Plant secondary metabolites play critical roles in plant-environment interactions. They are synthesized in different organs or tissues at particular developmental stages, and in response to various environmental stimuli, both biotic and abiotic. Accordingly, corresponding genes are regulated at the transcriptional level by multiple transcription factors. Several families of transcription factors have been identified to participate in controlling the biosynthesis and accumulation of secondary metabolites. These regulators integrate internal (often developmental) and external signals, bind to corresponding cis-elements — which are often in the promoter regions — to activate or repress the expression of enzyme-coding genes, and some of them interact with other transcription factors to form a complex. In this review, we summarize recent research in these areas, with an emphasis on newly-identified transcription factors and their functions in metabolism regulation.
Plant secondary metabolites, which for a long time were thought to be not directly involved in energy production, growth, reproduction, or other primary functions, are fascinating and important players in mediating plant responses to biotic and abiotic environmental factors. For example, they contribute to the flavor and taste of fruits and the colors of flowers, and serve as attractors for pollinators and seed dispersers. In addition, secondary metabolites also function as defensive compounds or toxins which guard against pathogens and herbivores, and protect plants from abiotic stresses such as ultraviolet (UV) light (Vogt 2010; Vranova et al. 2012). Furthermore, secondary metabolites have been widely utilized by humans as a source of natural fragrances and pharmaceutical medicines (He and Giusti 2010; Kroymann 2011; Duan et al. 2012).
Plant secondary metabolites are highly diverse in their chemical structures. Structurally and biosynthetically, they are classified into three major groups: terpenoids, phenolic compounds, and nitrogen-containing compounds. Terpenoids contain one or more C5 units, which are synthesized either via the cytosolic mevalonate pathway or the plastidial methylerythritol phosphate pathway. Phenolic compounds are highly diverse, and include phenylpropanoids, coumarins, stilbenes, and flavonoids. Phenylpropanoids contain at least one aromatic ring with one or more hydroxyl groups, and are synthesized via the shikimate pathway alone or in combination with the melonate pathway. The nitrogen-containing compounds are also highly diverse, and include alkaloids, non-protein amino acids, and amines. Their biosynthetic pathways usually have multiple routes, frequently starting from amino acids. To this point, more than 36,000 terpenoids, 12,000 alkaloids, and 10,000 flavonoids have been found, although this represents only a fraction of what exists in nature (Chen et al. 2007; Fang 2011).
Spatial and Temporal Patterns of Secondary Metabolism
The biosynthesis and accumulation of secondary metabolites are usually tissue- and developmental stage-specific. For example, in cotton, the phytoalexin gossypol accumulates largely in pigmented glands of aerial organs and in epidermal and subepidermal layers of roots (Xu et al. 2004). During embryo (seed) development, transcripts of genes encoding (+)-δ-cadinene synthase (cad1-C), the first enzyme committed to gossypol biosynthesis, and CYP706B1, a P450 monooxygenase that catalyzes the second step, are undetectable until 20 d post-anthesis. Pigmented glands, which accumulate gossypol, consistently appear shortly afterwards (Tan et al. 2000; Luo et al. 2001). Artemisinin, a sesquiterpene lactone found solely in Artemisia annua, is a highly-effective drug used in the treatment of malaria. Both the accumulation of artemisinin and the expression of its biosynthetic pathway genes, including amorpha-4,11-diene synthase (ADS), CYP71AV1, double bond reductase 2 (DBR2), and aldehyde dehydrogenase 1 (ALDH1), are highly active in the particular glandular trichomes distributed on leaves, stems, and inflorescences of Artemisia annua (Olsson et al. 2009). In Arabidopsis, root or aerial tissue-specific expression of terpene synthases determines the distribution of terpenoids. In flowers, where most terpenoids are detected, with (E)-β-caryophyllene being predominant, two sesquiterpene synthases of TPS21 and TPS11 are responsible for the formation of the sesquiterpenes, but with different tissue-specific expression patterns. TPS21 is expressed mainly in the stigma of open flowers and not much in sepals, whereas TPS11 is expressed in intrafloral nectaries and oveles. In roots, the volatile 1,8-cineole is produced by the specifically root-expressed monoterpene synthase AtTPS-Cin (Chen et al. 2004; Tholl et al. 2005; Tholl and Lee 2011). All of the above cases show secondary metabolites being biosynthesized and accumulated at the same place. Interestingly, the tobacco (Nicotiana tabacum) alkaloid nicotine is mainly accumulated in the leaf vacuoles, but its biosynthesis occurs in root tissues. This root-produced nicotine is then transported to the aerial parts of the tobacco plant and is finally stored in the central vacuoles of leaves (Hashimoto and Yamada 2003). Recently, the MATE-type transporter Nt-JAT1 was shown to be responsible for nicotine translocation in aerial parts of the plant, and its deposition in vacuoles (Shitan et al. 2009).
