > 5, 30′ post-treatment). The MJ-induced activation of the three TFs precedes that of the biosynthetic genes involved in the MEP-pathway and their early activation suggests a possible role of WRKYs and MYC2 in the upstream regulation of the MEP pathway. Based on the high conservation of these genes in the plant kingdom, AtWRKY18, AtWRKY40, and AtMYC2 were overexpressed in S. sclarea HRs, to establish their potential role in activating transcriptionally genes of the MEP-pathway in this plant species and, eventually, to enhance the synthesis of bioactive abietane diterpenes. Transformation and overexpression of AtWRKY18, AtWRKY40 and AtMYC2 TFs in S. sclarea HR lines. S. sclarea HRs overexpressing constitutively AtWRKY18, AtWRKY40 or AtMYC2 were generated by Agrobacterium rhizogenes-mediated transformation. Control lines were obtained by overexpressing a 174 bp fragment of A. thaliana GUS gene. Ten independent HR lines overexpressing each TF were obtained and kanamycin-resistant differentiated HRs were transferred into hormone-free medium and characterised for the expression of the transgenes, the absence of contaminating bacteria and the absence of any spurious amplification of endogenous TFs in the GUS control lines (Fig. S1a-d). Three independent HR lines overexpressing the three TFs were randomly selected for further experiments. The expression levels of the transgenes in the selected lines were determined by qRT-PCR ( Fig. 4a-c).
To ascertain potential growth impairment associated to the TF overexpression, HR growth, measured as dry weight at interval of one week after the inoculation in fresh liquid medium, was also monitored for four weeks ( Fig. 4d-f). All the overexpressing lines were characterised by a longer lag phase, particularly evident in the HR lines overexpressing the AtMYC2 TF. The final dry weight was dependent on the level of expression of AtWRKY18: HR line #6, characterised by the lowest level of the transgene, showed a final biomass not significantly different from the control HR line. Overexpression of AtWRKY40 did not cause any significant negative effect on the final HR growth, while a strong detrimental impairment was observed in all three HR lines overexpressing AtMYC2. Independently of the effects on the HR growth, the overexpressing HR lines have an evident red colour compared to the control HR line (Fig. 4g), which indicates indirectly the accumulation of abietane diterpenes, as reported by Vaccaro et al. 3 .

MEP-pathway biosynthetic genes are transcriptionally activated in AtWRKY-18 and -40 and
AtMYC2 overexpressing S. sclarea hairy roots. To prove whether or not the three TFs control the biosynthetic genes involved in the synthesis of abietane diterpenes, the expression level of endogenous SsDXS Except for the SsCMK, all the analysed genes were up-regulated in AtWRKY18 overexpressing HR lines, with a fold increase ranging from two to five (Fig. 5a). Interestingly, SsCPPS, encoding the first synthase responsible of conversion of GGPP into CPP, showed the highest expression level (about 5-fold increase) at least in two AtWRKY18 HR lines. In AtWRKY40 overexpressing lines, SsDXS (fold change > 3), encoding the first committed enzyme of the MEP-derived pathway, was preferentially up-regulated, together with SsCPPS (Fig. 5b). In all AtMYC2 overexpressing HR lines, the expression levels of SsDXS, SsDXR, SsHDS, SsGGPPS and SsCPPS were notably increased, indicating that AtMYC2 is a potent transcriptional activator of the MEP-pathway. Again, SsDXS (5-fold increase) and SsCPPS (more than 4-fold increase in the line AtMYC2-10) showed the highest transcriptional activation, while SsCMK transcript levels did not change significantly (Fig. 5c). Data represent mean values ± SD of three independent HR lines for each overexpressed TF, obtained from three technical replicates. Biomass production (mg g −1 hairy root dry weight) of three different AtWRKY18 (d) or AtWRKY40 (e) or AtMYC2 (f) overexpressing HR lines compared to control HR line during 4 weeks of culture are also reported. The asterisks indicate a significant difference between control and transgenic HRs (P < 0.001) according to two-way ANOVA with Bonferroni post hoc test. Phenotypes of representative transgenic HR lines, grown in hormone-free liquid medium (g). Ectopic expression of AtWRK-18, -40 and AtMYC2 enhances the production of abietane diterpenes in HR lines. The content of abietane derivatives was measured by targeted high-performance liquid chromatography with diode-array detector (HPLC-DAD) in acetone extracts obtained from S. sclarea HRs overexpressing the three TFs and compared to the content of the control HR line. Chemical structures of the most abundant S. sclarea abietane diterpenes and their typical chromatographic patterns are reported in Fig. S2, as comparison between the control and a representative transgenic line (AtMYC2, line #10). When examined collectively, the total abietane diterpene content varied with minimal fluctuations among the individual HR lines overexpressing each TFs overexpression (Fig. S3). AtWRKY18, AtWRKY40 and AtMYC2 boosted a significant increase in the total content of these compounds, expressed as mg g −1 dry weight (Fig. 6). Compared to the control HR line, AtMYC2 overexpressing HR lines accumulated a significant higher content of abietanes (fold-increase > 5), followed by AtWRKY40 overexpressing lines (fold-increase > 4) and by AtWRK18 lines (fold-increase > 2) ( Fig. 6 and Table S4).
