Introduction

Lonicera macranthoides Hand-Mazz is a widely distributed plant species in southern China and has been used in traditional Chinese medicine. It is also an ingredient for functional foods and cosmetics. Dried or fresh early-stage flowers of L. macranthoides are used as an antitoxin, antibacterial, and antiviral reagent. Chlorogenic acid (CGA, 3-caffeoylquinic acid) is the main active ingredient in L. macranthoides and exhibits a strong antioxidant property when used as food (Azzini et al. 2007). Previous studies have shown that CGA is rapidly absorbed in rats that were orally administered pure CGA (Lafay et al. 2006) or plant extract (Yang et al. 2004). Statistical analyses indicated that the administration of CGA resulted in a significant decrease in blood pressure (Onakpoya et al. 2015; Zhao et al. 2012). CGA also exhibits anti-inflammatory activity in human peripheral blood mononuclear cells (Albert et al. 2002; Hammer and Birt 2014). In addition, CGA is believed to play important roles in free radical scavenging (Tamagnone et al. 1998). These findings indicate that CGA provides significant benefits to human health. There are relatively high levels of CGA produced in L. macranthoides, which is almost two times higher than that in L. japonica (Zhou and Tong 2003). Therefore, L. macranthoides is considered a prime resource for natural CGA.

CGA is synthesized in various plant species, such as tomato, tobacco, eggplant, apple, pear, plum, coffee, and artichoke (Schutz et al. 2004; Tamagnone et al. 1998; Wang et al. 2003). Although major progress has been made recently in the understanding of CGA biosynthesis pathway, the key biosynthetic route of CGA remains controversial. First, the universal precursor, p-coumaroyl-CoA, is synthesized under the catalysis of phenylalanine ammonia-lyase (PAL), 4-coumarate CoA ligase, and cinnamate 4-hydroxylase. Subsequently, three metabolic pathways have been postulated. In the first route, hydroxycinnamoyl-CoA quinate hydroxycinnamoyl transferase (HQT) catalyzes the formation of CGA from caffeoyl-CoA and quinic acid (Niggeweg et al. 2004). Caffeoyl-CoA is supplied by the combined activities of hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyl transferase (HCT) and p-coumaroyl ester 3′ hydroxylase (C3′H) via a p-coumaroyl shikimate intermediate (Mahesh et al. 2007). The second proposed route is based on the synthesis of p-coumaroyl quinate by HCT, followed by hydroxylation via C3′H (Hoffmann et al. 2003; Zenk et al. 1980). In the third suggested route, caffeoyl glucoside serves as an activated intermediate (Villegas and Kojima 1986).To date, several genes in the above-mentioned routes have been investigated for their role in CGA biosynthesis in many species. Lepelley et al. (2007) determined the correlations between the gene expression of HCT, HQT, C3H1 and CCoAOMT1 and the CGA contents in different tissues at different development stages in coffee. They found that higher HQT expression appears to be more closely correlated with CGA accumulation. Genetic transformation results also demonstrated that PAL and HQT were key enzymes associated with CGA production in tobacco and tomato (Howles et al. 1996; Niggeweg et al. 2004). Peng et al. (2010) isolated the LjHQT gene and showed that the tissue distribution of HQT is correlated with the pattern of CGA abundance in L. japonica, indicating that HQT is an indispensable gene in CGA biosynthesis.

A survey of genetic resources of Lonicera identified a mutant called “Jincuilei” with several never-opening flowers. This mutant has a higher CGA content of 6.0% compared with 4.0% in the wild-type (Wang et al. 2009). The higher CGA content in “Jincuilei” indicates that this species can be used as a natural resource for CGA. Similarly, L. macranthoides is a plant species with high levels of CGA and may potentially serve as a prime plant model for investigations on CGA biosynthesis. According to our previous study (Chen et al. 2015), the LmHQT1 gene (CL74326 Contig1) is a candidate gene associated with CGA biosynthesis. Therefore, functional characterization of LmHQT1 can help us understand the CGA biosynthetic mechanism in L. macranthoides. In the present study, the LmHQT1 gene was cloned and transferred into L. macranthoides, and positive transgenic plants were identified. The CGA contents were investigated in LmHQT1-overexpressing L. macranthoides plants. The results of this study provide novel insight into the regulation of CGA biosynthesis and can be used to develop a method for increasing CGA accumulation via genetic engineering.