The biosynthesis of secondary metabolites is often in response to environmental factors, and the formation of defensive phytoalexins can be stimulated by herbivoral damage and microbial infection. The treatment of cotton plants or suspension-cultured cells with a fungal elicitor induced the expression of gossypol pathway genes and increased the accumulation of gossypol (Luo et al. 2001; Xu et al. 2004). Similar induction was also observed in ginkgo (Ginkgo biloba), with treatments of fungal elicitors resulting in the formation of greater amounts of bilobalide, ginkgolide A, and ginkgolide B (Kang et al. 2009). In rice (Oryza sativa), 13 sesquiterpenes were detected upon elicitor treatment, in comparison to trace amounts of sesquiterpenes under normal conditions (Cheng et al. 2007). The recently-discovered maize sesquiterpenoid phytoalexins, zealexins, also showed increased accumulation by a wide range of elicitors (Huffaker et al. 2011).
Transcription Factors Involved in Secondary Metabolism
The spatial, temporal, and inducible formation of secondary metabolites and the transcripts of corresponding biosynthetic genes are under tight regulation at different levels, in which transcriptional regulation via transcription factors has been investigated intensively. Transcription factors are sequence-specific DNA-binding proteins that interact with the regulatory (often promoter) regions of the target genes, and modulate the rate of transcriptional initiation by RNA polymerase (Vom Endt et al. 2002). They can integrate internal (often developmental) and external (environmental) signals to regulate enzyme gene expression, thus controlling the specific accumulation of secondary metabolites. Some transcription factors of different types form complexes to activate or suppress downstream gene expression. Several families of transcription factors have been described to be regulators of plant secondary metabolism (Table 1).
MYB transcription factors are characterized by varying numbers of the MYB DNA-binding domain, which consists of up to four imperfect repeats of 52 amino acids (Feller et al. 2011). The MYB family proteins can be divided into four families, and several members of the R2R3 family have been reported to be regulators of multiple biosynthetic pathways in various plant species. In Arabidopsis, AtMYB113, AtMYB114, AtMYB75, and AtMYB90 are involved in controlling anthocyanin content in vegetative tissues by activating the entire phenylpropanoid pathway (Borevitz et al. 2000; Stracke et al. 2001; Tohge et al. 2005; Gonzalez et al. 2008), whereas TT2 affects proanthocyanidins in the seed coat by regulating BANYULS (BAN) expression, which encodes nicotinamide adenine dinucleotide phosphate-dependent leucoanthocyanidin reductase (Nesi et al. 2001). Members of this family also regulate glucosinolate biosynthesis, with AtMYB29 and AtMYB76 regulating aliphatic glucosinolate biosynthesis in aerial issues (Gigolashvili et al. 2008), and AtMYB34, AtMYB51, and AtMYB122 controlling indolic glucosinolate production in roots and late-stage rosette leaves through regulation of tryptophan (Trp) metabolizing P450 genes of CYP79B2, CYP79B3, and CYP83B1 (Celenza et al. 2005; Gigolashvili et al. 2007; Dubos et al. 2010). In other plants, for example, tobacco (N. tabacum), MYBJS1 was found to be involved in phenylpropanoid regulation (Galis et al. 2006); and in the Asiatic hybrid lily (Lilium spp.), MYB6, and MYB12 are positive regulators of biosynthesis and determine the organ- and tissue-specific accumulation of anthocyanin (Yamagishi et al. 2010).