However, we found that the ectopic expression of the three TFs affected in a variable manner the final HR growth (Fig. 4). Therefore, we also determined the final yield of individual abietane diterpenes in the different TF overexpressing HR lines (Fig. 7). A general increase in the final yield of carnosic acid, salvipisone and aethiopinone was obtained by overexpressing AtWRK40, which was not associated to a significant decrease in HR final dry weight (Fig. 4d) and preferentially activated the expression of the DXS and CPPS genes (Fig. 5b). Aethiopinone (>6 mg L −1 ) was the diterpene whose synthesis was highly enhanced (P < 0.001) by AtWRK40 overexpression, with a 4.2 fold-increase above the basal level of the control HR line. As expected, the negative pleiotropic effect on final HR biomass due to the AtMYC2 overexpression heavily penalized the final yield of individual abietane diterpenes.

Discussion
Various biological activities have been reported for plant abietanes either of the abietane-phenolic type (e.g. carnosic acid and ferruginol) or abietane-quinone-type (aethiopinone, salvipisone and 1-oxoaethiopinone) 1,29 . Among them, aethiopinone has been shown to have anti-proliferative activity against HL-60 and NALM-6 leukemia cells 2 . We have also demonstrated that aethiopinone have an anti-proliferative activity in human melanoma cells, together with a safety profile for non-malignant cells 3 . Very low amount of these class pf diterpenes are synthesized in organs and tissues of several Salvia species and different biotechnological approaches (cell culture, hairy roots, elicitation) have been applied to enhance their content 3,5,30 . In the last few years, metabolic engineering strategies have been also attempted to increase the production of plant bioactive secondary metabolites, by tackling either single key-limiting genes or regulatory transcription factors, providing compelling evidence that it is feasible to boost the secondary metabolism in crop and medicinal plants [31][32][33] . Biosynthetic route of abietane diterpenes is tightly controlled by genes belonging to the well-conserved MEP-derived terpene pathway 34 . We have provided evidences that the content of abietane diterpenes can be enhanced by 2-3 times by overexpression of heterologous AtDXS and AtDXR genes in S. sclarea HRs 5 and more efficiently by MJ-elicitation (>20-fold increase), due to a coordinated transcriptional activation of several genes belonging to the plastidial MEP-derived terpene pathway 5 . These data indicate the existence of MJ-dependent transcriptional regulators of the MEP-pathway, as found for other metabolic pathways in several plant species 14,35 . It is well documented that metabolic engineering by targeting TFs is a powerful tool to manipulate the metabolic flux in plants and to eventually supersede the bottlenecks associated to targeting single genes 17,36,37 .
When we started our study on identifying putative regulatory TFs to boost the metabolic pathway of abietane diterpenes, information on the genome and transcriptome of S. sclarea and other Salvia spp were not publically available. Paralleling our previous results S. sclarea 5 , we found that biosynthetic genes of the MEP-pathway were transcriptionally activated by MJ in A. thaliana plantlets and that the promoter regions of A. thaliana MEP pathway genes contain G-and W-boxes, binding sites for MYC2 and WRKYs TFs. Noteworthy, in the case of W-boxes, hypergeometric distribution test revealed a significant enrichment of TTGACT/C motifs in the promoters of MEP pathway genes, corroborating the existence of a cooperatively regulation of these genes mediated by WRKY TFs in A. thaliana. Single or multiple copies of G-box and W-box were present in the promoter of the GGPPS gene, as already reported in the S. milthiorriza hortologous genes 38 , as well as in the promoter of genes  encoding enzymes acting up-stream of GGPP. This suggests that WRKY and MYC TF members might exert a putative common regulatory mechanism of the MEP-pathway genes in plants. MYC2, a basic helix-loop-helix (bHLH) transcription factor, has been shown to be directly or indirectly involved in regulating secondary metabolites, through the MJ-induced degradation of the negative regulatory Jaz proteins, leading to the release of the AtMYC2 and to the activation of its target genes biosynthesis 14 .
