Interplay of Sugar, Light and Gibberellins in Expression of Rosa hybrida Vacuolar Invertase 1 Regulation

(1)

Interplay of Sugar, Light and Gibberellins in Expression of Rosa hybrida Vacuolar Invertase 1 Regulation

Ame´lie Rabot

^1,5

, Virginie Portemer

^2,5,6

, Thomas Pe´ron

¹

, Eric Mortreau

¹

, Nathalie Leduc

³

, Latifa Hamama

^1,3,4

, Pierre Coutos-The´venot

²

, Rossitza Atanassova

²

, Soulaiman Sakr

¹

and Jose´ Le Gourrierec

^3,

*

1Agrocampus-Ouest, Institut de Recherche en Horticulture et Semences (INRA, Agrocampus-Ouest, Universite´ d’Angers), SFR 149 QUASAV, F-49045 Angers, France

2Universite´ de Poitiers, UMR 7267 CNRS/Universite´ de Poitiers E´cologie et Biologie des Interactions, e´quipe Physiologie Mole´culaire du Transport des Sucres chez les ve´ge´taux, 3 rue Jacques Fort, B31, 86 000 Poitiers, France

3Universite´ d’Angers, Institut de Recherche en Horticulture et Semences (INRA, Agrocampus-Ouest, Universite´ d’Angers), SFR 149 QUASAV, F-49045 Angers, France

4INRA, Institut de Recherche en Horticulture et Semences (INRA, Agrocampus-Ouest, Universite´ d’Angers), SFR 149 QUASAV, F-49071 Beaucouze´, France

5These authors contributed equally to this work.

6Present address: INRA, Institut Jean Pierre Bourgin, UMR 1318, F-78026 Versailles, France.

*Corresponding author: E-mail, jose.gentilhomme@univ-angers.fr; Fax,+33-2-41-735352.

(Received April 18, 2014; Accepted August 2, 2014)

Our previous findings showed that the expression of theRosa hybridavacuolar invertase 1 gene (RhVI1) was tightly correlated with the ability of buds to grow out and was under sugar, gibberellin and light control. Here, we aimed to pro- vide an insight into the mechanistic basis of this regulation. In situ hybridization showed thatRhVI1expression was localized in epidermal cells of young leaves of bursting buds.

We then isolated a 895 bp fragment of the promoter of RhVI1. In silico analysis identified putative cis-elements involved in the response to sugars, light and gibberellins on its proximal part (595 bp). To carry out functional analysis of theRhVI1promoter in a homologous system, we developed a direct method for stable transformation of rose cells. 5⁰deletions of the proximal promoter fused to theuidAreporter gene were inserted into the rose cell genome to study the cell’s response to exogenous and endogenous stimuli.

Deletion analysis revealed that the 468 bp promoter fragment is sufficient to trigger reporter gene activity in response to light, sugars and gibberellins. This region confers sucrose- and fructose-, but not glucose-, responsive activation in the dark. Inversely, the –595 to –468 bp region that carries the sugar-repressive element (SRE) is required to down-regulate theRhVI1promoter in response to sucrose and fructose in the dark. We also demonstrate that sugar/light and gibberellin/light act synergistically to up-regulateb-glucuronidase (GUS) activity sharply under the control of the 595 bp pRhVI1region.These results reveal that the 127 bp promoter fragment located between –595 and –468 bp is critical for light and sugar and light and gibberellins to act synergistically.

Keywords: Gibberellins Light Promoter Rosa hybrida Sugar Vacuolar invertase.

Abbreviations: GARE, gibberellin-reponsive element; GUS, b-glucuronidase; LRE, light-responsive element; MS, Murashige

and Skoog; MU, methylumbelliferone; RhVI1,Rosa hybridavacuolar invertase 1; SRE, sugar-repressive element.

Introduction

Invertases represent a family of enzymes that are classified into two main groups, alkaline/neutral invertases and acid invertases, that display different biochemical properties and subcel- lular localizations. Alkaline/neutral invertases have an optimal pH within a 6.5–8.0 range (Roitsch and Gonza´lez 2004) and are involved in carbon partitioning, plant development and stress responses (Gallagher and Pollock 1998,Qi et al. 2007,Jia et al.

2008,Barratt et al. 2009,Welham et al. 2009,Yao et al. 2009, Xiang et al. 2011). They are located in the cytoplasm and in various organelles (Lou et al. 2007,Murayama and Handa 2007, Szarka et al. 2008,Vargas et al. 2008,Vargas and Salerno 2010, Martin et al. 2013). Acid invertases are localized in the vacuole or in the apoplast and have an optimum pH of 3.5–5.5. Cell wall invertases play a key role in the control of sink activity (Miller and Chourey 1992,Sturm 1999,Tang et al. 1999,Goetz et al.

2001) and developmental processes (Borisjuk et al. 2004,Heyer et al. 2004). Vacuolar invertases are encoded by a small multi- gene family composed of two members inArabidopsis thaliana (Haouazine-Takvorian et al. 1997),Gossypium hirsitum(Wang and Ruan 2010), potato (Kumari and Das 2013) andRosasp.

(Rabot et al. 2012). High expression or activity of vacuolar invertases has mainly been reported in elongating tissues including grape berry (Davies and Robinson 1996), carrot taproot (Tang et al. 1999) and sugar beet petioles (Gonzalez et al.

2005), emphasizing their evident role in plant cell expansion (Gonzalez et al. 2005,Sergeeva et al. 2006). The function of vacuolar invertase in the establishment of sink strength has also been proposed on the basis of computational modeling (Nage¨le et al. 2010). While the expression level of vacuolar

Plant Cell Physiol.0(0): 1–15 doi:10.1093/pcp/pcu106, Advance Access publication on 9 August 2014, available online at www.pcp.oxfordjournals.org

!The Author 2014. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists.

Regular Paper

at INRA Institut National de la Recherche Agronomique on September 25, 2014http://pcp.oxfordjournals.org/Downloaded from

(2)

invertase is under the control of endogenous and exogenous stimuli (Dorion et al. 1996,Xu et al. 1996,Greiner et al. 1999, Kim et al. 2000,Andersen et al. 2002,Gonza`lez and Cejudo 2007, Proels and Roitsch 2009), nothing is known regarding the mechanistic basis behind that control. For example, water-stressed maize plants exhibited increased acid invertase activity and hexose (glucose and fructose) accumulation, linked to the up-regulation ofIvr2, a vacuolar invertase gene, in vegetative sink and source organs (Kim et al. 2000). Providing ABA to the root medium of hydroponically grown maize plantlets stimu- lated the activity of vacuolar invertases and increased the transcript levels ofIvr2in leaves and roots (Trouverie et al. 2004).

Invertase genes (Ivr1and2) are sugar modulated and fall into two classes with contrasting responses (Xu et al. 1996).Ivr2is up-regulated by glucose or sucrose supply, whileIvr1is down- regulated under the same experimental conditions. Despite these findings, many questions still remain unanswered: for example, nothing is known about the mode of action of different stimuli—synergistic/agonistic/antagonistic—on the expression of genes encoding sucrose-cleaving enzymes, or on the nature of thecis-elements putatively involved in their regulation. Present knowledge on the transcriptional regulation of invertase genes appears very scarce compared with the exten- sive studies on the fine-tuning mechanisms implied in the signaling pathways of different endogenous and exogenous stimuli and their target genes (Mishra et al. 2009, Das et al. 2012, Kushwah and Laxmi 2014). Moreover, even if the regulation of invertase expression has been reported in many physiological contexts, nothing is known within the context of bud outgrowth, a key process in plant branching.

As sessile organisms, plants have evolved a diversity of mechanisms to perceive and to integrate endogenous (hormones and nutrients) and exogenous (light, temperature) signals in their growth and development program (Alabadi and Blazquez 2009). Light is sensed by different photoreceptors that detect different light wavelengths (Eckardt 2003,Chen et al.

