Introduction

Catharanthus roseus is valuable medicinal plant, as it is a source of indole alkaloids, and represents a potential model for molecular and enzymological studies (Svoboda and Blake 1975). While there have been many attempts at genetic engineering and in vitro culturing to enhance indole alkaloid production in this plant, most have been unsuccessful. This is in part due to (1) the compartmentalization of indole alkaloid biosynthesis in planta, and (2) the lack of a suitable, stable genetic transformation protocol for C. roseus.

Transient gene expression systems including protoplast (Galun 1981), biolistic bombardment (Ferrer et al. 2000) or Agrobacterium-mediated transient gene expression by agroinfiltration (Batra and Kumar 2003) have been used for functional promoter analysis (Hellens et al. 2005) or to produce recombinant proteins (Joh et al. 2005); agroinfiltration has been applied to C. roseus gene expression studies using, for example, vacuum infiltration of tobacco leaves for transient expression of strictosidine synthase (STR) (Wang and Li 2002). Vacuum infiltration has been used to study the expression of recombinant indole alkaloids enzymes in C. roseus leaves (Di Forie et al. 2004).

The C. roseus DAT gene encodes the enzyme acetyl-CoA:deacetylvindoline-4-O-acetyltransferase involved in the last step of vindoline biosynthesis. Vindoline production requires the participation of at least two cell types (i.e., idioblasts and laticifers) and the intercellular translocation of a pathway intermediate, and is predominantly found in leaves, stems and young buds (St-Pierre et al. 1998, 1999). Although light is not required for the formation of idioblast and laticifer cell types (Vazquez-Flota et al. 2000), light has been shown to induce DAT transcription, along with other indole alkaloid biosynthesis genes such as deacetoxyvindoline 4-hydroxylase (D4H) and tryptophan decarboxylase (TDC) (Vazquez-Flota and De Luca 1998; Vazquez-Flota et al. 2000).

Earlier studies describing the functional analysis of C. roseus indole alkaloid biosynthetic gene promoters have focused on proximal promoters and related motifs and transcription factors and their role in biotic or abiotic stress responsiveness in suspension culture systems. For example, the basal and proximal promoters of C. roseus STR and TDC genes have been studied. The STR promoter was shown to be regulated by auxins, methyl jasmonate (MJ) and fungal elicitors (van der Fits and Memelink 2000). The TDC promoter was shown to contain three domains associated with elicitor response, -160 to -99 (the main binding site) and -99 to -87 and -87 to -37 (two cooperative domains) (Ouwerkerk and Memelink 1999). Meanwhile, the STR promoter was shown to be the target for a few positive activating factors, including CrBPF-1 (box P-binding factor) (van der Fits et al. 2000) and ORCA2/ORCA3 (Octadecanoid-Responsive Catharanthus AP2/ERF-domain) and negative activating factors, including ZCT1, ZCT2, and ZCT3 (members of the Cys2/His2-type zinc finger protein family from C. roseus) (Pauw et al. 2004) and CrGBF (G-box binding factor) (Memelink and Gantet, 2007). Crgbf1 and Crgbf2 factors were shown to be STR repressors by interacting with a B-box upstream of an ORCA interaction-binding site (Siberil et al. 2001). Other analysis demonstrated that CrMYC1, encoding a BHLH type transcription factor that binds to a G-box on the STR basal promoter is induced by MJ and elicitors (Chatel et al. 2003). The geraniol 10-hydroxylase (GH) gene promoter contains many domains between (-191 and -147), (-266 and -188), (-318 and -266) that are the target for many activators (Suttipanta et al. 2007). Recently, analysis of a ca. 1.8 kb fragment of the DAT promoter, fused with the reporter gene GUS was described (Wang et al. 2010). These authors demonstrated transient expression in C. roseus suspension cultures and characterized three TGACG motifs involved in MJ signaling.

In the present work, we report on our functional analysis of the DAT promoter, and extend the in silico analysis up to 2.4 kb of the 5′ upstream region as well as into the DAT open reading frame (ORF). In doing so, we identify numerous specific regulatory motifs, and use a promoter deletion/promoter-ORF fusion approach to explore the spatial expression of DAT, observed through the activity of the reported gene GUS, both transiently in C. roseus tissue and through stable transformation of tobacco.