bHLH transcription factors often interact with MYB family proteins to form a complex, and then regulate the downstream expression of target genes (Feller et al. 2011). A well-characterized example is the transcriptional regulation of anthocyanin, which is mostly studied in genes of the anthocyanin pathway in Arabidopsis. The bHLH proteins GL3, eGL3, and TT8 interact with MYB proteins in the presence of a WD40 repeat containing protein TTG1, forming a transcriptional regulation complex which activates anthocyanin biosynthetic genes (Gonzalez et al. 2008; Dubos et al. 2010).
Another important bHLH regulator is MYC2. MYC2 has widely been found to directly or indirectly participate in secondary metabolism in multiple plant species. In Arabidopsis, it interacts with DELLA proteins to integrate gibberellin (GA) and jasmonate (JA) signaling pathways to upregulate the expression of sesquiterpene synthase genes in flowers (Figure 1, Hong et al. 2012). In Catharanthus roseus, CrMYC2 binds directly to cis-elements in the ORCA3 gene promoter, thereby controlling the expression of several terpenoid indole alkaloid (TIA) biosynthesis genes, and TIA accumulation (Zhang et al. 2011). A similar pattern was also observed in common tobacco (N. tabacum): the NbMYC2a/b proteins enhanced nicotine biosynthesis by upregulating the ORCA-related NIC2 locus, which encodes an AP2/ERF family transcription factor (Shoji and Hashimoto 2011). Furthermore, three more bHLH transcription factors (NbbHLH1, NbbHLH2, and NbbHLH3) are correlated to nicotine accumulation according to expressed sequence tag screening via virus-induced gene silencing, of which NbbHLH1 and NbbHLH2 are positive regulators by their binding to G-box elements in the putrescine N-methyltransferase promoter, whereas BbbHLH3 is a negative regulator (Todd et al. 2010).
The APETALA2/ethylene response factor (AP2/ERF) family transcription factors are characterized by their DNA-binding AP2 domain, which consists of approximately 60 conserved amino acid residues (Mizoi et al. 2012). The ORCA proteins of C. roseus are the AP2/ERF transcription factors involved in secondary metabolism. ORCA3 controls the expression of multiple genes involved in the TIA biosynthesis pathway, including both the primary plastidial isopentenyl pyrophosphate pathway and the secondary TIA pathways (van der Fits and Memelink 2000). It activates strictosidine synthase (Str) gene expression by directly binding to the JA- and elicitor-responsive element (JERE) in its promoter region (van der Fits and Memelink 2001). Overexpression of this transcription factor has been found to promote accumulation of TIAs in suspension cells (van der Fits and Memelink 2000).
Recently, AP2/ERF proteins were reported to participate in secondary metabolism regulation in other plant species, too. In A. annua, two JA-responsive AP2/ERF proteins, AaERF1 and AaERF2, were shown to bind to the CRTDREHVCBF2 (CBF2) and RAV1AAT (RAA) motifs in promoters of genes encoding amorpha-4,11-diene synthase (ADS) and CYP71AV1, and to activate expression of both genes. Overexpression of AaERF1 or AaERF2 led to increased accumulation of artemisinin and artemisinic acids (Yu et al. 2012). In tobacco, a locus containing at least seven AP2/ERF genes is related to nicotine biosynthesis, of which ERF189 and ERF221 are most efficient (Shoji et al. 2010).