In the last few years WRKYs family is also emerging as another family of TFs controlling transcriptionally biosynthetic genes of metabolic pathways and as a valuable tool for enhancing the synthesis of different class of secondary metabolites 17 . For instance, CrWRKY1 regulates the terpenoid indole alkaloide (TIA) biosynthesis by modulating the expression of several TIA genes in Catharanthus roseus 39 . GaWRKY1 is a transcriptional activator of the gene CAD1-A, participating in cotton sesquiterpene biosynthesis 12 . Overexpression of TcWRKY1 increases the expression of 10-deacetylbaccatin III-10 β-O-acetyl transferase (DBAT), which plays a prominent role in the biosynthesis of taxol in Taxus chinensis 40 . Interestingly, publically available Arabidopsis microarrays revealed that about 30% (22 of 72) of WRKY TFs respond to jasmonate treatments 21 . Remarkablly, a recent genome-wide binding analysis in A. thaliana have indicated that AtWRKY18 and AtWRKY40 bind the promoter of several target genes, known to be transcriptionally regulated in a MJ-dependent fashion 23 .
In this study, we have confirmed that AtMYC2 and two WRKY TFs, AtWRKY18 and AtWRKY40, which contain MJ responsive elements (G/ATGTT) in their promoters, are transcriptionally regulated by MJ in Arabidopsis seedlings. In addition, W-box elements have been found in the promoters of AtWRKY18 and AtWRKY40 suggesting a possible autocatalytic regulation, in line with the well established auto-and cross-regulation mechanisms described for the WRKY family 16 . It has been also reported recently that several WRKY TFs are activated by MJ in S. miltiorrhyza HRs, which, in turn, accumulated high levels of tanshinones, a class of abietane-type diterpenoids 41 .
To establish their potential role in controlling the production of abietane diterpenes, S. sclarea HR lines overexpressing these three TFs were generated and we demonstrated that AtWRKY18, AtWRKY40 and AtMYC2 TFs were able to regulate the expression of several biosynthetic genes belonging to the plastidial MEP-derived pathway. AtWRKY18 activated significantly the expression of S. sclarea genes DXS, DXR, HDS, and GPPS genes, encoding enzymes involved in the synthesis GGPP, the universal precursor of diterpenes and also the expression level of the CPPS gene, encoding a synthase converting the cyclization of GGPP into copalyl diphosphate (CPP). It has been reported that CPP is the precursor of ferruginol and carnosic acid 42,43 , two abietane diterpenes, and might be the precursor also of aethiopinone, the main abietane diterpene found in S. sclarea roots, although this has to be further proved. Instead, overexpression of AtWRK40 was able to activate preferentially the transcription of DXS gene, encoding the first committed step of the MEP-pathway and the CPPS gene. Further studies are necessary to elucidate whether the transcriptionally activation of genes encoding enzymes of the MEP-pathway is due to a direct binding of these two WRKY TFs to the promoters of the targeted genes or through activation of up-streams TFs. However, in all the AtWRKY-18 and -40 overexpressing HR lines, the level of transcriptional activation of the MEP biosynthetic genes was far below the fold-increase that we have reported by MJ treatment 5 . This might be related to the occurrence of a concerted transcriptional activation of the whole metabolic pathway not necessarily guided by the action of a single TF, but rather by a combinatorial role for several TFs 44 . Actually, it has been reported that AtWRKY18 and AtWRK40 interact physically and functionally with each other to form homo-and heterodimers 22 , indirectly suggesting that these WRKY TFs may act in an additive or synergistic manner to drive metabolic flux towards specific compounds. Owing to their closely related functions, it would be interesting to ascertain the possibility to control the metabolic pathway of abietane diterpenes by co-expression of both these WRKY TFs in S. sclarea HRs.