Materials and methods

Plant materials

Seeds of C. roseus c.v. Pacifica Pink (Ball Ducrettet, Thonon les Bains, France), were surface sterilized by immersion in ethanol for 2 min followed by immersion in 1.25% Na-hypochlorite for 20 min and finally washed three times in sterile distilled water. Seeds were kept moist in the dark for 1 day, sown in Petri dishes containing hormone-free MS culture medium, and incubated at room temperature in the dark for 3 days. Emerged seedlings were grown under a 16 h light/8 h dark cycle for 6 days, and then leaves were excised or infected by Agrobacterium. For experiments with Nicotiana tabacum c.v. Petit Havana, foliar discs were obtained from in vitro-grown plants from our laboratory (Elbez et al. 2002)

Bacterial strains and plasmid constructs

Agrobacterium tumefaciens LB4404, AGL1 and A. rhizogenes 15834 strains were used to obtain shoots, hairy roots and the transient gene expression system using two plant systems: C. roseus and tobacco. All transformation experiments were carried out using electroporation. After each cycle of electroporation with E. coli or Agrobacterium, individual colonies were selected on appropriate antibiotic (kanamycin 50 mg/L) and used to extract the plasmids harboring the different constructs. The binary vector pCAMBIA1305.1, in which the T-DNA region harbors the reporter gene GUS Plus™ encoding β-glucuronidase as well as hygromycin phosphotransferase (hptII) (http://www.cambia.org/daisy/cambia/585.html), was used for all constructs. In the standard vector, these two genes are driven by two 35S promoters. For DAT promoter constructs, the 35S promoter was excised from the vector and replaced with appropriate DAT promoter fragments in line with the reporter and selection genes.

DAT promoter constructs

For DAT promoter constructs, genomic clones of DAT were inserted in pBluescript II SK+ (pBSIISK+) and used to amplify fragments of different sizes. These clones, including the pBS::λ#6/SalI fragment (13.9 kb), the pBS#6/E#1′-4 (EcoRI-EcoRI) fragment (6 kb), the pBS#6/E#1- 1 (EcoRI–EcoRI) fragment (2.8 kb), the pBS#6/E,S#2 (EcoRI–SalI) fragment (3.6 kb) of pBS::λ#6/SalI, were all cloned in pBSIISK+.

The pCAMBIA 1305.1 vector was used as a control, as well as a template to create all DAT promoter constructs (Fig. 1). The pCAMBIA pDAT 812 construct was realized by replacing the forward p35S::GUS Plus promoter with an 812 bp DAT promoter fragment (plus 49 bp downstream of the TIS) upstream of the GUS Plus reporter gene. This fragment was amplified by primers PDAT-2 (gcggaattcaagctt-AGGTGTTCTTCCCGACG) and PDAT-8 (cagatctaccatgg-TCTCTGTCTCAACCGATATT) and the template pBS#6/E # 1-1. HindIII and BglII restriction sites were added to PDAT2 and PDAT8 primers, respectively, to facilitate ligation. The construct pCAMBIA pDAT 812 Δ p35S was built from pCAMBIA pDAT 812 by deletion of the CaMV35S X2 (used to drive hygromycin R expression) with XhoI and SalI. The pCAMBIA pDAT 2.3 construct was obtained by replacing the forward p35S::GUS Plus promoter with a 2,480 bp fragment of the DAT promoter (−2,431:+49) upstream of the GUS Plus reporter gene. This fragment ≈2.3 kb (−2,431:+49 bp) of the 5′ upstream region of the DAT gene was amplified from pBS -#6/E # 1-1 by the primers PDAT8 and M13-Reverse 24-mer after adding XbaI and BglII restriction sites (GGAAACAGCTATGACCATGATTAC). pCAMBIA pDAT 2.3 kb+dat was constructed by replacing the forward p35S::GUS Plus promoter with a 2,480 bp fragment of the DAT promoter (-2,431:+49) ligated to the DAT ORF, upstream of the GUS Plus reporter gene. The DAT coding region was amplified from pBSDAT3 (St-Pierre et al. 1998) by the primers DAT20 (5′ GCGagatctACTCTCCAAAACGTTGATCA 3′) and DAT21 (5′ CCGagatctGTACCTCCATTAGAAACAAATTGAAGTAGC 3′). These primers integrate a BglII restriction site (AGATCT), as well as a few nucleotides and delete the DAT gene stop codon. All constructs were inserted into pCAMBIA1305.1 by ligation of the corresponding fragments and then transformed into E. coli, Agrobacterium tumefaciens strains LB4404 and AGL1 and Agrobacterium rhizogenes 15834 strain. These constructs were verified by sequencing and digestion after transformation.