Materials and methods

Plant materials

L. macranthoides (cv Yu lei 1#), grown in a greenhouse, was used in this study. The tissues at different developmental stages, including buds, young flowers (YF, 1 day after anthesis), mature flowers (MF, 5 days after anthesis), young leaves (YL, yellow green leaves), leaves (L, green leaves) and old leaves (OL, dark green leaves) were collected from 1-year-old seedlings. All collected samples were frozen in liquid N2 and stored at −80 °C until determination of the CGA content and qRT-PCR analysis.

Cloning the cDNA of LmHQT1 and sequence analysis

Total RNA was extracted from leaves of L. macranthoides using the TRIzol reagent (Invitrogen, Germany), following the manufacturer’s instructions. After DNaseI treatment, the first cDNA strand was generated using RevertAid™ First Strand cDNA Synthesis Kit (Fermentas, USA) according to the manufacturer’s protocol. The full length of the LmHQT1 gene was amplified by PCR using the following primers: forward, 5′-ATGGGAAGTGAAGGAAGTGTGAAGA-3′ and reverse 5′-TCAGAACTCGTACAAACACTTCTCAAA-3′. The PCR product was purified, cloned into the pMD18-T vector (TaKaRa) and sequenced.

For sequence analysis, the ortholog can be searched on the website by Blast alignment (http://www.ncbi.nlm.nih.gov/BLAST/) against our nucleotide or protein datasets. Conserved domain database (http://www.ncbi.nlm.nih.gov/Structure/cdd/) was used for functional annotation of the proteins. Multiple sequence alignment was performed using ClustalX (version 2.0.10) and the MEGA program (version 5.1) was utilized for molecular phylogenetic tree construction based on Neighbor-Joining method.

Plant transformation

Transgenic plants were obtained by Agrobacterium tumefaciens-mediated gene transfer system. To construct an overexpression vector, the coding sequence of LmHQT1 was ligated into the modified binary vector pLP100 under the CaMV35S promoter (kindly supplied by Prof. Zhengguo Li of Chongqing University), which carries the nptII selectable marker conferring kanamycin resistance. After validation by PCR amplification and sequencing of insert fragment, the plant expression vector was obtained and named as pLP100-35S-LmHQT1. This vector was transformed into A. tumefaciens strain EHA105 by using the freeze–thaw method. Plant transformation was conducted as follows: (i) Pre-culture. Young leaves of L. macranthoides were sterilized using 1% HgCl2 for 4 min and rinsed four times with sterilized distilled water. Subsequently, they were sliced into 2 cm2 pieces and pre-cultured on MB medium supplemented with 1.5 mg/L 2,4-dichlorophenozyacetic acid (2,4-D) and 1 mg/L kinetin (KT) for 2 days in the dark. MB medium, a modified solution based on Murashige and Skoog (MS) and B5 medium, contains 1,400 mg/L KNO3, 150 mg/L CaCl2·2H2O, 1000 mg/L Ca(C6H11O7)2, 220 mg/L Mg(NO3)2, 410 mg/L KH2PO4, 410 mg/L NH4NO3, 35 mg/L NaH2PO4·H2O, 200 mg/L (NH4)2SO4, 0.5 mg/L nicotinic acid, 100 mg/L inositol, 0.5 mg/L pyridoxine hydrochloride, 0.1 mg/L aneurine hydrochloride, and 2 mg/L glycine. (ii) Agrobacterium inoculation. The A. tumefaciens strain, EHA105, harboring the binary plasmid, pLP100-35S-LmHQT1, was cultured in liquid LB medium overnight at 28 °C. They were collected by centrifugation at 6000 rpm for 5 min and suspended in liquid MB medium supplemented with 1.5 mg/L 2,4-D, 1 mg/L KT, and 100 μmol/L acetosyringone (AS). The optimal cell density of Agrobacterium for infection was 0.4 at OD600. Explants were immersed in the bacterial suspension for 8 min under constant shaking and then co-cultured for 2 days in darkness. (iii) Shoot regeneration and kanamycin-resistance selection. After 2 days co-culture, the explants were transferred and placed on selection medium, which contains MB medium supplemented with 2.0 mg/L 6-benzyladenine (6-BA), 0.2 mg/L indolebutyric acid (IBA), 15 mg/L kanamycin (Kan), and 600 mg/L cefotaxime (Cef). The culture conditions were maintained at 25 ± 1 °C under a 12 h photoperiod with light intensity of 2000–3000 lx. The subculture period was maintained at 1-week intervals. The kanamycin-resistant calli/shoots were induced after subculture. (IV) Root induction. After about 50–70 days, the resistant adventitious shoots were transferred into 1/2 MB medium supplemented with 0.1 mg/L indole-3-acetic acid (IAA), 2.0 mg/L IBA, 100 mg/L activated carbon (AC), 15 mg/L Kan, and 600 mg/L Cef for root induction (see protocol in Fig. 1). Well-rooted explants were used for further analysis.