Biosynthetic genes encoding DXS, DXR, HDS, and GGPPS enzymes were also transcriptionally regulated in S. sclarea HR by overexpressing AtMYC2, in a quantitative manner as found in AtWRK18 overexpressing HR lines. This confirms MYC2 as a master regulatory factor in the MJ activation of the secondary metabolism, possibly by controlling other TFs or biosynthetic genes. In A. thaliana studies on AtMYC2 have been primarily focused on establishing its role in mediating the overall transcriptional changes in response to biotic and abiotic stresses 15 . However, a wealth of recent information is also uncovering the role of MYC2 in regulating specifically biosynthetic genes. In Catharanthus roseus, CrMYC2 regulates the expression of ORCA2 and ORCA3 TFs, which, in turn, tune the expression of alkaloid biosynthetic genes through a MJ-dependent manner 45 . Tobacco bHLH1 and bHLH2 TFs, similar to AtMYC2 and CrMYC2, were functionally identified as regulators of nicotine biosynthesis using virus-induced silencing in MJ-treated N. benthamiana, by decreasing the expression of PMT gene (putrescine N-methyltransferase) and MPO gene (N-methylpitrescine oxidase) and overexpression of N. tabacum counterpart genes also cooperatively activated the PMT2 gene 46,47 . AaMYC2 overexpression significantly activated the transcript levels of CYP71AV1 and DBR2, which resulted in an increased artemisinin and anthocyanin content in A. annua 27 . Recently, the regulatory role of MYC TFs have been proved in Salvia miltiorrhyza: SmMYC2a and SmMYC2b, which are however divergent from AtMYC2, are positive regulators of SmPAL, SmHPPR, SmHCT6, SmCYP98A14, SmMK, SmCMK, SmMCS, SmCPS and SmKSL, encoding enzymes of the biosynthetic route of tanshinones and phenolic acids in the roots of this species 48 .
Although the three Arabidopsis TFs were able to activate transcriptionally genes encoding enzymes acting upstream the GGPP as well as the CPPS gene in S. sclarea HRs, it remains to ascertain whether the observed transcriptional regulation is due to a direct binding of W-box and G-box present in the promoters of these biosynthetic genes and/or by activating other TFs, as demonstrated for AtMYC2 15 .
The transcriptional up-regulation of biosynthetic genes by AtMYC2, AtWRKY18 and AtWRK40 overexpression was associated to a general increased accumulation of total abietane diterpenes, compared to the control HR line. Considering the total abietane diterpenes content, the most boosting effect was obtained by overexpressing AtMYC2, which induced a coordinated up-regulation of four out of the five analysed biosynthetic genes belonging to the MEP-derived pathway as well as of CPPS gene. The content of aethiopinone, the compound for which we have reported a novel anti-proliferative activity against melanoma cells 3 , was enhanced consistently (8-fold increase) in the HR line with the highest transcript level of AtMYC2. Surprisingly, the same enhancing effect on the content of abietane diterpenes was not triggered by overexpressing AtWRKY18, which was shown to induce transcriptionally the same group of biosynthetic genes. We can only speculate that in AtWRKY18 overexpressing HR lines the ratio between precursor content and enzyme level in each step of this metabolic pathway was somehow unbalanced, preventing an expected accumulation of abietane diterpenes. This aspect has a great relevance in designing production platforms of secondary metabolites based on metabolic engineering 49 . Interestingly, the overexpression of AtWRKY40 in S. sclarea HR lines, which activated almost preferentially the activation of the DXS and CPPS genes, was sufficient to boost a significant accumulation of abietane diterpenes, especially aethiopinone and salvipisone content (6 and 3 fold-increase, respectively). These data confirm our previous findings evidencing that DXS and CPPS are limiting enzymes for the synthesis of abietane diterpenes in S. sclarea HR lines 3,5 .