Fig. 1
figure 1

Schematic representation of DAT gene promoter constructs. The constructs used in this study are derivatives of the pCAMBIA1305.1 vector, which contains both hygromycin resistance (HYG-R) and GUS plus genes under individual constitutive 35S promoters (pr35S, pr35S x2). a Control (p35S) construct. b pDAT 812, in which the 35S promoter driving GUS Plus has been replaced with an 812 bp fragment of the DAT promoter. c pDAT 812 Δ35S, represents the same construct as in (b), except the 35S(×2) HYG-R component of the pCAMBIA1305.1 vector has been deleted. d pDAT 2.3, in which the 35S promoter driving GUS Plus has been replaced with an 2,431 bp fragment of the DAT promoter. e pDAT 2.3+dat in which the 35S promoter driving GUS Plus has been replaced with an 2,431 bp fragment of the DAT promoter plus the entire DAT ORF

Full size image

Agrobacterium tumefaciens containing the various constructs were used for transient gene expression by vacuum agroinfiltration into C. roseus and for stable genetic transformation of tobacco while Agrobacterium rhizogenes containing pCAMBIA pDAT 2.3+dat and pCAMBIA1305.1 (control) was used to give rise to transformed hairy roots from C. roseus leaves. In all cases, approximately 10 plates, with each containing 15 explants (leaves or foliar discs) were subjected to transformation. After transformation, approximately 20–30 shoots/roots were kept on selective medium for several cycles to be sure that they were transgenic. Several small leaves or hairy roots (20–30) for used GUS staining.

Transient gene expression in Catharanthus roseus

Agroinfiltration was used to examine DAT promoter expression in a transient system (Di Forie et al. 2004). For this, A. tumefaciens AGL1, harboring the different DAT promoter constructs described above, was infiltrated into various C. roseus tissues submerged in A. tumefaciens cultures (OD600 = 0.8) under vacuum (PVK-600 vacu box, Suisse). DAT promoter constructs used included pCAMBIA pDAT 812, pCAMBIA pDAT 2.3, pCAMBIA pDAT2.3+dat and the plasmid pCAMBIA 1305.1 as control. For each construct, 10 negative pressure cycles (10 min each) were applied. After agroinfiltration, explants were placed on imbibed filter papers in Petri dishes for 3 days at 25°C ± 1°C with long day light period (16 h:8 h) before GUS assay.

Stable transformation of Nicotiana tabacum

Tobacco (Nicotiana tabacum cv Petit Havana) leaf disks were infected with a disarmed A. tumefaciens LB4404 strain harbouring different DAT promoter constructs (including, pDAT 812, pDAT 2.3, pDAT 2.3+dat and pCAMBIA1305.1) according to Elbez et al. (2002). Foliar discs derived from young leaves were immersed in a suspension culture (OD600 = 0.1) for 10 min on a shaker (150 rpm). Leaf disks were co-cultured with A. tumefaciens on MS solid medium for 2–3 days, after which the disks were washed twice with sterile distilled water, (15 min/wash with shaking) and once with MS medium supplemented with cefotaxime (500 mg/L). Washed explants were placed on MS solid medium containing NAA (0.1 mg/L), BAP (1 mg/L) and cefotaxime (500 mg/L). After 3 weeks, emerging shoots were transferred to new plates (MS medium supplemented with 50 mg/L hygromycin) for selection. Shoots were subcultured every 3 weeks. Only shoots resistant to antibiotic for at least 3 months (3–4 cycles of selection on hygromycin-containing media) were used for subsequent applications.