Fig. 1
figure 1

Optimized protocols of Agrobacterium tumefaciens-mediated genetic transformation system in Lonicera macranthoides Hand-Mazz

Full size image

Transgenic lines were selected by GUS (β-glucuronidase) staining and PCR detection. In brief, the β-gluc gene was amplified by PCR from seven kanamycin-resistant seedlings using the following primers: forward, 5′-ACCTCTCTTTAGGCATTGGTTTC-3′ and reverse 5′-GCACACTGATACTCTTCACTCCAC-3′. To validate its accuracy, another insert sequence, which contains the 35S promoter and part of the LmHQT1 gene, was amplified using the following primers: forward, 5′-GCTCCTACAAATGCCATCATTGC-3′ and reverse 5′-CGAGAGCAGACCTCA AGTACTCAT-3′.

Real-time quantitative PCR

Total RNA was extracted from different tissues at different developmental stages in L. macranthoides or transgenic lines. Total RNA isolation and cDNA synthesis were performed following the above-mentioned methods. Gene-specific primers for LmHQT1 (forward, 5′-CCAACCAGACGAGCAAGTTA-3′ and reverse, 5′-CCATAG GTGTGGCTGTGAAC-3′) were designed by online software (https://www.genscript.com/ssl-bin/app/primer). qRT-PCR was carried out as described previously (Chen et al. 2015), and three biological replicates were performed in all qPCR experiments. Relative expression levels were calculated based on the 2−ΔΔCt method using tubulin (GenBank Accession No. KR233012) as a reference gene and the primers were as follows: forward, 5′-CCACATCTGTTGTTGAGCCT-3′ and reverse, 5′-GCGCCTGCAGATATCATA GA-3′.

Determination of the CGA content by high-performance liquid chromatography (HPLC)

HPLC was employed to investigate the CGA contents from the materials including flowers and leaves at different developmental stages in L. macranthoides, or leaves in wild-type and transgenic lines. The detailed determination methods and data processing were carried out as described previously (Chen et al. 2015).

Results

Isolation and sequence analysis of the LmHQT1 gene

Based on the annotated sequence of the HQT-homologous gene identified from transcriptome data in L. macranthoides (Chen et al. 2015), a 1344 bp full-length cDNA of the HQT gene was obtained by RT-PCR and tentatively designated as LmHQT1 which was deposited in GenBank under accession number KR233011. The cDNA of LmHQT1 contained a 1344 bp open reading frame (ORF), which encoded a protein of 447 amino acid residues using Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, USA). Further analysis via NCBI CCD and the online tool Pfam (http://pfam.xfam.org/search/sequence) showed that the LmHQT1 protein harbored a transferase domain (Accession No. PLN02663 or pfam02458). This structural domain encompassed amino acids 1–429 and was involved in CoA-dependent acyltransferase (Fig. 2). To search homologs of LmHQT1 in L. macranthoides, BLAST analysis was performed against the NCBI database. NCBI BLASTP indicated that the LmHQT1 protein was most closely homologous to the HQT gene in Lonicera japonica (LjHQT, ACZ52698.1). It has been also found that the HQT genes in Nicotiana sylvestris and Solanum lycopersicum showed high similarity to LmHQT1. Pairwise alignment displayed that LmHQT1 showed 99, 72 and 71% amino acid identity with LjHQT, NsHQT (XP_009800587.1) and SlHQT (NP_001234850.1), respectively (Fig. 2).