2004,Azari et al. 2010). Downstream from the photoreceptors, positive and negative regulation of light-responsive transcription factors occurs (Jiao et al. 2007,Leivar and Monte 2014).

These light-responsive transcription factors were identified through genetic analyses of mutants deficient in their response to light and through screens for light-responsivecis-element- binding proteins. The GATA-box, G-box, GT-1-box and I-box are important in light-regulated gene expression (Donald and Cashmore 1990, Teackle et al. 2002, Jeong and Shih 2003, Choudhury et al. 2008). Plant hormones are also perceived and initiate a transduction signal which is a powerful driver for the regulation of plant growth and development (Nemhauser 2008). Their effect can occur through transcriptional regulation of gene expression; a huge number of promoter motifs have already been identified as hormone responsive (Abe et al. 2003,Chen et al. 2006). Many investigations show that the effect of hormones can be affected by light, supporting a cross-talk between light and hormone signaling pathways (Alabadi and Blazquez 2009, Halliday et al. 2009, Hornitschek et al. 2012, De Lucas and Prat 2014). Light and auxin interact cooperatively to control adventitious root

formation (Sorin et al. 2006), while light and gibberellin antag- onize each other to control hypocotyl elongation (Alabadi and Blazquez 2009,Claeys et al. 2014). Sugar availability also affects plant growth, as sugars serve both as carbon and energy sources for cell metabolism and as effective signaling molecules (Smeekens et al. 2010, Bihmidine et al. 2013, Smeekens and Hellman 2014). Sugars can regulate the expression level of a variety of genes, whose promoters contain many sugar-responsivecis-elements (Ishiguro and Nakamura 1992,Grierson et al.

1994,Cakir et al. 2003,Sun et al. 2003,Morikami et al. 2005).

Different sugar signaling pathways have been described (Hanson and Smeekens 2009, Li et al. 2011) and show that the expression of certain genes is under the control of both sugar and light (Joeng et al. 2010). A reciprocal relationship between circadian clock and sugar signalling was also recently demonstrated (Bolouri Moghaddam and Van den Ende 2013).

The core central oscillatorCCA1,TOC1andGIgenes are stimu- lated by sucrose (Knight et al. 2008,Dalchau et al. 2011), and both invertase enzyme activity and gene expression are under the control of the circadian clock (Gonzalez et al. 2005). On the other hand, it was recently reported that soluble sugars increase the concentration of reactive oxygen species (ROS) scavengers (Bolouri-Moghaddam et al. 2010,Keunen et al. 2013).

Among sucrose-cleaving enzymes, a vacuolar invertase (RhVI1) is positively correlated with the potential outgrowth of buds under light in rose buds (Girault et al. 2010,Rabot et al.

2012). This is not the case forRhVI2(Rabot et al. 2012), or for cell wall invertase (Girault et al. 2010); this can reveal great flexibility in the control of source–sink relationships in plants (Tymowska-Lalanne and Kreis 1998). Pharmacological experiments based on the exogenous application of sucrose or its non-metabolizable analog (palatinose) to in vitro cultured buds showed that light and sucrose act synergistically to induce bothRhVI1expression and total vacuolar invertase activity (Rabot et al. 2012). Similarly, the effect of light was abolished following the addition of a gibberellin biosynthesis inhibitor to the growth medium (Choubane et al. 2012); stimulation ofRhVI1under light therefore requires gibberellin neo- biosynthesis in buds. These findings reveal thatRhVI1expres- sion is under the synergistic effect of light and sugar, and also of light and gibberellin, and offer an original framework to address the molecular mechanism behind this regulation. They make it a potential candidate for the integration of light, sucrose and gibberellin signaling during the early phase of bud outgrowth.

As a first step towards deciphering the molecular mechanism that controls RhVI1expression in bud outgrowth, we cloned the 895 bp promoter ofRhVI1 and carried out its functional analysis in a homologous system consisting of rose cell suspension cultures. We developed a direct procedure to achieve stable transformation of rose cells, usingAgrobacterium tumefaciens to avoid somatic embryogenesis which remains very time-consuming for many woody plants. Our results demonstrate that the 468 bpRhVI1promoter fragment is sufficient to confer up-regulation in response to sugars, light and gibberellins independently of each other. Finally, we propose the –595 bp to –468 bp region as being required for light and sugar and light and gibberellin to act synergistically.

(3)

Results

RhVI1 exhibits restricted expression in epidermal cells during bud burst under light

Our early data indicated thatRhVI1 expression was up-regulated in bursting buds (Girault et al. 2010,Rabot et al. 2012). To localize it more precisely within bud tissues, we conducted in situ hybridization experiments on bursting buds. Such an ap- proach has previously only been used onZea maysroots (Kim et al. 2000) and cotton fibers (Wang et al. 2010). We first identified different bud tissues (Fig. 1A), and then we hybri- dized longitudinal sections of bursting buds with specific antisense probes in situ (positive hybridization,Fig. 1C, D).RhVI1 transcripts were found in the epidermal cells of young leaves and of primordial leaves, in a non-continuous manner (Fig. 1C, D). Under the same experimental conditions, no staining was found after hybridization with the sense probe as a negative control (Fig. 1B), supporting a tissue-specific localization ofRhVI1.

Identification of cis-elements involved in sugar, gibberellin and light signaling in the RhVI1 promoter

We previously isolated two cDNA sequences encoding two different isoforms of vacuolar invertase (RhVI1 and RhVI2) from rose buds and characterized them regarding the regulation of their expression by light, sugars and gibberellins (Choubane et al. 2012,Rabot et al. 2012). UnlikeRhVI2,RhVI1 responds to sugar and light (Rabot et al. 2012), that is why we focused the present study on the transcriptional control of RhVI1. To elucidate further the regulation ofRhVI1, 895 bp of its 5⁰regulatory region upstream of the translation start codon was isolated.

In silico analysis of the 895 bppRhVI1revealed the presence of 39 motifs potentially involved in the response to light, sugars and hormones (Table 1), suggesting that transcription ofRhVI1 might be controlled by various exogenous and endogenous factors. We found a TATA-box and a CAAT-box, which allow for transcription initiation, at –89 bp and –100 bp upstream of the translation start codon, respectively. Among these 39 motifs, we found two types of cis-elements related to sugar regulation. The first one is a sugar-repressive element (SRE; TTATCCA; Tatematsu et al. 2005) responsive to both sucrose and glucose starvation (Toyofuku et al. 1998,Wang et al. 2007). It is localized at –505 bp, and corresponds to an SRE initially characterized as a binding site for three MYB proteins involved in sugar and hormonal regulation ofa-amylase in rice (Fig. 2, highlighted in gray) (Lu et al. 2002).

The second one, a WBOXHVISO1 motif characterized by the TGACT sequence (Fig. 2, highlighted in red), has been previously demonstrated to be the binding site of a SUgar SIgnaling in Barley 2 (SUSIBA2) transcription factor responsible for a positive response of isoamylase to sucrose during barley seed development (Sun et al. 2003,Sun et al. 2005). WBOXHVISO1 is present in two copies, localized at –315 and –156 bp upstream of the first codon. We identified 18 motifs involved in

transcriptional regulation by hormones along the 895 bp promoter sequence of thepRhVI1gene (Table 1). Among them, 17 were putatively involved in the gibberellin signaling pathway and distributed as follows: four copies of a gibberellin-responsive element (GARE) cis-element (C/TAACC/GG/AA/CC/A) known to bind a GAMyb transcription factor (Woodger et al.