Another aspect of great relevance in metabolic engineering is that a single TF may be involved in multiple molecular response and cross-talk among different metabolic pathways. The overexpression of a TF might have a strong negative effect on plant cell, organ and tissue growth and final biomass, as reported for overexpression of ORCA3 in C. roseus 50 . Although AtMYC2 overexpressing HR lines accumulated the highest content of each abietane compound, the growth of these transgenic HR was drastically impaired, affecting negatively the final yield of this class of compounds. This is not surprisingly given the multitasking role of MYC2 as a point of crosstalk between distinct signalling cascades, including light, abscisic acid and jasmonic acid 51 . Possible negative effects of constitutive AtMYC2 overexpression could be circumvented, using an inducible promoter to drive its expression. In our hands, the overexpression of AtWRK40 induced the highest final yield in abietane diterpenes, especially aethiopinone, due to a trade-off between the expression level of biosynthetic genes and HR growth.In conclusion, the data presented herewith demonstrated the feasibility of enhancing the content of abietane diterpenes by ectopic expression of heterologous Arabidopsis TFs, which were able to regulate in a coordinated manner the expression of several genes of the MEP-derived terpene pathway and the CPPS gene. To our knowledge, this is the first example of metabolic engineering of abietane diterpenes in S. sclarea using WRKY TFs. The present results are also informative and instrumental to enhance the synthesis of abietane diterpenes derived from the plastidial MEP-derived terpenoid pathway in other Salvia species. Genome-wide scanning of G-box and W-box motifs was carried out by the online tool FIMO (Find Individual Motif Occurrences) 54 . The genomic Arabidopsis promoter dataset containing 37694 sequences (approximatively 1000 bp upstream the transcription starting site) was obtained from Genomatix (version TAIR10_1). Promoter Motif scanning was carried out by searching for MYC and WRKY core motifs and accepted variants as reported by Godoy et 55 and Eulgem et al. 56 , respectively. Data were expressed as number of a given cis-element per promoter (cis-element frequency) and percentage of promoters with a given cis-element.

Material and Treatments
Plasmid construction, HR transformation and growth analysis. The plasmids containing the full-length cDNA of A. thaliana WRKY18 (Accession n. U14890), WRKY40 (Accession n. C105126) e MYC2 (Accession n. U12679) were obtained from the Arabidopsis Biological Resource Center (https://abrc.osu.edu). The coding sequences of the genes were amplified by polymerase chain reaction (PCR) using a High-Fidelity DNA Polymerase (Pfx, Invitrogen, Carlsbad, CA, USA) with specific primers (Table S5), characterised by the presence of the short sequence CACC at 5′ for the cloning in pENTR/ kit D-TOPO ® (Invitrogen) to generate an Entry-Clone. The correct insertion and the absence of mutations were verified by sequencing (Primm Biotech, Milan, Italy). Subsequently, coding sequences of the transcription factors were subcloned through the LR reaction (The Gateway ® LR Clonase ™ enzyme mix kit, Invitrogen) in the gateway Destination vector pGBW17 (kindly provided by Prof. Nakagawa 57 , driven by the constitutive strong viral 35SCaMV promoter, and containing the C-terminal 4xmyc-tag and kanamycin resistance selectable marker 58 , through a site-specific recombination 59  standard freeze/thaw cycle and CaCl 2 method 60 and used to transform leaf disks from 20 day-old S. sclarea axenic plantlets, as previously reported 3 . After several transfers to medium containing 50 mg l −1 kanamycin and decreasing concentrations of cefotaxime (from 100 down to 50 mg l −1 ), HRs developing from the infected areas were individually excised, sub-cultured several times on hormone-free medium and maintained at 23 °C in the dark. Kanamycin HR independent lines, with no bacterial contamination, were selected, and sub-cultured into 250 ml flasks, containing a hormone-free liquid MS medium supplemented with 50 mg l −1 kanamycin and kept on a gyratory shaker at 120 rpm at 23 °C in the dark. The HR lines were sub-cultured every week.
HR growth was analysed by inoculating equal amounts (0.5 g) of untransformed or TF overexpressing HR lines into MS hormone-free liquid medium. Dry weight was monitored during one month at one-week intervals.