Fig. 2
figure 2

Sequence analysis of LmHQT1 in Lonicera macranthoides. Multiple sequence alignment of LmHQT1 (KR233011), SlHQT (NP_001234850.1), and NsHQT (XP_009800587.1). The conserved domain (PLN02663 and pfam02458) is indicated by a straight line. The two conserved motifs are marked with red frames

Full size image

Given that LmHQT1 belongs to a plant transferase subfamily, the phylogenetic tree was constructed with 20 transferase gene sequences from several plant species. The results displayed that LmHQT1 and HCT show the highest homology. In addition, LmHQT1 exhibited close homology with hydroxycinnamoyl-CoA:hydroxyanthranilate N-hydroxycinnamoyltransferase (HHT) and anthranilate N-hydroxycinnamoyl/benzoyltransferase protein (HCBT) (Fig. 3).

Fig. 3
figure 3

LmHQT1 belongs to plant acyltransferase family. The phylogenetic tree was constructed using MEGA software (version 5.1) based on the neighbor-joining method. Values above the branches are bootstrap percentages (1000 replicates). Abbreviations and gene ID were as follows. LmHQT1, L. macranthoides hydroxycinnamoyl-CoA quinate hydroxycinnamoyl transferase; LjHQT (ACZ52698), L. japonica HQT; NtHQT (CAE46932), N. tabacum HQT; AtHCT (NP_199704.1), A. thaliana hydroxycinnamoyl-CoA: shikimate/quinate hydroxycinnamoyltransferase; NtHCT(CAD47830), N. tabacum HCT; AtHCBT (NP_200592.1), A. thaliana anthranilate N-hydroxycinnamoyl/benzoyl transferase; DcHCBT (CAB06430), D. caryophyllus HCBT; AtSHT (NP_179497.1) A. thaliana spermidine hydroxycinnamoyl transferase; AsHHT1 (BAC78633), A. sativa hydroxycinnamoyl-CoA: hydroxyanthranilate N-hydroxycinnamoyltransferase; AtCHAT (NP_186998.1), A. thaliana acetyl CoA: (Z)-3-hexen-1-ol acetyltransferase; VlAMAT(AAW22989), V. labrusca anthraniloyl-CoA: methanol acyltransferase; AtHHT (NP_568587.2), A. thaliana omega-hydroxypalmitate O-feruloyl transferase; AtSDT (NP_179932.1), A. thaliana spermidine disinapoyl acyltransferase; AtPMAT2 (NP_189609.1), A. thaliana phenolic glucoside malonyltransferase 2; At5AT1 (NP_200924.1), A. thaliana anthocyanin 5-aromatic acyltransferase 1; At3CAT1(NP_171890.1), A. thaliana coumaroyl-CoA: anthocyanidin 3-O-glucoside -6″-O-coumaroyltransferase 1; BanAAT (CAC09063), banana alcohol acyltransferase; HvACT (AAO73071), H. vulgare agmatine coumaroyltransferase; CbBEBT (AAN09796), C. breweri benzoyl-CoA: benzylalcohol O-benzoyltransferase; Lp3MAT1 (AAS77404), L. purpureum malonyl-CoA: flavonol 3-O-glucoside-6′′-O-malonyltransferase; LaHMT/HLT (BAD89275), Lupinus albus (−)-13 alpha-hydroxy multiflorine/(+)-13alpha-hydroxylupanine O-tigloyltransferase. The LmHQT1 protein was labeled in red frame

Full size image

CGA contents in different tissues

The CGA contents in flowers and leaves at different developmental stages were quantified. The CGA content was highest in young leaves, that is, 54.01 mg/g FW (fresh weight). However, it rapidly declined during leaf aging. In addition, the CGA concentration in young leaves was higher than that in flowers at different developmental stages, and the highest levels of CGA in flowers were found in buds, followed by young flowers and mature flowers (Fig. 4a).