2003), located at –859, –524, –408 and –111 bp, respectively (Fig. 2, dark box), nine motifs carrying the GTCA/TGAC nu- cleotide sequence, known to bind the rice WRKY71 transcriptional repressor of the gibberellin signaling pathway (Zhang et al. 2004), located along thepRhVI1promoter (–698, –640, –559, –521, –336, –311, –185, –152 and –108 bp;Fig. 2, underlined) and four pyrimidine boxes (C/TCTTTT) known to bind a DOF transcription factor (Isabel-La Moneda et al. 2003) located at –734, –576, –255 and –128 bp, respectively (Fig. 2, yellow highlighted). Finally, an ABA-responsivecis-regulatory element (called MYB2consensusAT), located at –319 bp, corresponds to a MYB recognition site found in the promoter of the dehydra- tion-responsive generd22in Arabidopsis (Abe et al. 2003) and required for ABA induction. In addition,pRhVI1carried numer- ous light-responsivecis-acting elements (LREs) (the GT-1-box, Fig. 1 In situ hybridization usingRhVI1probes on bursting buds. (A–D) Paraffin-embedded longitudinal sections of buds 72 h after stem severing and culture under white light (200 mmol m²s¹, 16 h/8 h).

(A) Toluidine blue O-stained section showing the different bud parts.

(B) Section showing the absence of in situ hybridization signal obtained with theRhVI1sense probe. (C and D) Sections showing the in situ hybridization signal with theRhVI1antisense probe. Arrows point to the hybridization signal. Abbreviations: Am, axillary meristem; Lp, leaf prim- ordium; Yl, young leaf. Scale bars = 50mm.

(4)

GATA-box, I-box and E-box) (Borello et al. 1993,Zhou 1999, Toledo-Ortiz et al. 2003). Among them, five of the seven GT-1- boxes we identified were grouped in the distal part of the promoter (–603, –577, –557, –500, –493, –180 and –128 bp), and so were the two GATA-boxes located at –582 and –500 bp upstream of the start codon and the I-box that overlapped with the GATA-box at position –500 bp. We further performed the functional analysis of the proximal part (595 bp) of the RhVI1promoter since it includes all SREs and the most light- and gibberellin-responsive elements. These findings showed that most interesting cis-regulatory elements (i.e. related to sugars, light or hormonal signaling) were primarily found in the 595 bp proximal region.

Rosa hybrida cell culture and its direct genetic transformation

To gain further insight into the physiological significance of the cis-acting elements necessary for sugar, light and gibberellin regulation ofRhVI1expression, an R. hybridacell culture was initiated and the stabilized culture was transformed with dif- ferentRhVI1595 bp proximal promoter deletions. To avoid the limited efficiency and the hazardous character of transient

expression in cells and the difficulties associated with their co-transformation with a second reporter gene as an internal control, we developed a method for direct stable transformation of rose cells in suspension usingA. tumefaciens. For that purpose, a rose cell suspension was initiated fromR. hybridacv.

‘Orange’ leaves and petioles by inducing cell dedifferentiation and callogenesis. The friable calli were added to a liquid culture medium under shaking to produce a rose cell suspension. The cultured cells were transformed usingA. tumefaciensstrain EHA 105 carrying binary vectors with the chimeric fusions of the p35S::uidA andpRhVI1::uidA reporter genes. The cell suspensions were acclimated to light or dark conditions.

Whether under dark or light conditions, the rose cell culture grew exponentially from the first day to the fourth day (50–220 mg of dry cells) followed by a stationary phase from the fifth day to the seventh day (Fig. 3A, dotted curves). This hetero- trophic system used the sucrose initially present in the medium as the main carbon source. The sucrose content dropped sharply and the culture medium was completely depleted at day 2 (38 down to 0 mM). Conversely, the concentration of hex- oses, derived from sucrose hydrolysis, increased during the first 2 d and then progressively declined. Glucose decreased more Table 1 Sugar-, hormone- and light-responsive elements found in silico in the 895 bp RhVI1promoter sequence

Element Sequence Function Copies Position

Sugars SRE TTATCCA Alpha-amylase; MYB proteins; gibberellin;

sugar starvation;

1 –505

WBOXHVISO1 TGACT SUSIBA2 bind to W-box element 2 –315; –156

Hormones WRKY71OS GTCA/TGAC Binding site of rice WRKY71, a transcriptional repressor of the gibberellin signaling pathway

9 –698; –640; –559; –521; –336;

–311; –185; –152; –108

GARE C/TAACC/GG/

AA/CC/A

Transcription factor involved in the activation of signal mediated by gibberellin

4 –859; –524; –408; –111 PYRIMIDINE BOX C/TCTTTT Pyrimidine box found in rice (Oryza

sativa) alpha-amylase (RAmy1A) gene;

gibberellin-responsecis-element of GARE and pyrimidine box are partially involved in sugar repression

4 –734; –576; –255; –128

MYB2CONSENSUSAT CAGTTG ABA 1 –319

Light CICADIANELHC GATCTAGTTG/

CAATGCATC/

GATCTAGTTG

Region necessary for circadian expression of tomato (Solanum lycopersicum) Lhc gene

3 –866; –418; –336

I-box GATAA Conserved sequence upstream of light-

regulated genes; conserved sequence upstream of light-regulated genes of both monocots and dicots

1 –500

E-box CANNTG E-box of napAstorage-protein gene of

Brassica napus; this sequence is also known as the RRE (R response element)

5 –879; –416; –401; –330; –317

GT-1 consensus GATAAT/

ATTTCC/

ATTTTC/

TTTTCC/

GAAAAA

Consensus GT-1 binding site in many light-regulated genes, e.g. RBCS from many species, PHYA from oat and rice, spinach RCA and PETA, and bean CHS15

7 –603; –577; –557; –500; –493;

–180; –128

GATA GATA Three GATA box repeats were found in

the promoter of Petunia hybridaChl a/b-binding protein, Cab22 gene;

required for high level, light-regulated and tissue-specific expression

2 –582; –500

(5)

rapidly than fructose, and glucose was totally depleted at day 5, which coincides with the beginning of the stationary phase.

To gain insight into the transcriptional regulation mechanisms ofRhVI1, we stably transformedRosacells in suspension culture with thep595RhVI1,p468RhVI1,p307RhVI1andp148RhVI1pro- moters fused to theuidAreporter gene (Fig. 3B). Ab-glucuronidase (GUS) histochemical assay also revealed that the cells had been successfully transformed (Fig. 3C). We conducted the expression analysis 48 h after cell subculturing, i.e. in the middle of the exponential phase. The cells were then rinsed three times with Murashige and Skoog (MS) medium without sugars. At that point of cell growth, the culture medium was completely sucrose depleted, which is probably due to metabolic activity of the cell culture, and appeared to be the most suitable time for treatment with different effectors (Fig. 3A).

Transcriptional regulation of pRhVI1 is modulated by sucrose and fructose in the absence of light

To assess the sugar responsiveness of pRhVI1, we tested the proximal promoter region and its deletions (Fig. 4A) for their response to different metabolizable sugars [sucrose, glucose and fructose at 58 mM] after 24 h incubation in the dark (Fig. 4). We further analyzed the expression levels conferred by these promoters by quantifying GUS activity (Fig. 4B). We used mannitol (58 mM) as an osmotic agent and considered the resulting reporter gene activity as a control.

In the dark, GUS reporter gene expression conferred by the proximal promoter (595pRhVI1) was low [100 pmol

methylumbelliferone (MU) min¹mg¹protein], and statistic- ally similar for all our metabolizable sugars (sucrose, glucose and fructose) and mannitol. Conversely, the first deletion (468pRhVI1), that caused the loss of the SREcis-repressive element, promoted a significant induction of GUS activity by sugars.

Sucrose and fructose induced the expression of theRhVI1pro- moter by 3- and 2.5-fold, respectively, compared with mannitol (sucrose, 205 pmol MU min¹mg¹protein; fructose, 165 pmol MU min¹mg¹protein). Interestingly, we observed no significant induction of GUS activity when the cells were incubated with glucose (Fig. 4B).