Formation of hairy roots

Hairy roots of C. roseus were obtained as described by Guillon et al. (2006a, b) by wounding in vitro cultured leaves along the main vein with a scalpel previously dipped in solid colonies of A. rhizogenes 15834 harboring either pCAMBIA 1305.1 (as control) or pDAT 2.3+dat. Up to seven plates, with each plate containing 15 leaves, were inoculated. Once emerged, hairy roots were selected on MS medium containing NAA (0.1 mg/L), BAP (2 mg/L) and hygromycin (50 mg/L). Only hairy roots resistant to antibiotic for at least 3 months (3–4 subcultures of selection on hygromycin containing media) were used for subsequent applications.

Histochemical GUS assay

Explants of agroinfiltrated C. roseus, shoots of stably transformed tobacco and hairy roots of C. roseus were used for GUS activity measurements, according to Jefferson (1987). Plants were immersed in staining solution: Na-phosphate buffer (50 mM, pH 7.0), 0.5 mM potassium ferricyanide, 0.5 mM potassium ferrocyanide, 1 % Triton X-100, 1 mg/mL 5-bromo-4-chloro-3-indolylglucuronide, and incubated at 37°C in the dark for 1 day. Later, chlorophyll was removed by immersion of the plant material in 70% ethanol. Stained plant material was viewed by light microscope (Olympus BX51) and photographed with a DP50 camera.

Results

In silico analyses of the DAT gene promoter

For our initial in silico analysis, we examined the sequence spanning ca. 2.4 kb upstream of the translational start site of the DAT gene, using publically available databases including PlantCARE (Lescot et al. 2002) and PLACE (Higo et al. 1999). Subsequent analysis of putative regulatory elements within the coding sequence of the DAT gene was accomplished using PLACE, PlantCARE Transfac (Wingender et al. 2000), softberry (Shahmuradov et al. 2003) and MIRAGE (Ghosh 2000). Such databases allow the prediction of spatial and temporal acting elements and their putative role in response to biotic and abiotic stresses.

For the 5′ upstream region, within the first ca. 1.8 kb, we found many of the motifs described by Wang et al. (2010), and extend this analysis by an additional 600 bp. From the PlantCARE and PLACE databases, common basal promoter elements such as the TATA-box and CAAT-box were readily identified within the 2.4 kb 5′ UTR region analyzed (Fig. 2). In addition, many regulatory motifs associated with plant responses to biotic and abiotic stresses, as well as regulators of spatial specific expression, and expression enhancers were identified. For example, numerous light (G-box, I-box, AE-box, AT1-motif, GT1-motif, ATC-motif, ATCC-motif, ATCT-motif, MRE, LAMP-element, chs-CMA1a, chs-CMA2a, TCT-motif) (Arguello-Astorga and Herrera-Estrella 1998), abscissic acid (ABRE, TCE), MJ (CGTCA, TGACG), and ethylene (ERE) motifs, as well as motifs associated with response to fungal elicitors (Box-W1, ELI-box3), gibberellins (P-box, TATC-box) and other elements involved in response to stress (e.g., temperature, wounding and to drought) were identified (Fig. 2).

Fig. 2
figure 2

In silico analysis of the 5′ upstream region (ca. 2.4 kb) of the DAT gene. Regulatory motifs (shaded) were identified using PlantCARE (Lescot et al. 2002) and PLACE (Higo et al. 1999). The transcriptional start site (+1) is based on Wang et al. (2010)

Full size image

Analysis of the coding region of DAT using PLACE, Plant CARE, Transfac, Softberry and MIRAGE identified putative motifs involved in stomata specific expression, such as DOFs, in a few positions (Table 1). Other motifs such as WRKY, C1, SEFs and (TGA1a and TGA1b) (Katagiri et al. 1989) were identified in the DAT coding region.

Table 1 Regulatory elements and transcription factor binding sites in the coding region of the Catharanthus roseus DAT gene
Full size table

Analysis of DAT gene promoter-driven GUS expression

All five DAT gene constructs, including the control (p35S), pDAT 812 (in which the 35S promoter driving GUS Plus has been replaced with an 812 bp fragment of the DAT promoter), pDAT 812 Δ35S (same as pDAT 812 except the 35S::Hygromycin component of the pCAMBIA1305.1 vector has been deleted), pDAT 2.3 (in which the 35S promoter driving GUS Plus has been replaced with an 2,431 bp fragment of the DAT promoter) and pDAT 2.3+dat (in which the 35S promoter driving GUS Plus has been replaced with an 2,431 bp fragment of the DAT promoter plus the entire DAT ORF) (Fig. 1a–e) were verified by re-isolation from E. coli, digestion with appropriate restriction enzymes and gel analysis (data not shown). Isolated fragments were also sequenced to verify their identity (data not shown).