Fig. 4
figure 4

The CGA contents and expression levels of LmHQT1 in different tissues of Lonicera macranthoides. a CGA contents in floral bud (Bud), young flower (YF), mature flower (MF), young leaf (YL), leaf (L) and old leaf (OL) were determined by HPLC. Three replications were performed for each examination, Duncan’s multiple range test was used to analyze the significance, and the different lowercase letters indicate significant (P < 0.05) differences between samples. b QRT-PCR was employed to analyze the expression patterns of LmHQT1 in different tissues including Bud, YF, MF, YL, L and OL

Full size image

Expression patterns of the LmHQT1 gene in different tissues

LmHQT1 transcripts in different flowers and leaves at different developmental stages were detected. qRT-PCR results showed that the LmHQT1 gene was differentially expressed in different tissues at different developmental stages. In brief, LmHQT1 was expressed at extremely high levels in young leaves and showed significantly lower levels during leaf senescence. Similarly, in flowers, LmHQT1 displayed a relatively strong expression at the bud stage and a slight decline as the tissues aged. Compared with young leaves, LmHQT1 was expressed at low levels in flowers at all developmental stages (Fig. 4b). Moreover, the transcripts of the LmHQT1 gene were positively correlated with the CGA content, indicating its important role in CGA biosynthesis.

Generation and identification of LmHQT1 transgenic plants

To characterize the function of the LmHQT1 gene, an overexpression vector was constructed and transgenic plants were generated via A. tumefaciens-mediated transformation developed previously in our laboratory. Throughout the processes including resistant calli induction, shoots induction, and rooting induction, approximately 17.2% of the explants were allowed to differentiate adventitious shoots in the presence of kanamycin (Fig. 5).

Fig. 5
figure 5

Transgenic plants generated by Agrobacterium tumefaciens-mediated transformation. a Callus induction for 4 weeks under selective callus-inducing medium. b Shoot induction for 2 weeks after transferring to selective shoot-inducing medium. c Shoot induction for 6 weeks cultured on selective shoot-inducing medium. d Root induction for 2 weeks after shoots were transferred into selective root-inducing medium

Full size image

To identify transgenic plants, 12 well-rooted seedlings (transgenic lines) selected randomly were used to detect insertion of the GUS gene. A specific 500 bp band was amplified from DNA of eight seedlings, indicating that the GUS fragment was integrated into the host genome successfully (Fig. 6a), which was also validated by GUS staining (Fig. 6c). An approximately 1100 bp band, which was attributed to the partial sequence of CaMV35S–LmHQT1 gene fusion, was observed. This finding suggested the integration of LmHQT1 into the genome of the detected well-rooted seedlings (Fig. 6b).

Fig. 6
figure 6

Identification of transgenic plants of Lonicera macranthoides. a PCR detection of GUS gene in candidate transgenic plants and controls. WT indicates wild-type plants, FP represents the seedlings with no roots on the selective medium, 1–12 indicate the 12 well-rooted seedlings on the selective medium, N is the negative control and P is the positive control which using plasmid of pLP100-35S-LmHQT1 as template in PCR reaction. b PCR amplification of the inserted target gene. WT indicates wild-type plants, N is the negative control, and 3, 4, and 5 are the three randomly selected positive transgenic plants used in the amplification of the GUS gene. c GUS staining of transgenic plants in Lonicera macranthoides, and arrows pointing at the blue parts indicate sites of GUS activity

Full size image

Overexpression of the LmHQT1 gene increased CGA contents in L. macranthoides

qRT-PCR analysis was employed to investigate the constitutive expression of LmHQT1 in transgenic lines. LmHQT1 transcripts accumulated in three individual lines, which were about 6- to 20-fold higher than that in wild type (Fig. 7a). To characterize the roles of the LmHQT1 gene in CGA biosynthesis, the CGA contents were determined in transgenic plants. Results revealed that the CGA contents significantly increased in transgenic plants, with 63.2, 74.9 and 66.0 mg/g FW in the leaves of overexpression lines 3, 4 and 5, respectively. These values were largely higher than that in wild type, which had a CGA content of 45.2 mg/g FW (Fig. 7b). Therefore, the results indicated that the LmHQT1 gene plays an important role in CGA biosynthesis.