Many WRKY family transcription factors play important roles in mediating plant responses to stress factors, and the inducible expression patterns of WRKY genes supports their involvement in the regulation of biosynthesis of defense-related secondary metabolites. In cotton, GaWRKY1 is a positive regulator of gossypol biosynthesis, and it binds to the promoter of one of the (+)-δ-cadinene synthase genes (Xu et al. 2004). In A. annua, AaWRKY1 is also a positive regulator of artemisinin biosynthesis by binding to and activating the corresponding sesquiterpene synthase gene promoter (Ma et al. 2009). In tobacco, two transcription factors, WRKY3 and WRKY6, were found to be related to volatile terpene production (Skibbe et al. 2008). Aside from functioning in terpenoid biosynthesis pathways, the WRKY family proteins also participate in the regulation of other pathways. For example, WRKY33 regulates the biosynthesis of camalexin, the most important phytoalexin of Arabidopsis (Qiu et al. 2008).
The plant-specific NAC (for NAM, ATAF1/2, and CUC2) domain-containing proteins represent one of the largest transcription factor families in plants. They are known to control multiple processes in plants, including developmental programs and abiotic/biotic stress responses (Olsen et al. 2005). Recently, an Arabidopsis NAC family protein ANAC042 was reported to be a regulator of camalexin, which represents the first NAC family transcription factor involved in plant secondary metabolism regulation. The anac042 mutant, which upon elicitation failed to accumulate as much camalexin as the wild type, was highly susceptible to Alternaria brassicicola infection, possibly due to deficient expression of camalexin biosynthetic genes CYP71A12, CYP71A13, and CYP71B15 (Saga et al. 2012).
The SQUAMOSA Promoter Binding Protein-Like (SPL) family transcription factors, targeted by microRNA 156 (miR156), participate in a broad range of developmental processes in Arabidopsis (Wang et al. 2009). A recent investigation showed that SPL9 is a negative regulator of anthocyanin accumulation. By interfering with the MYB-bHLH-WD40 transcriptional–activation complex and then DFR gene expression, SPL9 and its regulator miR156 control anthocyanin metabolism in an age-dependent manner (Gou et al. 2011).
Transcriptional Regulation of Representative Secondary Metabolite Groups
Anthocyanin accumulation in plants is modulated by the WD-repeat/bHLH/MYB complex. In Arabidopsis, the WD-repeat protein TTG1 recruits bHLH transcription factors of GL3, TT8, or EGL3, as well as MYB transcriptional factors of MYB75, MYB90, MYB113, and MYB114 to form a complex to upregulate dihydroflavonol 4-reductase, anthocyanin synthase, and glucosyltransferase gene expression as well as anthocyanin accumulation (Dubos et al. 2010). Interestingly, through tissue- or developmental stage-specific expression of the complex component homologues, accumulation of anthocyanin can be specified, leading to colorful leaves, flowers, and fruits. For example, in apple (Malus domestica), anthocyanin accumulation in fruit flesh is determined by the MYB family transcription factor MYB10, whereas in fruit skin, it is controlled by another skin-specific transcription factor MYB1 (or MYBA), although both of these are in the same family and share high sequence identity (Takos et al. 2006; Ban et al. 2007; Lin-Wang et al. 2010). In lily, the accumulation of anthocyanin pigmentation in tepals, filaments, and styles is determined by tissue-specific expression of LhMYB12, whereas expression of LhMYB6 correlates with anthocyanin spots in tepals and light-induced pigmentation in leaves (Yamagishi et al. 2010). In grape (Vitis vinifera), seasonal expression of three MYB transcription factors of VvMYBPA1, VvMYB5a, and VvMYB5b in response to light conditions is correlated to anthocyanin accumulation in the skin and seeds (Matus et al. 2009). Furthermore, insertion of a Copia-like retrotransposon adjacent to the MYB transcription activator Ruby was shown to result in cold-dependent accumulation of anthocyanins in blood oranges (Butelli et al. 2012).