Compared with468pRhVI1, the307pRhVI1and148pRhVI1 truncated promoters reduced reporter gene expression:

GUS activity dropped significantly to 30 and 10 pmol MU min¹mg¹protein, respectively. These results suggest that a WBOXHVISO1-positive sugar cis-acting element may play a crucial role in the control ofpRhVI1transcriptional activity.

Light and sugar affect pRhVI1 expression synergistically

Previous data showed up-regulation of RhVI1 by light in the buds of beheaded plants (Girault et al. 2010) and this effect was enhanced by sugars (Rabot et al. 2012). As the response to sugars was abolished by the 307pRhVI1and 148pRhVI1dele- tions but was maintained by the proximal promoter595pRhVI1 and the 468pRhVI1deletion, we only studied these last two transformants for their response to the combined effects of light and sugars (Fig. 6). First, we analyzed the response of Fig. 2 Sequence of the 895 bp 5⁰upstream region of the promoter sequence of theRosavacuolar invertase 1 (pRhVI1) gene encompassing the maincis-regulatory elements: (i) sugar repression (SRE, highlighted in gray) and activation (WBOXHVISO1, highlighted in red); (ii) gibberellin signaling (WRKY71OS, underlined; GARE, fully outlined; pyrimidine box, yellow highlighted); (iii) ABA signaling (blue); (iv) light signaling (CICADIANELHC, bold and italic; I-box, orange; E-box, dot outlined; GATA-box, dot underlined, GT-1-box, wavelet outlined; and DOF, yellow highlighted); and (v) L1 layercis-regulatory element (L1box, double outlined). The CAAT-box and TATA-box are double underlined.

(6)

pRhVI1to light only by incubating cultured cells with mannitol as the sole carbon source under light. As shown inFig. 5, similar GUS activity was observed for both the 595pRhVI1 and 468pRhVI1 constructs under light, and was 1.5- and 2-fold higher than in the dark, respectively. The light-responsivecis- acting elements (E-box and GT-1-box) present in the 468pRhVI1promoter region seem to be sufficient to confer a similar induction of GUS expression to the level conferred by the proximal 595pRhVI1promoter in response to light. This result may account for high expression ofRhVI1in buds kept under light compared with darkness (Rabot et al. 2012).

Interestingly, the two constructs were significantly induced by metabolizable sugars combined with light (Fig. 6). Under the control of the proximal595pRhVI1promoter, we noted a significant enhancement of GUS expression in the presence of sucrose and light (3-fold, 500 pmol MU min¹mg¹ protein). We observed a similar effect, yet at a lower level, when we combined glucose or fructose with light (1.5- and 1.2-fold, respectively, compared with mannitol). The 468pRhVI1 promoter induced weaker GUS expression than595pRhVI1 (300 pmol MU min-

1mg¹ protein) when sucrose and light were combined.

Altogether these results indicate that sugars, especially sucrose, combined with light are required for the strong up-regulation of pRhVI1expression, and that the –595 to –468 bp region may be

necessary for this synergistic effect. Such a region is marked by the overlapping of one motif for sugar repression (SRE) and different LREs (GATA-box, I-box and GT-1-box) (Fig. 2,Table 1).

pRhVI1 is under the control of the synergistic effects of light and gibberellins

Gibberellins and light modulate bud burst in Rosa sp.

(Choubane et al. 2012). Moreover, the effect of light onRhVI1 expression involves gibberellin biosynthesis in buds (Choubane et al. 2012). In order to investigate the effect of gibberellins on theRhVI1promoter, we incubatedRosacell cultures with mannitol and exogenous GA3under light or dark conditions.

The effect of gibberellins on RhVI1 might involve the WRKY71OS or GAMYB cis-elements carried by the RhVI1 proximal promoter (Table 1). In the dark, we found similar GUS activity levels for the595pRhVI1and468pRhVI1constructs, with 2.0- and 2.8-fold, respectively, higher reporter gene expression in the presence of gibberellin than in the mannitol control (Fig. 7).

Compared witho595pRhVI1, this higher induction (2.8-fold) conferred by 468pRhVI1 might reflect the loss of two WRKYOS71 gibberellin-repressive elements. GUS expression conferred by 595pRhVI1and468pRhVI1was 2-fold higher under light than in the dark, but similar to gibberellin treatment in the dark (Fig. 7).

Fig. 3 Characterization ofRosa hybridacell suspension and validation of its genetic transformation by chimeric constructs of thepRhVI1::GUS reporter gene. (A) Cell growth curves (dotted lines) and sugar concentrations (full lines) in the culture medium under light (yellow lines) or dark (black lines) conditions. The concentrations of different sugars [sucrose (diamonds), glucose (squares) and fructose (triangles)] were measured.

Error bars represent the SEM. (B) PCR control of cell transformation with chimeric constructs of the pRhVI1::GUS reporter gene.

(C) Histochemical GUS staining of transformed cells withp35S::GUS(control) andp595RhVI1::GUS5 d after subculturing.

(7)

However, under combined light and gibberellin conditions, the proximal promoter induced a sharp increase in GUS activity, up to 500 pmol MU min¹mg¹ protein (2.5-fold) compared with gibberellin in the dark (Fig. 7). In contrast, GUS activity induced by 468pRhVI1 increased

weakly (1.3 -fold; 260 pmol MU min¹mg¹protein) compared with the dark conditions (Fig. 7).

Our results indicate that the combination of gibberellin and light strongly up-regulates pRhVI1and that the –595 to –468 bp region might be required for the synergistic effect of Fig. 4 Functional analysis ofRhVI1promoter deletions in stably transformedRosacells in response to sugar treatment. (A) Schematic repre- sentation of chimericpRhVI1::GUS constructs. Putativecis-elements of interest are represented by three symbols (light , gibberellin , sugars ), and the TATA-box ( ) and CAAT-box ( ). Numbers indicate distances from the translation start codon (+1). (B) Response ofpRhVI1to various sugars in the dark. Transformed cells were incubated with mannitol (MAN), sucrose (SUC), glucose (GLC) or fructose (FRU) for 24 h. GUS activity for each construct corresponds to the means of at least three independent experiments. Error bars represent the SEM. Significant differences between values are indicated by a, b, c, d, e and f, while bc or cd indicate no differences between the b and c groups or the c and d groups, respectively. GUS activity is expressed in pmol methylumbelliferone (MU) min¹g¹protein.

(8)

light and gibberellin. This region encompasses both gibberellin- and light-responsive elements: two repeats of WRKYOS71 gibberellin-cis-repressing elements and one GAmyb-cis-activating element, surrounded by four GT-1-boxes and two GATA-boxes (Table 1) which could work together in this response.

Discussion

Soluble sugars as well as gibberellins contribute to the effects of light on bud burst and vacuolar invertase through the regulation ofRhVI1 expression and to its effects on the activity of vacuolar invertases (RhVIs) (Girault et al. 2010,Choubane et al.

2012,Rabot et al. 2012). To decipher the functional mechanism of the regulation ofRhVI1expression, we isolated the 895 bp promoter ofRhVI1. Based on in silico analysis, we investigated the regulation of its 595 bp proximal promoter region that bears most of thecis-elements responsive to light, sugar and gibberellins in stably transformed R. hybrida cell culture.

The proximal promoter region has been reported to be

sufficient to control the expression of Vitis vinifera hexose transporter (VvHT1) in cell culture or to confer organ-specific and light-regulated transcription to the rice rubisco activase gene (Rca) in transgenic Arabidopsis (Atanassova et al. 2003, Yang et al. 2012).

