Molecular physiology of adventitious root formation in Petunia hybrida cuttings: involvement of wound response and primary metabolism


Author for correspondence:
Mohammad Reza Hajirezaei
Tel:+49 39482 5266
Fax: +49 39482 5515


  • • Adventitious root formation (ARF) in the model plant Petunia hybrida cv. Mitchell has been analysed in terms of anatomy, gene expression, enzymatic activities and levels of metabolites. This study focuses on the involvement of wound response and primary metabolism.
  • • Microscopic techniques were complemented with targeted transcript, enzyme and metabolite profiling using real time polymerase chain reaction (PCR), Northern blot, enzymatic assays, chromatography and mass spectrometry.
  • • Three days after severance from the stock plants, first meristematic cells appeared which further developed into root primordia and finally adventitious roots. Excision of cuttings led to a fast and transient increase in the wound-hormone jasmonic acid, followed by the expression of jasmonate-regulated genes such as cell wall invertase. Analysis of soluble and insoluble carbohydrates showed a continuous accumulation during ARF. A broad metabolite profiling revealed a strong increase in organic acids and resynthesis of essential amino acids.
  • • Substantial changes in enzyme activities and metabolite levels indicate that specific enzymes and metabolites might play a crucial role during ARF. Three metabolic phases could be defined: (i) sink establishment phase characterized by apoplastic unloading of sucrose and being probably mediated by jasmonates; (ii) recovery phase; and (iii) maintenance phase, in which a symplastic unloading occurs.


Roots formed de novo from differentiated cells (e.g. of stem or hypocotyl tissues) are defined as adventitious roots (Casson & Lindsey, 2003). Adventitious root formation (ARF) in leafy stem cuttings is a crucial physiological process for propagation of many ornamental plant species. Despite intensive control of environmental factors in the modern propagation industry, high economic losses still occur as a result of insufficient rooting (Sorin et al., 2006). It has been repeatedly shown that ARF of cuttings can be improved via conditioning of the donor plant, for example by the control of nitrogen supply (Druege et al., 2000) or by application of arbuscular mycorrhizal fungi (Druege et al., 2006). However, the lack of knowledge of the molecular physiological basis hampers the development and use of reliable technologies.

The formation of adventitious roots is a complex process that involves successive developmental phases requiring different hormonal signals and other factors (De Klerk et al., 1999). Severance of a cutting from the donor plant has two consequences: isolation from functional integrity of the ‘whole plant’ and injury. After detachment of the shoot, basipetal polar transport of auxin contributes to auxin accumulation in the stem base (Garrido et al., 2002) and the rise of free auxin in the basal stem very probably contributes to the early events of ARF (Blakesley, 1994; De Klerk et al., 1999; Sorin et al., 2005). There is increasing evidence that ARF is also dependent on the action of ethylene (Clark et al., 1999; Shibuya et al., 2004), production of which is caused by wounding during the cutting process (Blakesley, 1994).

In the majority of plants wounding also leads to an increase in jasmonates (Schilmiller & Howe, 2005) and the role of jasmonates in ARF is almost unknown. Jasmonates, such as jasmonic acid (JA) and its various metabolites, have been recognized as being signals in plant responses to most of biotic and abiotic factors (Wasternack, 2007). They are synthesized via the octadecanoid pathway (Feussner & Wasternack, 2002) and the allene oxide cyclase (AOC) is regarded as crucial for JA biosynthesis (Wasternack & Hause, 2002). Endogenous increase in jasmonate levels following wounding leads to specific gene expression. Those genes code for enzymes of JA biosynthesis itself (Wasternack, 2007) or for enzymes involved in sugar metabolism such as cell wall invertases, thereby influencing source–sink relations in plants (Roitsch & González, 2004; Schaarschmidt et al., 2006).