Terpenoid indole alkaloids
Catharanthus roseus is well known for its production of TIAs including ajmalicine, catharanthine, serpentine, and vindoline, which are efficient inhibitors of microtubule polymerization and are used to treat certain cancers. Tryptamine is catalyzed by Trp decarboxylase (TDC) hybrid with the MEP pathway product secologanin by strictosidine synthase to form strictosidine, which is then transformed to the monomeric and dimeric TIAs (Suttipanta et al. 2011).
Production of TIAs and of known biosynthetic genes is inducible by JA treatment. Two cis-elements, BA and JERE (JA and ethylene-responsive element), were identified near the TATA box, of which the JERE region is absolutely required for JA induction whereas the BA region enhances induction. An elicitor-inducible but JA-independent MYB-like factor CrBPF1 binds to this BA region (van der Fits and Memelink 2000), whereas the JA-dependent AP2/ERF domain-containing ORCA transcription factors (ORCA2, ORCA3) can bind to the JERE region of the strictosidine synthase gene promoter (van der Fits and Memelink 2001). Recently, the bHLH family protein CrMYC2 was found to bind to the JA-responsive element in the ORCA3 promoter and activate gene expression, in turn regulating a subset of alkaloid biosynthesis genes (Zhang et al. 2011).
By yeast one-hybrid screening with the elicitor responsive part of the TDC promoter, three members of the Cys2/His2 type (transcription factor IIIA-type) zinc finger proteins, ZCT1, ZCT2, and ZCT3, were isolated. They were shown to exhibit binding activity to the TDC and STR promoters in a sequence-specific manner, and repress their activity. In addition, they have also been shown to repress the activating activity of ORCAs on the STR promoter (Pauw et al. 2004). Furthermore, the root-specific and phytohormone responsive WRKY transcription factor CrWRKY1 is also able to bind to the W-box in the TDC promoter and activate its expression. Interestingly, overexpression of CrWRKY1 in C. roseus hairy roots upregulated several key TIA pathway genes and the transcriptional repressors ZCT1, ZCT2, and ZCT3, but repressed the activators ORCA2, ORCA3, and CrMYC2, thus increasing the level of serpentine accumulation threefold but significantly decreasing catharanthine and tabersonine. This suggests that CrWRKY1 plays a key role in determining the root-specific accumulation of serpentine in C. roseus plants (Suttipanta et al. 2011).
Glucosinolates and camalexin
The indolic secondary metabolite camalexin is the major phytoalexin of Arabidopsis. It shares indole-3-acetaldoxime (IAOx) as a precursor with indolic glucosinolates and auxin biosynthesis, which is produced from Trp by cytochrome P450 CYP79B2 and CYP79B3 (Glawischnig et al. 2004). IAOx is further converted to indole-3-acetonitrile (IAN) by CYP71A13 (Nafisi et al. 2007) and CYP71A12 (Millet et al. 2010), and then conjugated to glutathione by glutathione S-transferases (Chen et al. 2012). The subsequent hydroxylation and final conversion to camalexin via dihydrocamalexic acid are catalyzed by P450 CYP71B15/PAD3 (Bottcher et al. 2009).