We generated transgenicR. hybridacells and stabilized them within 8 weeks, and used them as a powerful tool for studying the effects of different treatments (light, sugars and hormones) onpRhVI1expression. Stably transformed cell suspensions appeared as a suitable model for promoter functional study, and added further details to our previous analysis on isolated rose buds cultivated in vitro (Rabot et al.2012) (Fig. 3A). The suc- cessful use ofA. tumefaciensfor genetic transformation of different Rosa species has already been reported (Firoozabady et al. 1994, Dohm et al. 2001, Condliffe et al. 2003, Vergne et al. 2010), but ours is the first direct method for stable genetic transformation ofRosacells in culture.

Sugars, gibberellins and light independently regulate the response of RhVI1 in a Rosa cell homologous system

To identify the promoter regions that respond to sugars, gibberellins and light independently, we fused 5⁰-end deletions of the RhVI1proximal promoter to the GUS reporter gene and used the construct to transformRosacells (Fig. 4A). In silico analysis of theRhVI1895 bp promoter revealed the presence of many sugar-, gibberellin- and light-relatedcis-acting elements (Table 1). The two sugar-responsive motifs found in this promoter (one SRE and two WBOXHVISO1) have previously been reported. The SRE was first identified in the promoter of the rice a-amylase geneamy7(Morita et al. 1998) and then in many a-amylase promoters in rice and barley (Yu et al. 1996,Perata et al. 1997). This motif is generally functionally active in one copy (Lu et al. 1998), and was recognized by the OsMYBS1, OsMYBS2 and OSMYBS3 transcription factors, leading to the induction ofa-amylase expression under sucrose and glucose starvation in rice (Morita et al. 1998,Toyofuku et al. 1998,Lu et al. 2002). Nothing is known about induction ofa-amylase Fig. 6 Combined effect of light and sugar onpRhVI1expression. Cells

containing595pRhVI1and468pRhVI1were incubated with mannitol (MAN), sucrose (SUC), glucose (GLC) or fructose (FRU) for 24 h under light. GUS activity for each construct corresponds to the means of at least three independent experiments. GUS activity is expressed in pmol methylumbelliferone (MU) min¹g¹protein. Error bars represent the SEM. Significant differences between values are indicated by a, b and c.

Fig. 7 Combined effect of light and gibberellin onpRhVI1expression.

Stably transformed cells were incubated with 5mM GA3for 24 h under dark and light conditions. GUS activity for each construct corresponds to the means of at least three independent experiments. GUS activity is expressed in pmol methylumbelliferone (MU) min¹g¹protein.

Error bars represent the SEM. Significant differences between values are indicated by a, b, c and d.

Fig. 5 Effect of light onpRhVI1expression. Stably transformed cells were incubated during 24 h under dark and light conditions. GUS activity for each construct corresponds to the means of at least three independent experiments. GUS activity is expressed in pmol methylumbelliferone (MU) min¹g¹ protein. Error bars represent the SEM. Significant differences between values are indicated by a, b and c.

(9)

expression by fructose. The promoter of 20 rice genes, whose expression is significantly controlled by sucrose starvation, contains this SREt (Wang et al. 2007). Moreover, down-regulation of the Arabidopsis dormancy-associated geneAtDRM1by sugar is related to the presence of an SRE on theAtDRM1promoter (Tatematsu et al. 2005). In agreement with this finding, we observed increased activity of theRhVI1promoter in response to sucrose and fructose but not to glucose in the dark after deleting the promoter region located between –595 and –468 bp encompassing the SRE (located at –505 bp). The sugar- repressive role of SRE has been reported for potato vacuolar invertase (StvacINV1) under light (Ou et al. 2013), whose proximal promoter (500 bp) contains seven SREs in comparison with one SRE in the promoter ofRhVI1.

The Wbox (WBOXHVISO1) was first identified in the promoter of the isoamylase geneiso1in barley; it is recognized by SUSIBA2, a transcription factor belonging to the WRKY family, which can also bind to the SUREcis-element (Sun et al. 2003).

This Wbox exists in one copy in the iso1 promoter and is involved in sugar signaling under light (Sun et al. 2003).

However, many promoters contain two, close or distant, Wboxcis-elements that bind WRKY transcription factors as homo- or heterodimers (Mare` et al. 2004, Ciolkowski et al.

2008). Similarly, two copies of the WBOXHVISO1 sugar-responsive motif exist on the proximal promoter ofRhVI1(at –315 and –156 bp of a 468 bp fragment) (Table 1) and might be required for induction by sucrose and fructose in the dark (Fig. 4B). SRE and WBOXHVISO1 were also recently identified in the promoter of the bamboo vacuolar invertase gene (Bofruct3). However, nothing is known about their functional activity in response to sugars (Liao et al. 2013). These findings confirm that sugars controlRhVI1expression independently of light, and may at least partially explain the ability of sucrose, and to a lesser extent fructose, to induceRhVI1expression in Rosavegetative buds under darkness (Rabot et al. 2012).

Gibberellins regulate many phases of plant development such as seed germination or flowering transition and then control the expression of many functionally different genes including vacuolar and cell wall invertase genes during seed germination (Mitsuhashi et al. 2004), a-amylase genes in barley embryos (Morita et al. 1998), the sucrose synthase gene in cotton fibers (Bai et al. 2014) and the vacuolar invertase geneRhVI1inRosasp. (Rabot et al. 2012). The proximal promoter ofRhVI1(595pRhVI1) contains two different types ofcis- acting elements potentially involved in the response to gibberellins, i.e. three GAREs and seven WRKY71OSs (Table 1). The GARE is targeted by the R2R3-type positive transcription factor GAMYB in barley, and up-regulatesa-amylase independently of sugars in barley (Gubler et al. 1995) and rice (Chen et al. 2006).

This GARE motif is also responsible for the gibberellin-mediated up-regulation of parietal invertase lin5 expression in tomato (Proels et al. 2003). More recently, eight and one GARE cis- elements were identified in the promoters of the vacuolar invertase genes StvacINV1 and Bofruct3, respectively, and StvacINV1was the only one whose expression was proposed to be up-regulated by gibberellin through the GARE motif under light (Liao et al. 2013,Ou et al. 2013). In our conditions,

the GARE motif could potentially be involved in the gibberellin- mediated stimulation ofRhVI1expression in the dark (Fig. 7), suggesting that such a gibberellin signaling pathway could be independent of light conditions. Besides the GARE motif, the promoter ofRhVI1contains seven WRKY71OScis-elements in the –595 to –468 bp region (Table 1), that were first identified in the promoter ofAmy32bin rice and are recognized by the transcriptional WRKY repressor (Zhang et al. 2004). The gibberellin-triggered increase in468pRhVI1activity (2.8-fold) was higher than for595pRhVI1(2-fold) that bears two copies of the WRKY71OS repressive cis-element. Therefore this gibberellin signaling pathway could also be involved in the regulation of RhVI1expression.

In silico analysis of the RhVI1 proximal promoter (595pRhVI1) also revealed the presence of 13cis-acting elements potentially involved in the response to light [GT-1-box (5), E-box (3), GATA-box (2), I-box (1) and circadian LHC (2)]

(Sarokin and Chua 1992, Lu et al. 2002, Sun et al. 2003) (Table 1,Fig. 2). These motifs were identified in the promoters of many light-responsive genes (Lam and Chua 1990,Borello et al. 1993,Teakle et al. 2002), but only scarce data are available for genes encoding sugar metabolism enzymes. The GATA-box LRE has functional significance in the regulation of the promoter of banana SPS (sucrose phosphate synthase) under light (Choudhury et al. 2008). More recently, it was found that light-induced expression of Bofruct3 could involve seven LREs including SORLIP-1 (2), SORLIP-2 (1), SORLREP-3 (2), TATTCT (1) and GT-1 (1) (Liao et al. 2013), and only GT-1 is shared with theRhVI1promoter. This suggests that all thesecis- elements are functional and that they might work together.

