Higher plants produce a large variety of secondary metabolites, which are commonly classified as phenylpropanoids, aromatic polyketides, terpenoids, and alkaloids. These structurally diverse chemicals help protect the plant against pathogens or herbivore attacks and attract pollinators and are also utilized by humans as dyes, flavorings, and pharmaceuticals (Pichersky and Gershenzon, 2002). In particular, alkaloids, which are nitrogen-containing compounds, are often used as important pharmaceuticals, stimulants and narcotics due to their strong biological activity (Facchini, 2001). However, the biosynthetic pathways of many alkaloids and the regulatory mechanisms of their biosynthesis are largely uncharacterized because the distribution of alkaloids is limited to specific plant species.

Among alkaloid biosynthesis, the biosynthetic pathways of several alkaloids have been well investigated at the molecular level due to their chemical uniqueness as well as their economic importance (Dewey and Xie, 2013; Hagel and Facchini, 2013; Sato, 2013; Zhu et al., 2014). Examples of these alkaloids include nicotine in Nicotiana tabacum (Solananceae), monoterpenoid indole alkaloids (MIAs), vinblastine and vincristine in Catharanthus roseus (Apocynaceae), isoquinoline alkaloids (IQAs), berberine in Coptis japonica (Ranunculaceae), sanguinarine in Eschscholzia californica (Papaveraceae) and morphine in Papaver somniferum (Papaveraceae). Among these alkaloids, the biosynthesis of berberine, which is relatively broadly distributed in the plant kingdom and often produced in cultured cells, has been the most intensively investigated (Sato and Yamada, 1984; Ikuta and Itokawa, 1988; Hagel and Facchini, 2013; Sato, 2013). Thus, genes involved in berberine biosynthesis in cultured C. japonica cells have been identified and thoroughly characterized (Sato, 2013). These include genes involved in the condensation of dopamine and 4-hydroxyphenylacetaldehyde to (S)-norcoclaurine by (S)-norcoclaurine synthase (NCS; Minami et al., 2007), the conversion of (S)-norcoclaurine to (S)-reticuline by the sequential reactions of (S)-norcoclaurine 6-O-methyltransferase (6OMT; Sato et al., 1994; Morishige et al., 2000), (S)-coclaurine-N-methyltransferase (CNMT; Choi et al., 2002), (S)-N-methylcoclaurine 3′-hydroxylase (CYP80B2; Ikezawa et al., 2003), and (S)-3′-hydroxy-N-methylcoclaurine-4′-O-methyltransferase (4′OMT; Morishige et al., 2000), then (S)-reticuline to berberine by berberine bridge enzyme (BBE; Minami et al., 2008), (S)-scoulerine-9-O-methyltransferase (SMT; Takeshita et al., 1995), (S)-canadine synthase (CYP719A1; Ikezawa et al., 2003) and (S)-tetrahydroprotoberberine oxidase (THBO; Matsushima et al., 2012; Figure 1).

Using cultured C. japonica cells, specific transcription factors (TFs) involved in berberine biosynthesis have been isolated and characterized because TFs are key for the regulation of gene expression. Specifically, two general transcriptional activators, CjWRKY1 and the non-MYC2-type CjbHLH1, were shown to regulate the expression of almost all of the berberine biosynthetic enzyme genes present in C. japonica cells (Kato et al., 2007; Yamada et al., 2011a). Involvement of a WRKY-type TF in IQA biosynthesis has also been confirmed by the heterologous expression of the Arabidopsis WRKY1 gene in E. californica (Papaveraceae) and by the expression of native PsWRKY in P. somniferum (Apuya et al., 2008; Mishra et al., 2013); these WRKYs are group I type and contain two WRKY domains, in contrast to CjWRKY1, which belongs to the IIc group and contains a single WRKY domain. Involvement of the non-MYC2-type basic helix-loop-helix (bHLH) TF in IQA biosynthesis was also confirmed by the identification of two CjbHLH1 homologs in E. californica, EcbHLH1-1 and EcbHLH1-2, which were shown to participate in the regulation of sanguinarine, a benzophenanthridine alkaloid (Yamada et al., 2015).