Fig. 7
figure 7

The transcript levels of LmHQT1 and CGA contents in LmHQT1 overexpression plants in Lonicera macranthoides. a Gene expressions of LmHQT1 were performed by qRT-PCR. WT indicates wild-type materials, while OX-3, OX-4, and OX-5 represent three well-rooted explants on selective medium. Materials for each replication were harvested from three independent explants, and three independent replications were performed for each sample. b CGA contents were determined using HPLC. Student’s t test was used to determine significance, and two asterisks (**) indicate significant (P ≤ 0.01) differences

Full size image

Discussion

Lonicera macranthoides is an exceptionally rich source of CGAs and believed to be of fundamental importance to human health because CGA is associated with antioxidant activity and proved to have anti-inflammatory and anti-cancer properties (Cai et al. 2004; Chagas-Paula et al. 2011). The accumulation of CGAs is a precisely controlled process that varies considerably depending on tissues, developmental stages, and environmental conditions. Our previous study demonstrated that the CGA contents differed significantly among the tissues including leaves, stems, and flowers of L. macranthoides (Chen et al. 2015). In this study, the levels of CGA were extremely higher in the earliest stage tested (young leaves), but decreased sharply during aging and senescence of leaves (Fig. 4a). In addition, growth conditions exerted a remarkable effect on CGA accumulation in leaves. Clé et al. (2008) found that CGA levels rose in leaves when tomato plants were transferred from the growth room to a tunnel with higher light intensity. Moreover, the CQA content increases in response to abiotic stresses (Dixon and Paiva 1995). A previous study reported a large increase in diCQAs accumulation in the leaves when exposed to UV-C radiation in different globe artichoke genotypes (Moglia et al. 2008).

The CGA biosynthetic pathway has been proposed. HQT is considered to be one of the important enzymes for CGA synthesis in plants, which can act directly on caffeoyl CoA and quinic acid to generate CGA or synthesize p-coumaroyl quinate from p-coumaroyl CoA and quinic acid, and then convert to CGA via C3′H (Niggeweg et al. 2004).

HCT/HQT enzymes are encoded by a gene family composed of at least seven members in L. macranthoides (Chen et al. 2015).Of these members, LmHQT1 was cloned in the present study. The predicted protein showed a high level of identity to its tobacco and tomato orthologs (Fig. 2). We found that LmHQT1 belongs to the plant acyltransferase family, which has two conserved peptide motifs, including HXXXD and DFGWG (Hoffmann et al. 2003; St-Pierre et al. 1998). It is also observed that all three HQT genes used in Fig. 2 contained the same sequence (HTLSD) in the highly conserved HXXXD motif (Fig. 2), whereas an asparagine in place of a threonine was found in the conserved HTLSD box of the HQT gene in coffee (Lepelley et al. 2007). Phylogenetic analysis of the acyltransferase family indicated that LmHQT1 displayed a high level of similarity to HCBT (Fig. 3), which is a related enzyme of HCT and belongs to the subgroup D in Arabidopsis. Hydroxycinnamoyl groups can be transferred to acceptors generated from the shikimate pathway such as shikimate, quinate, and anthranilate via the two proteins (Hoffmann et al. 2003). Thus, the results provide fundamental information on characterizing the functions of LmHQT1.

In recent years, numerous studies have focused on the regulation of the CGA biosynthesis pathway at the transcription level. Reports revealed a relationship between HQT gene expressions and CGA levels in many plant species. DiCQAs, isomers of CGA, were observed to be in higher amounts at the early stage of grain development and then in decline as development progressed. Correspondingly, the transcripts of the CcHQT gene displayed the highest levels at the small green stage and then decreased as the grain matured, thereby indicating an association between CcHQT gene expression and CGA accumulation during grain development in Coffea canephora (Lepelley et al. 2007). Our previous study also showed positive correlations between the gene expressions of five HCT/HQT homologs and CGA contents in different tissues of L. macranthoides (Chen et al. 2015). In the present work, we found that the LmHQT1 gene was widely expressed and displayed a large variation in expression among different tissues or developmental stages. The transcript levels of LmHQT1 were highest in the young leaves and then fell sharply during senescence. Similarly, CGA accumulation decreased during the aging process of leaves (Fig. 4b), suggesting a potentially important role of LmHQT1 in CGA biosynthesis. However, Lepelley et al. (2007) reported that no significant variations in HQT transcripts at four different stages of leaf development in coffee. We believe that this difference was due to plant species. In addition, HQT transcripts can be modulated in response to abiotic stress. Clè et al. (2008) reported the involvement of the HQT gene in phenolic acids accumulation under UV stress in tomato leaves. Comino et al. (2009) demonstrated that the transcription of the HQT gene increased under UV-C treatment in globe artichoke. UV-C application leads to large increases in leaf DCQs (isomers of CGA) in globe artichoke (Moglia et al. 2008), indicating the central role of the HQT gene in CGA production in globe artichoke.