Since ARF, an energy-requiring process, needs carbon skeletons, it relies on adequate supply of carbohydrates to the region of root regeneration (Veierskov, 1988). Low carbohydrate levels in cuttings at the beginning of rooting limit the speed or intensity of subsequent ARF (Veierskov et al., 1982; Druege et al., 2004), whereas application of sugars to the rooting medium increases subsequent root formation (Eliasson, 1978; Li & Leung, 2000; Takahashi et al., 2003). Moreover, application of different sugars suggests distinct functions of glucose, sucrose and starch in the different phases of ARF (Li & Leung, 2000; Corrêa et al., 2005) and points to a regulatory role of sugars (Druege et al., 2000; Corrêa et al., 2005). However, how carbohydrate metabolism affects this developmental process remains unsolved.

Despite the central role of biochemical and physiological events during ARF, only a limited number of molecular studies of ARF have been performed. In order to obtain molecular markers for genetic analysis (Sorin et al., 2006), differentially expressed genes (Brinker et al., 2004) and gene products (Sorin et al., 2006) were identified. Despite intensive studies, no conclusive models concerning the relationships between distinct transcript accumulations, enzyme activities, metabolic patterns and ARF have been reported.

The genus Petunia is a plant of high economic importance in world-wide horticulture and provides important qualities to serve as model plant for studying plant development (Gerats & Vandenbussche, 2005). The aim of the present investigation was to establish Petunia hybrida as model system for studying ARF in leafy stem cuttings and to analyse molecular and biochemical processes involved. The focus was set on the contribution of JA as important regulator of the wound response and of primary metabolism including sucrose and starch accumulation. Time-course analyses of transcript levels, enzyme activities and metabolite levels were combined with histological study of ARF in order to address the questions about the roles JA, carbohydrates, amino acids and related metabolic pathways might play during ARF. In that way, molecular and physiological markers were identified which could be used for characterization and improvement of ARF.

Materials and Methods

Plant material, growth and harvesting conditions

Excised leafy cuttings of P. hybrida cv. Mitchell harboring four to five leaves of similar size were placed in plastic trays containing Perlite (‘Perligran A’, particle size 0–6 mm; Knauf Perlite GmbH, Dortmund, Germany). Trays containing cuttings were covered to maintain a humid environment, put in a phytotron and cultivated under the following conditions: 20°C (night) and 22°C (day), humidity 60% (night) and 85% (day), 10 h light (250 µmol m−2 s−1) and 14 h dark. At specific developmental stages of ARF, 5 mm of each cutting base (rooting zone) were immediately frozen in liquid N2 and stored at −80°C or fixed in a solution of formalin, ethanol and acetic acid for anatomical investigation (Gerlach, 1984).

Anatomical investigation of cutting base and histological staining of starch

The anatomical examination was performed as described by Haensch (2004). Stem segments were embedded in hydroxyethylmethacrylate (Histo-Technique-Set Technovit 7100; Kulzer, Wehrheim, Germany) and cut into 6 µm sections using a Jung CM 1800 microtome with type 818 disposable microtome blades (Leica Instruments, Nussloch, Germany). Sections were stained with 0.05% (w : v) toluidine blue O (Serva, Heidelberg, Germany) in 1% (w : v) sodium tetraborate decahydrate buffer (Hutchinson et al., 1996) and covered with Entellan (Merck, Darmstadt, Germany). In parallel, transverse sections, 150 µm thick, were produced using a Vibratome VT 1000S (Leica) and incubated for 10 min in iodine (Sigma-Aldrich, Germany). Microscopic analyses were performed using an AxioImager A1 microscope in combination with an AxioCam MRc5 camera (Carl Zeiss, Jena, Germany).

Carbohydrate measurement

Soluble and insoluble sugars were determined as described by Trethewey et al. (1998) with minor modifications. Starch hydrolysis was carried out by incubation the aliquots in a buffer containing 50 mm sodium acetate, pH 5.2 and 7 units mg−1 of amyloglucosidase (Roche Diagnostics GmbH, Mannheim, Germany). Determination of produced glucose was performed according to Hajirezaei et al. (2000).