Characterization of nicotine and MIA biosynthesis pathways allowed the identification of other families of TFs such as AP2/ERF, an MYC2-type bHLH and WRKY that are involved in biosynthesis regulation (Yamada and Sato, 2013; Zhou and Memelink, 2016). Whereas no AP2/ERF-type TFs have been reported to play a role in IQA biosynthesis so far, octadecanoid derivative responsive Catharanthus AP2-domain 3 (ORCA3) and NtERF189, which belong to the Group IX AP2/ERF subfamily, were isolated and characterized as transcriptional activators in MIA and nicotine biosynthesis, respectively (van der Fits and Memelink, 2000; van der Fits and Memelink, 2001; Shoji et al., 2010). Nicotine and MIA biosynthesis are also regulated by Arabidopsis MYC2-type bHLH TFs, such as NbbHLH1, NtMYC2, and CrMYC2, which show little similarity to CjbHLH1 and function in a jasmonate (JA) signaling cascade together with the CORONATINE INSENSITIVE 1 (COI1) and Jasmonate-ZIM domain proteins (JAZs) (Todd et al., 2010; Shoji and Hashimoto, 2011; Yamada et al., 2011b; Zhang et al., 2011, 2012; De Geyter et al., 2012). In addition to ERF and MYC2-type bHLH, WRKY proteins such as CrWRKY1 (which belongs to a different group than CjWRKY1) were also found to be involved in the regulation of MIA biosynthesis (Suttipanta et al., 2011).

These differences in TF activity found in alkaloid biosynthesis may reflect biosynthetic pathway differences or simply be an artifact resulting from the lack of characterization. Because TFs regulate gene expression through the specific binding to cis-acting elements in the promoter region of target genes, we aimed to characterize the promoter sequences of biosynthetic enzyme genes. Previous work indicates that WRKY proteins can specifically recognize the W-box DNA sequence motif (TTGACC/T), while bHLH and AP2/ERF TFs mainly recognize the E-box hexanucleotide consensus sequence (CANNTG) and the GCC-box DNA sequence motif (CGCCGCC), respectively (Atchley and Fitch, 1997; Rushton et al., 2010; Mizoi et al., 2012). We isolated the promoter sequences of three berberine biosynthetic enzyme genes (CjCYP80B2, Cj4′OMT, and CjCYP719A1) and found that CjWRKY1 activated the CYP80B2 promoter and CjbHLH1 activated the 4′OMT and CYP719A1 promoters in a transient luciferase (LUC) reporter assay (Kato et al., 2007; Yamada et al., 2011a); we then characterized these promoter sequences and their interaction with CjWRKY1 and CjbHLH1.

In this study, we predicted several cis-acting elements such as W-box, E-box and GCC-box motifs in the CYP80B2, 4′OMT, and CYP719A1 promoter regions and characterized the role of these elements using a trans-activation assay involving deletion/mutation CYP80B2 promoters and co-expression with CjWRKY1. Similarly, the effects of CjbHLH1 expression were investigated using deleted 4′OMT and CYP719A1 promoters. Our results suggest the involvement of a putative W-box in the regulation of biosynthetic enzyme genes via CjWRKY1 and the involvement of a putative E-box via a CjbHLH1 function. Additionally, we found unexpected transcriptional activation of biosynthetic enzyme genes via non-W-box sequence interaction with CjWRKY1 and the possible involvement of a GCC-box in berberine biosynthesis in C. japonica. These results suggest that more research is needed to obtain additional data regarding the transcriptional regulation of alkaloid biosynthesis.

Secondary metabolites often mediate defense mechanisms against pathogen or herbivore attacks in plants (Pichersky and Gershenzon, 2002); these functions are likely regulated by defense response genes. In fact, TFs found to be associated with alkaloid biosynthesis belong to the WRKY, ERF, and bHLH groups, which are known to be involved in defense responses. In berberine biosynthesis, two TFs, CjWRKY1 and CjbHLH1, were isolated and characterized in C. japonica (Kato et al., 2007; Yamada et al., 2011a). Although the regulation of berberine biosynthetic enzyme genes by CjWRKY1 and CjbHLH1 had been previously studied, little is known regarding gene regulation because we lack information on the promoter of these biosynthetic enzyme genes. Thus, in this report, we characterized the promoter regions of three biosynthetic enzyme genes using truncations or mutations of promoter sequences.

Although the lengths of isolated promoter sequences were rather short, the promoter regions of CYP80B2 (646 bp upstream sequence from the translation initiation site), 4′OMT (1215 bp upstream sequence) and CYP719A1 (911 bp upstream sequence) managed to activate the LUC reporter gene in C. japonica protoplasts, and two of them showed an induction of LUC activity by either CjWRKY1 or CjbHLH1 co-expression (Figures 4 and 7). These results suggest that either isolated promoter regions might not be sufficient or that endogenous expression of TFs may be sufficient for the induction. A clear reduction of transcriptional activity of the 163 bp CYP80B2 promoter region between -251 and -88 with a mutated E-box (CACGTG→ATCGAC) and GCC-box-like (GCAGCC→GCTTCA) elements suggests that the latter interpretation may be more likely (Supplementary Figure S4). This observation and previous reports suggest that the biosynthetic enzyme genes in the berberine biosynthesis pathway may be regulated by relatively short promoter regions.