To identify the key genes associated with the CGA biosynthetic pathway, genetic transformation was performed. This technique has been recently applied in many species. Niggeweg et al. (2004) demonstrated that recombinant tobacco HQT shows higher affinity for quinate than shikimate. Downregulation of HQT gene in tomato resulted in a 98% reduction of the CGA content in leaves, suggesting that HQT was the principal route for CGA accumulation in Solanaceous species. Moreover, overexpression of HQT in tomato was observed to elevate the CGA levels up to 85% in leaves, and the transgenic plants displayed enhanced resistance to oxidative stress and bacterial pathogen. Sonnante et al. (2010) identified two novel HQT genes (HQT1 and HQT2) in artichoke. They noted that both of the genes showed much higher affinity for quinate over shikimate; hence quinate was used as an acyl acceptor in preference to shikimate. Moreover, HQT1 transcription was positively correlated with the CGA content, and overexpression of HQT1 in Nicotiana causes plants to accumulate higher amounts of CGA and cynarin (1,3-dicaffeoylquinic acid), suggesting that HQT1 is devoted to CGA synthesis. In the present study, to characterize the role of LmHQT1 in CGA production, the recombinant protein was successfully transformed into L. macranthoides based on an Agrobacterium-mediated transformation system established in our previous work (Fig. 1). This technique is a highly efficient and credible method for the genetic transformation of L. macranthoides. It has generated transgenic plantlets within 3 months and about 17% of explants can be well rooted on kanamycin selection medium (Fig. 5).The results of PCR analysis and GUS staining also demonstrated that the ORF of LmHQT1 was integrated into L. macranthoides plants (Fig. 6). qRT-PCR analysis showed that the transcript levels of LmHQT1 extremely increased in different transgenic lines by about 6- to 20-fold than that in wild types (Fig. 7a). Accordingly, the CGA contents in leaves of LmHQT1 overexpression lines increased by 40–60% compared with those of wild types (Fig. 7b). Our results confirmed the crucial role of LmHQT1 on the CGA biosynthetic pathway in L. macranthoides. However, Clé et al. (2008) considered that HQT activity was not rate limiting for CGA production in tomato under certain conditions, such as high light intensity rooms. Relatively, the 1.4- to 1.6-fold increase in CGA levels that we achieved in L. macranthoides was lower than that in tomato (up to twofold) (Niggeweg et al. 2004), thereby indicating that the difference in the regulation of HQT on CGA biosynthesis may be due to plant species. The overexpression of PAL in tobacco can lead to much higher levels of CGA (up to threefold) (Howles et al. 1996). However, PAL is the first step in the phenylpropanoid biosynthesis pathway, and its overexpression elevates other intermediate secondary metabolites. By contrast, HQT is the final step in CGA biosynthesis and only affects the amounts of CGA and its isomers. Therefore, our strategy to increase CGA level was considered more efficient. HCT showed very close homology to HQT. The reports of Comino et al. (2007) suggested that HCT may contribute to CGA biosynthesis in artichoke. However, HCT was observed to be insignificant to CGA accumulation in tomato (Niggeweg et al. 2004) and Arabidopsis (Hoffmann et al. 2003), but played a central role in lignin synthesis in tobacco (Hoffmann et al. 2004). Based on this controversial issue, further studies are necessary to characterize the role of HCT in CGA biosynthesis in L. macranthoides. Transcription factors were very recently found to regulate CGA synthesis. Li et al. (2015) reported that the constitutive expression of AtMYB11 and AtMYB12 in tobacco results in the increase in CGA levels by up to fivefold and twofold, respectively. Moreover, the gene expressions of HQT and HCT were upregulated in AtMYB11-overexpressing transgenic plants, indicating that HQT and HCT may be regulated byAtMYB11in tobacco plants. Our previous study also showed the coexpression of several transcription factors and HCT/HQT homologs in different tissues of L. macranthoides (unpublished data). Subsequent research could focus on the characterization of transcription factors that regulate CGA production.

Author contribution statement

ZXC and NT designed this work. ZXC, NT and GHL performed the experiment. ZXC and YQL prepared the original draft the manuscript. NT and ZQX revised the manuscript. All authors read and approved the final manuscript.