Transient expression in seedlings of C. roseus after agroinfiltration

Agroinfiltration was used to transiently express the different pDAT::GUS constructs in C. roseus in order to localize specific DAT expression (using GUS enzyme activity as a proxy for gene expression) and potentially identify cis-acting regulatory sequences. Agroinfiltration of whole seedlings with the p35S GUS Plus construct revealed strong GUS activity levels throughout the plantlet (Fig. 3a). By contrast, whole plantlets agroinfiltrated with pDAT 812 GUS Plus (Fig. 3b), pDAT 812 Δ35S GUS Plus (data not shown) and pDAT 2.3 GUS Plus (Fig. 3c) constructs showed weaker and more organ specific GUS activity. That is, GUS accumulated in cotyledons and the hypocotyl region (Fig. 3b, c), but not in roots. Inclusion of the DAT ORF in GUS constructs (pDAT2.3+dat GUS Plus) greatly diminished overall GUS accumulation throughout agroinfiltrated plantlets (Fig. 3d). In contrast to stably transformed tobacco leaves (see below), higher magnification of leaves of C. roseus transiently expressing pDAT2.3+dat GUS Plus constructs did not reveal any apparent tissue specificity (Fig. 3e); indeed, GUS activity was effectively non-detectable in these tissues.

Fig. 3
figure 3

β-Glucuronidase (GUS) activity in Catharanthus roseus and Nicotiana tabacum: a–e Agroinfiltrated seedlings of Catharanthus roseus harbouring different pDAT constructs: a p35S, b pDAT 812, c pDAT 2.3, d, e pDAT 2.3+dat. f–i Leaves of Nicotiana tabacum stably transformed with f p35S, g pDAT 812, h pDAT 2.3, i pDAT 2.3+dat. j, k Hairy roots of Catharanthus roseus harbouring different pDAT constructs: j p35S, k pDAT 2.3+dat. Arrows in (e) and (i, inset) point to guard cells. f–k ×100, e, i (inset), ×200

Full size image

Expression in stably transformed Nicotiana tabacum

Due to the difficulty of obtaining stably transformed C. roseus whole plants, a tobacco heterologous expression system was used for pDAT GUS promoter construct analysis. Tobacco leaves from plants transformed with p35S GUS Plus showed strong GUS activity (Fig. 3f). Similarly, transformation with the pDAT812 GUS Plus (Fig. 3g) and pDAT2.3 GUS Plus (Fig. 3h) constructs resulted in GUS activity in most cell types throughout leaves, albeit with less intensity than observed for the p35S GUS Plus construct. By contrast, transformation with the pDAT2.3+dat GUS Plus construct resulted in strong, specific GUS activity in stomatal guard cells and vascular tissues cell types (Fig. 3i).

Expression in C. roseus hairy roots

The construct pDAT2.3+dat GUS Plus and the positive control p35S GUS Plus were transformed into Agrobacterium rhizogenes, strain15834. Although DAT expression was restricted to specific cells types found in leaves and young buds (St-Pierre et al. 1999), it has also been expressed in C. roseus hairy roots under the regulation of the 35S promoter (Magnotta et al. 2007). Herein, we succeeded in expressing the GUS reporter gene under the control of the DAT promoter, in a C. roseus hairy root system. First, we expressed GUS in C. roseus hairy roots derived from A. rhizogenes transformed with our p35S::GUS construct, and found relatively strong GUS activity in all tissues, but especially the vascular tissue and meristem regions (Fig. 3j). By contrast, expression of the reporter gene (as revealed by GUS activity) in C. roseus hairy roots derived from A. rhizogenes transformed with our pDAT 2.3+dat GUS Plus construct was restricted to the vasculature (Fig. 3k). As with the transient expression in C. roseus plantlets and stable expression in N. tabacum, GUS activity was weak in the tissue transformed with the pDAT 2.3+dat GUS Plus construct.