Further PLACE and truncation analysis of promoter sequences showed the existence of many putative cis-acting elements targeted by WRKY, bHLH and AP2/ERF TFs and that putative elements such as the W-box at the -227 or -138, the GCC-box at -148 and the E-box at -177 of the CYP80B2 promoter, the W-box at -285, the GCC-box at -201, and the E-box at -232 of the 4′OMT promoter may be important for promoter activities (Figure 3). Interestingly, our analysis suggests the involvement of GCC-box and ERF in the regulation of berberine biosynthesis. Although our preliminary transient RNAi assay using available AP2/ERF EST clones (Yamada et al., 2011a) did not show an effect on the transcription of berberine biosynthetic enzyme genes, more careful investigation of this group of TFs is needed because several AP2/ERF TFs such as ORCA3, ERF189, and tomato ERF have been reported to participate in the regulation of MIA biosynthesis in Catharanthus roseus, nicotine biosynthesis in N. tabacum and steroidal glycoalkaloid biosynthesis in tomato and potato through binding to GCC-box elements (van der Fits and Memelink, 2000; Shoji et al., 2010; Cárdenas et al., 2016).

The importance of the predicted W-box and E-box elements found in the promoter regions was further confirmed by the LUC activity analysis with truncated promoters and co-expression with CjWRKY1 and CjbHLH1, respectively (Figures 4 and 7). More detailed mutation analysis of W-box elements found in the -217 CYP80B2 promoter sequence further supports the involvement of -138, and -116 W-box elements, although some remaining induction by co-expression with CjWRKY1 was observed even in the null putative W-box reporter construct.

This remaining activity may be the result of induction by non-conserved W-box sequences or by indirect effects on transcription/translation, including the effects of the reference gene. Because CjWRKY1 was post-transcriptionally regulated by phosphorylation and protein degradation (Yamada and Sato, 2016), overexpression of CjWRKY1 might interfere/modify the assay system. Furthermore, rice OsWRKY13 is reported to bind not only W-box elements but also the PRE4 element (TGCGCTT), barley SUSIBA2 (HvWRKY46) can bind to a sugar-responsive element (TAAAGATTACTAATAGGAA) as well as the W-box, and NtWRKY12 can bind to the sugar-responsive-like element but not the W-box (Grierson et al., 1994; Sun et al., 2003; Cai et al., 2008; van Verk et al., 2008). While these non-W-box-type elements are not found in the CYP80B2 promoter, the possibility of finding non-W-box elements via CjWRKY1 needs more careful molecular and biochemical characterization. A potential indirect effect on promoter activity has also been reported in nicotine and MIA biosynthesis (Zhou and Memelink, 2016).

Our study also indicates that promoter activity can be modified by a combination of cis-elements. The truncation of the promoter sequences (Figures 3, 4 and 7), the duplicated promoter construct (Figure 5C), and the mutation of putative cis-elements (Figure 5B; Supplementary Figure S4) showed that the effects of putative cis-elements are likely to be additive. In fact, when CjWRKY1 and CjbHLH1 were co-expressed with the 4′OMT promoter, the effects of these two TFs were additive (Figure 7C). This suggests that further careful investigation of promoter sequences and additional information regarding the preferential interaction of cis-elements with TFs may allow us to further dissect the regulation of alkaloid biosynthesis gene expression. Zhang et al. (2011) reported that CrMYC2, a MYC2-type bHLH TF involved in the regulation of MIA biosynthesis, regulates the expression of ORCA3, an AP2/ERF transcriptional activator of several MIA biosynthesis enzyme genes. Because the regulation of gene expression is complex and varies among alkaloid species, it is clear that we need additional genome information to study it further. In fact, until now, we did not know how the gene expression of CjWRKY1 and CjbHLH1 was regulated. Recent progress in next generation sequencing technology can help us obtain whole genome sequences of non-model plants such as C. japonica and thus help us elucidate the mechanisms of signal transduction as a result of interactions between TFs and gene promoters in IQA biosynthesis.

YY, TY, and SY prepared materials and carried out the preliminary experiments. YY performed the experiments. YY and FS designed the research and wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.