Reporter gene activity conferred by the RhVI1 promoter increased by 2-fold under light compared with darkness (Fig. 5) in mannitol-treated cells whatever the construct used (595 and 468 bp). These results suggest that the GT-1-box and E-box found in the 468 bp region are sufficient for light responsiveness, independently of sugar and gibberellin.

The synergistic effects of light and sugar and light and gibberellin are triggered through cis-acting elements localized in the RhVI1 proximal promoter

The synergistic effect of light and sugars on RhVI1promoter activity (Fig. 6) was obvious for the proximal promoter, which bears an SRE (TTATCCA, position –505 bp) between –595 and –468 bp, and a few copies of two different light-responsive motifs, i.e. GT-1 (–577, –557, –500 and –493 bp) and GATA- boxes (–582 and –500 bp). One possible explanation is that light may remove the repression effect related to the sole SRE-mediated sugar, either due to competition of the two transcription factors for the overlapping cis-elements in pRhVI1 (Lodish et al. 2000) or by steric interference with other transcription factors binding to their own consensus sequences (Lee et al. 2003). In Arabidopsis, the interaction of the ABI4 transcription factor with the ABI4 binding site could prevent the binding of G-box binding factor to the closely associated G-box motif present in theRBCSpromoter (Acevedo-Hernandez et al.

(10)

2005), leading to the down-regulation ofRBCSunder light. The synergistic effect of light and sugar (i.e sucrose) on pRhVI1- conferred transcription is in accordance with their impact on RhVI1expression in vegetative buds (Rabot et al. 2012), and with their promoting action on cell elongation in growing organs (Wu et al. 1993, Weber et al. 1996, Stewart et al.

2011). Such a synergistic effect of light and sucrose might be achieved through PIF transcription factors, which are G-box- (E- box) binding proteins (Ni et al. 1998), that act as integrators in light- and sugar-mediated responses (Huq et al. 2004,Kumar et al. 2004,Stewart et al. 2011).

Similarly to light and sucrose, light and gibberellins strongly up-regulated595pRhVI1(Fig. 7), in agreement with our previous results inRosasp. (Choubane et al. 2012). The same synergistic effect of light and gibberellins is well documented for seed germination (Yamaguchi and Kamiya 2002) and differs from that reported during hypocotyl elongation (Peng and Harberd 1997). On the other hand, functional analysis of the a-amylase promoter in barley highlights the existence of a gibberellin-responsive regulatory complex named the GARC domain (Isabel-La Moneda et al. 2003) made of three sequence motifs: a GAREt, the pyrimidine box C/TCTTTT, and an SRE (Skriver et al. 1991, Gubler and Jacobsen 1992, Rogers et al.

1994) working in concert. Similar regulation cannot be ruled out for ourRhVI1promoter, due to the presence of one pyrimidine box (C/TCTTTT, at –576 bp) that may play a role in RhVI1 regulation by gibberellin together with the GAREcis- element (at –524 bp) and the SRE (at –505 bp). It is well established that the light–gibberellin interaction relies on a nuclear protein cascade involving light, PIF and gibberellin signaling factors such as DELLA (a gibberellin signaling repressor) and GID1 (a gibberellin receptor) (De Lucas et al. 2008,Nunez-Flores et al. 2010,Leivar and Quail 2011). Interestingly, a connection beween DELLA and sucrose has recently been established, making DELLA a new key component in sucrose signaling (Li et al. 2014) and placing DELLA at the center of a network connection of light, sugar and gibberellin.

Our results support the finding that the synergistic effects of light and gibberellin on the one hand, and light and sugar on the other hand, onpRhVI1expression require a 127 bp region located between –595 and –468 bp (Fig. 8). While vacuolar invertase is subjected to many types of regulation including post-translational regulation (McKenzie et al. 2013, Tauzin et al. 2014), we propose here a working model suggesting that these synergistic effects on pRhVI1activity may be due to the ability of light to remove the negative responses to sugar (via SRE) and to gibberellin (via WRKY710S). It will be interesting to identify the transcription factor network involved in the response ofRhVI1gene expression to light, gibberellin and sugars, and to unravel the molecular mechanisms behind this transcriptional regulation.

RhVI1 is expressed specifically in the epidermis cells of expanding tissues in Rosa buds

In situ hybridization showed that RhVI1 mRNA was mainly restricted to the epidermal cells of young leaves and leaf

primordia of Rosa sp. buds (Fig. 1), which is in accordance with the presence of two epidermis tissue-specificcis-elements (L1box, –847 and –355 bp) on the promoter ofRhVI1(Fig. 2).

The epidermal localization ofRhVI1expression supports its role in the expansion of young leaves and leaf primordia during bud outgrowth. Epidermal cells are formed from the outermost L1 layer in the shoot apical meristem (SAM), and are well known to be the growth-limiting cell layer in plants (Savaldi-Goldstein et al. 2007,Savaldi-Goldstein and Chory 2008,Nobusawa et al.

2013). Epidermal cells control organ growth by modulating the cell division activity of the ground tissue in response to the mechanical constraints from the epidermis (Fleming et al.

1997) or to the spatial distribution of cytokine biosynthesis within organs (Nobusawa et al. 2013).

Bud outgrowth ofRosasp. requires light perception by buds (Girault et al. 2008), and the morphogenetic effect of light is associated with an important flux of sucrose to the axillary buds involving the activity ofRhSUC2(Henry et al. 2011) along with stimulation of gibberellin neo-biosynthesis within the buds (Choubane et al. 2012). It is likely that under light, part of the bud-imported sucrose acts synergistically with light to up-regulate the expression ofRhVI1and its protein activity (Rabot et al.

2012;Fig. 6), in epidermal cells. This is in line with data of Mason et al. (2014) showing that sucrose, and not auxin, is the initial regulator of bud outgrowth in pea. Moreover, vacuolar invertase is well known to promote sink strength (Nage¨le et al. 2010) and cell expansion in a range of growing tissues (Davies and Robinson 1996,Tang et al. 1999,Andersen et al.

2002). Epidermal cells also exhibit high sensitivity to exogen- ously applied hormones, such as gibberellin (Steffens and Saute 2005), that promotes the elongation activity of the pre-formed leaves at the flank of the meristem (Hay et al. 2002).

These findings and our previous data (Girault et al. 2010, Henry et al. 2011,Choubane et al. 2012,Rabot et al. 2012) shed light on the synergistic effects of light and sugar and light and gibberellin on the expression of epidermis-localized RhVI1.

Fig. 8 Hypothetical working model for the proposed mechanism of pRhVI1expression regulation involving light and sugar and light and gibberellin synergistic interactions.

(11)

They also pave the way for uncovering thetrans-acting factors that contribute to light and sugar and/or gibberellin signaling and how epidermal cells could perceive and integrate all these endogenous and exogenous cues.

Materials and Methods

Plant material and growth conditions

TwoR. hybridacultivars were used for this study.Rosa hybridacv. ‘Radrazz’ was used to isolatepRhVI1.Rosa hybridacv. ‘Orange’ was used to generate the cell suspension. The cells were maintained in MS medium supplemented with 2,4-D (1 mg ml¹) and kinetin (1 mg ml¹) at 27C on an orbital shaker (150 r.p.m.), and subcultured weekly. To keep a selection pressure on the transformed cells and to eliminateAgrobacterium, two antibiotics, kanamycin (300 mg ml¹) and cefotaxime (500 mg ml¹), were added to the medium.

Isolation and characterization of pRhVI1

A genomic DNA library was used to isolate the promoter of theRhVI1gene (Rabot et al. 2012). The isolation of the genomic fragment was performed with a

‘BD Genome Walker System Universal’ kit (BD Bioscience) according to the manufacturer’s instructions and using the following primers forRhVI1:pRhVI1/

1 (5⁰ATGGACACTAGTACTTCTGCCTACGCTCCC) and AP1 (5⁰CCATCCTAAT ACGACTCACTATAGGGC).