Nucleic acid purification, PCR and RT-PCR analyses. Genomic DNA was extracted from HR lines by the cetryl trimethyl ammonium bromide (CTAB) method as described in Doyle and Doyle 61 and used as template in PCR reactions to establish stable integration of AtMYC2, AtWRKY18 and AtWRKY40 in transgenic HRs, using a forward primer located at the promoter of the PGBW17:35S 2 and a reverse primer located at the 3′ end of the respective genes. Amplification reactions were set using 2.5 units of Taq polymerase (Invitrogen, Carlsbad, CA, USA), and run as follows: 94 °C for 5 min, followed by 30 cycles at 94 °C 30 sec, annealing 30 sec, at the different T m , according to the primers melting temperatures, 72 °C 1 min, and a final extension 72 °C for 5 min. To detect the absence of contaminating Agrobacterium, the HR genomic DNA from control and transformed HRs was used as template to amplify the virD2 gene, using specific primers 62 .
Total RNA, from A. thaliana plantlets or S. sclarea HRs, was extracted using the plant RNA/DNA Purification kit (Norgen Biotek Corporation Ontario, Canada), according to the manufacturer's protocol. For semi-quantitative RT-PCR, complementary DNA was synthesized from 1 μg total RNA, previously treated with RNase-free DNAse I, using random hexamers and the Superscript III RT (Invitrogen, Carlsbad, CA, USA) at 50 °C for 50 min. In the PCR reactions one microliter of cDNA was used as template with specific primers (Table S5) and 2.5 units of Taq polymerase (Invitrogen, Carlsbad, CA, USA), using the following conditions: initial denaturation at 94 °C 1 min, followed by 30 cycles denaturation at 94 °C 30 sec, annealing 30 sec at the different T m , according to the primers melting temperatures, extension at 72 °C for a time depending on the length of the DNA target and final extension at 72 °C for 5 min.
For qRT-PCR, complementary DNA was synthesized as described above. The reactions were performed in a 20 µl volume, containing several cDNA dilutions and 0.5 µM primers, specific for each gene. Cycler-DNA Master SYBR Green I mix (Roche Diagnostics Ltd, Lewes, UK). The reactions were run in a Light Cycler rapid thermal cycler system (Roche Diagnostics Ltd, Lewes, UK) under the following fast cycling steps: initial denaturation for 2 min at 94 °C, followed by 40 cycles at 94 °C for 2 s, 58 °C for 30 s. In addition, melting curves (20 min; from 58 °C to 90 °C) were generated to check any spurious amplification products. To normalize RNA levels, S. sclarea or A. thaliana 18S rRNA was used as internal reference control. The sequences of all the primers used are listed in Table S5. At least three technical repeats from three biological replicates were carried out. Level of gene expression of the three exogenous TFs and of endogenous biosynthetic genes of the MEP-pathway in different overexpressing hairy root lines and the control lines were represented as 2 -deltaCt 63 .
Qualitative and quantitative determination of abietane diterpenes. Abietane diterpenes were identified and quantified as previously reported 3,5 . Briefly, lyophilized and powdered hairy roots (0.5 g) were extracted with acetone for 72 h at room temperature and the obtained residues dissolved in methanol and subjected to HPLC-DAD analysis (Agilent1200 Series, G1312A binary pump, G1329A automatic sample injector, G1315D diode array detector). The HPLC fingerprint was carried out on a C 8 column (Agilent, Zorbax eclips C 8 250 × 4.6 mm) with a sample injection volume of 50 µl. The mobile phase was a gradient elution of water acidified with 0.1% formic acid (solvent A) and acetonitrile (solvent B), starting with 35% B and rising to 100% B after 30 min, at a flow rate of 1.0 ml min −1 . The different diterpenes were detected at 280 nm by comparing with standard purified compounds and concentration calculated by the interpolation of the peak areas with calibration curves, constructed over the range 10-200 µg ml −1 of purified compound. Content of diterpenoids in roots was expressed as mg g −1 of HR dry weight. Statistical analysis. All data are represented as mean ± SD of at least three independent biological experiments performed in triplicate. The statistical significance of HR growth rate and diterpene content values was established by the two-way analysis of variance, with Bonferroni post-hoc test analysis, using the GraphPad Prism 5 software. Differences between control and transgenic HR lines were considered to be significant at least when P < 0.05.
Statistical analyses for discovering the enrichment of W box and G cis-element sequences in the promoters of Arabidopsis MEP pathway genes were performed by Hypergeometric test 64 . The test was performed by counting and comparing the number of the promoters of MEP pathway genes containing a given G-and W-box motif (number of success) with respect to the all Arabidopsis promoters (number of success in the population).
Data availability. All data generated or analysed during this study are included in this published article and its Supplementary Information files.