Discussion

The expression of genes is normally thought to be regulated by the untranslated sequence upstream of the ORF; however, gene expression can also be affected by numerous other elements, which can be located in the upstream (i.e., promoter) region, coding region (including intron and exon elements), as well as in the 5′UTR and 3′UTR (see, for example, Curie and McCormick 1997; Buzeli et al. 2002; Ito et al. 2003; Menossi et al. 2003; Fiume et al. 2004; Gowik et al. 2004; Reddy and Reddy 2004; Ng et al. 2006). The DAT gene has been shown experimentally to be responsive to light and MJ (St-Pierre et al. 1998; Hernandez-Dominguez et al. 2004; St-Pierre and De Luca 1995; St-Pierre et al. 1998; Vazquez-Flota and De Luca 1998). Consistent with this, Wang et al. (2010) characterized three TGACG motifs involved in MJ signaling in the first 1.8 kb of the DAT promoter, while our in silico analysis of the DAT promoter region also identified numerous motifs known to be light responsive elements. We also found many other motifs known to be involved in the acclimation and/or response to biotic and abiotic stresses. These factors participate in the regulation of many processes such as the synthesis of storage proteins, regulation of the light response of carbohydrate metabolism genes, several defense mechanisms and gibberellin and auxin responses (Yanagisawa and Schmidt 1999; Plesch et al. 2001). However, none of these imparted any specific tissue expression in either transient or stable gene transformation systems. Indeed, it was not until the DAT ORF was incorporated into the reporter gene construct that tissue specificity (i.e., guard cell specific expression) was observed. At this stage, it is difficult to pinpoint exactly which elements within the ORF are responsible for the explicit guard cell-specific expression observed, but DOF (DNA binding with one finger) elements at 457, 745, 1007, 1010, 1048 and 1123 bp within the ORF are likely involved.

The DAT construct pDAT 2.3+dat GUS Plus imparted stomatal guard cell specific expression of GUS (viewed by proxy through GUS enzyme activity) in stably transformed tobacco leaves and predominantly vascular expression in stably transformed C. roseus hairy roots. By contrast, little tissue specificity of DAT expression was observed in constructs that did not contain the DAT ORF. The DAT gene is normally expressed in idioblast cells in C. roseus (St-Pierre et al. 1998, 1999), however, neither tobacco leaves, nor C. rosues roots contain idioblasts. Nevertheless, a similar stomatal guard cell specific expression was reported for other promoters associated with idioblast-specific gene expression [e.g., TGG1 (2.5 kb) and Brassica napus Mir1 Bn1 (2.9 kb)] expressed in Arabidopsis thaliana (Thangstad et al. (2004)). We speculate that the stomatal expression of DAT was governed in part by elements in the DAT ORF. Specific gene expression driven by elements within the ORF of the corresponding gene in other related cell types have been described for A. thaliana. For example, the OASA1 promoter drives reporter gene expression in glandular trichomes in Arabidopsis, provided domains between -266 and -66 and +112 and +375 are present. When expressed in a heterologous system (e.g., tobacco, peppermint) expression remained restricted to non-glandular trichomes as long as the ORF domain was present (Gutierrez-Alcala et al. 2005).

In previous work, the DAT coding region was expressed in a hairy root system, albeit under the control of a constitutive (35S) promoter (Magnotta et al. 2007). In the present work, we demonstrated the expression of the DAT coding region (through GUS enzyme activity) under the control of its own promoter in both C. roseus hairy roots and tobacco leaves. From this, it was clear that the coding region of the DAT gene, in conjunction with elements in the 5′ upstream region, play a role in its cell/tissue specific expression. However, in order to fully elucidate which elements or motifs (e.g., cis acting elements and their trans-acting factors) are responsible for the spatiotemporal expression of the DAT gene under different environmental and physiological conditions, additional, shorter constructs, derived from the pCAMBIA pDAT 2 3+dat construct must be created and tested. Additionally, the role of DOFs located within the DAT coding region need to be elucidated by gel shift analysis, footprinting and chromatin immunoprecipitation.