In silico analysis of theRhVI1promoter sequence was performed using the PlantPAN program (http://plantpan.mbc.nctu.edu.tw/seq_analysis.php;Chang et al. 2008). Sequence alignment was performed using the ClustalW program (http://www.ebi.ac.uk/Tools/msa/clustalw2/;Larkin et al. 2007).

For 5⁰-end deletions ofpRhVI1, the promoter sequence was PCR-amplified using the following primers: 595pVI_S (5⁰AAAAAAGCAGGCTAGGCTGTCAA CCAAGACAAC), 468pVI_S (5⁰AAAAAA GCAGGCTTGTGCAGATTTGCTCAC C), 307pVI_S (5⁰AAAAAAGCAGGCTTCCTTTAGGTGAACATGCC), 148pVI_S (5⁰AAAAAAGCAGGCTCCTTCTCCGATCTCTCTC) and 21pVI_R (5⁰AAGAAA GCTGGGTGGGGGGTGGGTTTAAATAGG) carrying B1 and B2 GATEWAY^Õ recombination sites (Invitrogen). The PCR products were cloned according to the manufacturer’s instructions by BP recombination into pDONR207 and subsequently LR recombined into the destination vector pBI101R1R2 forRosa cell transformation (Baudry et al. 2006)

Rosa cell culture initiation and transformation

A rose cell suspension was initiated.fromR. hybridacv. ‘Orange’ leaves and petioles by inducing callogenesis on solid MS medium (Duchefa), supplemented with 2,4-D (1 mg l¹), kinetin (0.5 mg l¹) and sucrose (30 g l¹), pH 5.7, at 24C in the dark. Undifferentiated and friable calli were successfully obtained, and allowed for the production of a rose cell culture by callus disag- gregation in MS liquid culture medium supplemented with the same above- mentioned hormones, under orbital shaking at 160 r.p.m., at 27C in the dark.

Rose cells were transformed usingA. tumefaciensstrain EHA 105 carrying binary vectors with the chimeric fusion of theCauliflower mosiac virus(CaMV) 35S promoter::uidA reporter gene. Stable transformation was achieved by inoculating 4 ml of the rose cell suspension at the exponential phase (5 d after subculturing) with 100ml of the bacterial suspension (virulence was previously induced by treatment with 100mM acetosyringone) in a 5 cm diameter Petri dish. Co-culture was maintained for 48 h in the liquid MS medium, supplemented with 2,4-D (1 mg l¹) and kinetin (0.50 mg l¹), at 27C in the dark.

After this co-culture period that allowed for the growth of both plant and bacterial cells, rose cells were washed three times with the same liquid medium supplemented with the antibiotics kanamycin (300 mg l¹) and cefotaxime (500 mg l¹) purchased from Duchefa.

The cells were further transferred and grown on a three-layer medium consisting of: (i) a bottom solid layer of MS medium; (ii) a middle semi-solid layer of 1% (w/v) low-melting agarose in MS liquid medium containing the treated cells; and (iii) an upper liquid layer of MS medium to avoid cell dehy- dration. Each layer was supplemented with the same antibiotics at the above- mentioned concentrations. Thus, immersedR. hybridatransformed cells were

grown and selected at 24C in the dark for at least 2 months. The small transgenic calli were re-suspended in liquid MS medium with the same antibiotic concentrations, and cultured for several weeks at 27C on an orbital shaker at 160 r.p.m. in the dark to form a homogenous cell suspension. Integration of the transgene into the genome ofRosacells was confirmed by histochemical GUS assay.

The transgenic rose cell culture obtained by this new procedure and the untransformed control cell suspension were maintained by weekly subculturing of the cells in fresh medium [10 : 40 (v/v)], with or without antibiotics, respectively, under light and dark conditions.

GUS fluorimetric assay and histochemical assay

The GUS fluorimetric assay was performed on the different stably transformed cell suspensions with eachpRhVI1deletion according toJefferson et al. (1987).

Fluorescence emission was measured using a ‘Fluostar Omega’ microplate fluor- ometer (BMG Labtech).

To perform the fluorimetric assay, the cells suspensions were treated with different sugars or sugar analogs and a hormone (GA3) as follows: after 2 d of subcul- ture, the cells were rinsed three times with sugar-free MS medium, and then 58 mM sugars (mannitol, sucrose, glucose and fructose) were added. The effect of gibberellins was also studied by using 5mM GA3, a bioactive form of gibberellin.

Statistical analysis

All results reported here are means±SEM of at least three independent experiments. We performed an analysis of variance (ANOVA) to test the effect of the different treatments on the GUS activity of the vacuolar invertase promoter. ANOVA was carried out using Statgraphicsplus^ÕSoftware.

In situ hybridization

Preparation of digoxigenin (DIG)-labeled RNA probes for in situ hybridization. DIG-labeled RNA probes were prepared using an in vitro transcription kit (Riboprobe Combination Systems, Promega) according to the manufacturer’s instructions. The riboprobes were synthesized from the full-length RhVI1clones. Antisense and sense probes were transcribed from SP6 or T7 RNA polymerase promoters after linearization of the vector withNcoI orPstI, respectively. Full-length probes were treated by alkaline hydrolysis, as described previously byWinzer-Serhan et al. (1999), to produce 250 bp fragments.

Cytochemical methods for light microscopy. Harvested buds were fixed on ice under vacuum in 4% (w/v) paraformaldehyde in 1phosphate-buffered saline (pH 7) for 3 h, then dehydrated in an ethanol series, transferred to an embedding solvent (Histochoice, Sigma) and finally saturated with paraffin (seePe´ron et al. 2012).

Longitudinal sections (10mm) were stained with 0.05% toluidine blue O in NaPO4

buffer (pH 5.5) for 10 min before removing the paraffin (Lozano-Baena et al. 2007). The slides were washed with distilled water and dried at room temperature. The paraffin was removed with Histochoice and the slides were mounted with Eukitt (Sigma).

In situ hybridization forRhVI1transcript localization on bud sections was conducted as previously described inPe´ron et al. (2012).

Funding

This work was supported by the Ministe`re de l’Agriculture et de la P^eche (France).

Acknowledgments

We thank M. Laffaire, C. Bouffard, A. Lebrec, B. Dubuc and S. Chalain for technical assistance in supplying plant material of excellent quality, and R. Gardet for greenhouse management.

We also thank B. Dubreucq (UMR1318 INRA-AgroParisTech) for providing pBI101R1R2 vector.

(12)

Disclosures

The authors have no conflicts of interest to declare.

References

Acevedo-Hernandez, G.J., Leon, P. and Herrerra-Estrella, L.R. (2005) Sugar and ABA responsiveness of a minimal RBCS light-responsive unit is mediated by direct binding of ABI4.Plant J.43: 506–519.

Abe, H., Urao, T., Ito, T., Seki, M., Shinozaki, K. and Yamaguchi-Shinozaki, K.

(2003) Arabidopsis AtMYC2 (bHLH) and AtMYB2 (MYB) function as transcriptional activators in abscisic acid signaling.Plant Cell15: 63–78.

Alabadı´, D. and Bla´zquez, M.A. (2009) Molecular interactions between light and hormone signaling to control plant growth.Plant Mol. Biol.

69: 409–417.

Andersen, M.N., Wu, Y., Asch, F., Jensen, C.R., Næsted, H., Mogensen, V.O.

et al. (2002) Soluble invertase expression is an early target of drought stress during the critical, abortion-sensitive phase of young ovary development in maize.Plant Physiol.130: 591–604.

Atanassova, R., Leterrier, M., Gaillard, C., Agasse, A., Sagot, E., Coutos- The´venot, P. et al. (2003) Sugar-regulated expression of a putative hexose transport gene in grape.Plant Physiol.131: 326–334.