Biosynthesis of camalexin can be induced by infection with bacterial pathogens such as Pseudomonas syringae and Erwinia carotovora, and fungi such as Alternaria brassicicola and Botrytis cinerea, as well as through exogenous factors like herbicides, heavy metal ions, and UV-B irradiation. The DNA-binding-with-one-finger (DOF) transcription factor OBP2 is expressed in Arabidopsis organs, leaves, roots, flowers, and petals, and its expression is responsive to herbivores, pathogens, and the phytohormone JA. Overexpression of OBP2 was shown to activate expression of CYP83B1, whereas silencing OBP2 led to suppression of the expression of CYP83B1, suggesting that OBP2 is part of a regulatory network that regulates glucosinolate biosynthesis in Arabidopsis (Skirycz et al. 2006). MYB transcription factor ATR1 (AtMYB34), identified in a screen for altered Trp metabolism, controls production of glucosinolates in Arabidopsis. Its dominant overexpression allele atr1D confers elevated expression of genes encoding CYP79B2, CYP79B3, and CYP83B1. Overexpression of ATR1 increased indolyl glucosinolate and IAA accumulation but not aliphatic glucosinolates, whereas the recessive atr1–2 mutant showed reduced gene expression and a lower accumulation of indolyl glucosinolates (Celenza et al. 2005).
The WRKY transcription factor WRKY33 was found to bind to the promoter regions of CYP71B15/PAD3 in response to P. syringae infection (Qiu et al. 2008). The wrky33 mutant was found to be highly susceptible to B. cinerea and A. brassicicola infection, and its overexpression increased resistance to these pathogens (Zheng et al. 2006). The mitogen-activated protein (MAP) kinase cascade plays a critical role in WRKY33 regulation. In response to P. syringae infection, MPK4 is activated to phosphorylate MKS1, releasing WRKY33 from the MKS1/WRKY33 complex, which targets the CYP71B15 promoter and stimulates camalexin biosynthesis (Qiu et al. 2008). Moreover, WRKY33 is also phosphorylated by MPK3/MPK6 in response to B. cinerea infection (Mao et al. 2011). It is likely that different pathogen infection signals activate various MAP cascades and converge to WRKY33 activation, leading to the induction of camalexin biosynthesis.
Recently, the Arabidopsis NAC transcription factor ANAC042 was shown to participate in the regulation of camalexin biosynthesis. The anac042 mutant failed to accumulate camalexin at the level of wild-type plants, and was highly susceptible to A. brassicicola infection. Expression of biosynthetic pathway genes CYP71A12, CYP71A13, and CYP71B15/PAD3 were reduced in the mutants, indicating that the camalexin defects were at least partly a result of reduced expression of these P450 genes. Although ANAC042 expression can be induced by pathogen infection, kinase inhibitor treatment and mutant analysis showed that ANAC042 might function independently of the WRKY33-dependent signaling pathways (Saga et al. 2012).
Transcription factors play a paramount role in regulating genes involved in likely all aspects of plant growth and development, including secondary metabolism. In recent years, an increasing number of transcription factors and the underlying mechanisms of plant secondary metabolism regulation have been elucidated. However, other mechanisms regulating specific pathways do exist. For example, the DnaJ Cys-rich domain containing plastid protein OR enhances the differentiation of non-colored plastids into chromoplasts for carotenoid accumulation in cauliflower (Brassica oleracea) (Lu et al. 2006); and in soybean (Glycine max), the dominant inhibition of seed coat pigmentation is due to inverted clusters of three chalcone synthase (CHS) genes, which results in the silencing of the expression of all CHS genes (Tuteja et al. 2009). Investigation of the network controlling the biosynthesis, transportation, accumulation, and release of secondary metabolites will further our understanding of plant adaptation to changing environments and the interaction among plants and other organisms.
Many secondary metabolites are highly valuable to humans. However, due to their low concentrations in plant tissues, direct extraction of these metabolites is usually expensive and inefficient. Engineering biosynthetic pathways in microbes or other plant species via a synthetic approach is a recent phenomenon that holds promise, but successful examples of this approach are still limited, partially due to the complexity of biosynthesis pathways and the frequent toxicity of metabolic intermediates. Increasing metabolite production by engineering transcription factors is a promising alternative. By ectopic expression or overexpression of regulatory proteins, we can expect to upregulate the entire pathway.
(Co-Editor: Xiaoquan Qi)
This work was supported by the State Key Basic Research Program of China (No. 2007CB108800), the National Natural Science Foundation of China (No. 30630008).