Azari, R., Reuveni, M., Evenor, D., Nahon, S., Shlomo, H., Chen, L. et al.

(2010) Overexpression of UV-DAMAGED DNA BINDING PROTEIN 1 links plant development and phytonutrient accumulation in high pig- ment-1 tomato.J. Exp. Bot.61: 3627–3637.

Bai, W.Q., Xiao, Y.H., Zhao, J., Song, S.Q., Hu, L. et al. (2014) Gibberellin overproduction promotes sucrose synthase expression and secondary cell wall deposition in cotton fibers.PLoS One9: e96537.

Barratt, D.H., Derbyshire, P., Findlay, K., Pike, M., Wellner, N., Lunn, J. et al.

(2009) Normal growth of Arabidopsis requires cytosolic invertase but not sucrose synthase.Proc. Natl Acad. Sci. USA106: 13124–13129.

Baudry, A., Caboche, M. and Lepiniec, L. (2006) TT8 controls its own expression in a feedback regulation involving TTG1 and homologous MYB and bHLH factors, allowing a strong and cell-specific accumulation of flavonoids in Arabidopsis thaliana.Plant J.46: 768–779.

Bihmidine, S., Hunter, C.T. III, Johns, C.E., Koch, K.E. and Braun, D.M. (2013) Regulation of assimilate import into sink organs: update on molecular drivers of sink strength.Front. Plant Sci.4: 177.

Bolouri-Moghaddam, M.R., Le Roy, K., Xiang, L., Rolland, F. and Van den Ende, W. (2010) Sugar signalling and antioxidant network connections in plant cells.FEBS J.277: 2022–2037.

Bolouri Moghaddam, M.R. and Van den Ende, W. (2013) Sugars, the clock and transition to flowering.Front. Plant Sci.4: 22.

Borello, U., Ceccarelli, E. and Giuliano, G. (1993) Constitutive, light- responsive and circadian clock-responsive factors compete for the different l box elements in plant light-regulated promoters.Plant J.4:

611–619.

Borisjuk, L., Rolletschek, H., Radchuk, R., Weschke, W., Wobus, U. and Weber, H. (2004) Seed development and differentiation: a role for metabolic regulation.Plant Biol.6: 375–386.

Cakir, B., Agasse, A., Gaillard, C., Saumonneau, A., Delrot, S. and Atanassova, R. (2003) A grape ASR protein involved in sugar and abscisic acid signaling.Plant Cell15: 2165–2180.

Chang, W.C., Lee, T.Y., Huang, H.D., Huang, H.Y. and Pan, R.L. (2008) PlantPAN: plant promoter analysis navigator, for identifying combina- torial cis-regulatory elements with distance constraint in plant gene groups.BMC Genomics9: 561.

Chen, M., Chory, J. and Fankhauser, C. (2004) Light signal transduction in higher plants.Annu. Rev. Genet.38: 87–117.

Chen, P.W., Chiang, C.M., Tseng, T.H. and Yu, S.M. (2006) Interaction between rice MYBGA and the gibberellin response element controls tissue-specific sugar sensitivity of alpha-amylase genes.Plant Cell18:

2326–2340.

Choubane, D., Rabot, A., Mortreau, E., Legourrierec, J., Pe´ron, T., Foucher, F.

et al. (2012) Photocontrol of bud burst involves gibberellin biosynthesis in Rosa sp.J. Plant Physiol.169: 1271–1280.

Choudhury, S.R., Roy, S., Das, R. and Sengupta, D.N. (2008) Differential transcriptional regulation of banana sucrose phosphate synthase gene in response to ethylene, auxin, wounding, low temperature and different photoperiods during fruit ripening and functional analysis of banana SPS gene promoter.Planta229: 207–223.

Claeys, H., De Bodt, S. and Inze´, D. (2014) Gibberellins and DELLAs: central nodes in growth regulatory networks.Trends Plant Sci.4: 231–239.

Condliffe, P.C., Davey, M.R., Brian, P.J., Koehorst-van Putten, H. and Visser, P.B. (2003) An optimised protocol for rose transformation ap- plicable to different cultivars.Acta Hort.612: 115–120.

Ciolkowski, I., Wanke, D., Birkenbihl, R.P. and Somssich, I.E. (2008) Studies on DNA-binding selectivity of WRKY transcription factors lend structural clues into WRKY-domain function.Plant Mol. Biol.68:

81–92.

Dalchau, N., Baek, S.J., Briggs, H.M., Robertson, F.C., Dodd, A.N. et al. (2011) The circadian oscillator gene GIGANTEA mediates a long-term response of the Arabidopsis thaliana circadian clock to sucrose.Proc.

Natl Acad. Sci. USA108: 5104–5109.

Das, P.K., Shin, D.H., Choi, S.B., Yoo, S.D., Choi, G. and Park, Y.I. (2012) Cytokinins enhance sugar-induced anthocyanin biosynthesis in Arabidopsis.Mol. Cells34: 93–101.

Davies, C. and Robinson, S.P. (1996) Sugar accumulation in grape berries.

Cloning of two putative vacuolar invertase cDNAs and their expression in grapevine tissues.Plant Physiol.111: 275–283.

De Lucas, M., Davie`re, J.M., Rodrı´guez-Falco´n, M., Pontin, M., Iglesias- Pedraz, J.M., Lorrain, S. et al. (2008) A molecular framework for light and gibberellin control of cell elongation.Nature451: 480–484.

De Lucas, M. and Prat, S. (2014) PIFs get BRright: PHYTOCHROME INTERACTING FACTORs as integrators of light and hormonal signals.

New Phytol.202: 1126–1141.

Dohm, A., Ludwig, C., Nehring, K. and Debener, T. (2001) Somatic embryogenesis in roses.Acta Hort.547: 341–347.

Donald, R.G. and Cashmore, A.R. (1990) Mutation of either G box or I box sequences profoundly affects expression from the Arabidopsis rbcS-1A promoter.EMBO J.9: 1717–1726.

Dorion, S., Lalonde, S. and Saini, H.S. (1996) Induction of male sterility in wheat by meiotic-stage water deficit is preceded by a decline in invertase activity and changes in carbohydrate metabolism in anthers.Plant Physiol.111: 137–145.

Eckardt, N.A. (2003) A component of the cryptochrome blue light signaling pathway.Plant Cell15: 1051–1052.

Firoozabady, E., Moy, Y., Courtney-Gutterson, N. and Robinson, K. (1994) Regeneration of transgenic rose (Rosa hybrida) plants from embryo- genic tissue.Nat. Biotechnol.12: 609–613.

Fleming, A.J., McQueen-Mason, S., Mandel, T. and Kuhlemeier, C. (1997) Induction of leaf primordia by the cell wall protein expansion.Science 276: 5317–1415.

Gallagher, J.A. and Pollock, C.J. (1998) Isolation and characterization of a cDNA clone from Lolium temulentum L. encoding for a sucrose hydro- lytic enzyme which shows alkaline/neutral invertase activity.J. Exp. Bot.

49: 789–795.

Girault, T., Abidi, F., Sigogne, M., Pelleschi-Travier, S., Boumaza, R., Sakr, S.

et al. (2010) Sugars are under light control during bud burst in Rosa sp.

Plant Cell Environ.33: 1339–1350.

Girault, T., Bergougnoux, V., Combes, D., Viemont, J.D. and Leduc, N.

(2008) Light controls shoot meristem organogenic activity and leaf primordia growth during bud burst in Rosa sp.Plant Cell Environ.31:

1534–1544.

Goetz, M., Godt, D.E., Guivarch, A., Kahmann, U., Chriqui, D. and Roitsch T. (2001) Induction of male sterility in plants by metabolic engineering of the carbohydrate supply.Proc. Natl Acad. Sci. USA98:

6522–6657.