Isolation of AOC from P. hybrida (PhAOC)

Total RNA from wounded leaves of P. hybrida was isolated using RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer instructions. Synthesis of first-strand cDNA was performed with BD SMART RACE cDNA amplification kit (BD Biosciences Clontech, Erembodegem, Belgium) using 5′-RACE CDS primer and SMART II A oligo. Resulting cDNA served as template for PCR reactions with 5′-RACE CDS primer and a degenerated primer [5′-TT(CT)TA(CT)TT(CT)GG(CG)GA(CT)TA(CT)GG-3′] based on the conserved sequences of known AOCs of Solanum lycopersicum (AF384374), Nicotiana tabacum (AJ308487) and Solanum tuberosum (AY135641). The resulting DNA fragment of c. 430 bp contained the 3′end. Specific primers (primer 1: 5′-AATGCCGGATCCACCAGTGAC-3′; primer 2: 5′-TGCCGGATCCACCAGTGACTGC AAGATATG-3′) were used to amplify the 5′ end resulting in a 972 bp cDNA of PhAOC, which includes the full-length coding region.

Real-time (RT-)PCR for PhAOC

Plant material was pooled from nine different cuttings and 100 mg FW of each sample was used for isolation of RNA by RNeasy Mini Kit (Qiagen) followed by treatment with DNaseI (Qiagen). Using M-MLV reverse transcriptase (Promega, Mannheim, Germany) cDNA synthesis was performed with total RNA and oligo(dT)19 primer at 42°C for 50 min. The level of PhAOC mRNA was investigated by RT-PCR as described by Miersch et al. (2008) except that primers designed from PhAOC and Cytoplasmic ribosomal protein S13 of P. hybrida (PhCyRiPro; CV297717) were used: PhAOC 5′-CGGCATTTTTGCAGGAGTTT-3′ and 5′-CCAGCAACTCA GATGGCAGAT-3′ resulting in a fragment of 115 bp; PhCyRiPro 5′-AAGCTCCCACCTGTCTGGAAA-3′ and 5′-AACAGATTGCCGGAAGCCA-3′ resulting in a fragment of 102 bp. Each reaction mix contained a 15 ng RNA equivalent of cDNA and 1 pm gene-specific primers. All assays were performed on at least three biological replicates in three technical replicates each. The ΔCt (threshold cycle) values were calculated by subtracting Ct values of PhAOC from the arithmetic mean of the Ct value of PhCyRiPro. The ΔΔCt values were calculated by subtracting ΔCt values obtained for each samples from ΔCt value of the plant material obtained directly after excision of cuttings.

RNA Isolation and Northern blot analysis

Total RNA was extracted from petunia cutting base as described by Logemann et al. (1987). Thirty micrograms per sample were separated on a 1.5% (w : v) formaldehyde-agarose gel (Sambrook et al., 1989), transferred to a nitrocellulose membrane (GeneScreen, NEN Life Science Products, Waltham, MA, USA) and fixed by UV cross-linking. Radioactive labeling of cDNA fragments was performed using the High Prime kit (Roche Diagnostics, Mannheim, Germany) and [α-32P]-dCTP. Hybridization was carried out as described previously (Herbers et al., 1994) and signals were detected by exposure to Kodak X-ray films (Sigma, Taufkirchen, Germany). The following probes were used: Petunia cyclin B1 (AJ250315) was amplified from Petunia cDNA derived from the total RNA with specific primer pairs (CycB1 for primer: 5′-AGGTACCAGCCAAGAAGAAGG-3′; CycB1 rev primer: 5′-TGCGCTAATGCCAACTAACTG-3′) resulting in a fragment of 449 bp; fragments of the genes for potato sucrose synthase (SuSy, c. 900 bp, P10691), tomato cell wall invertase (800 bp, AAM28823) and potato vacuolar invertase (800 bp, ABF18956) were isolated from the appropriate vectors (Zrenner et al., 1995); for monosaccharide transporter STP4, a cDNA fragment (1000 bp) was isolated using a petunia cDNA library.

Determination of AOC enzyme activity and quantitative analyses of JA and 12-oxo-phytodienoic acid (OPDA)

One gram (FW) of homogenized plant material pooled from nine different cuttings was used for determination of AOC activity in a coupled enzyme assay according to Stenzel et al. (2003). The AOC protein levels were determined by immunoblot analyses using an antibody raised against tomato AOC (Stenzel et al., 2003). The amount of JA and OPDA was determined as described by Miersch et al. (2008).

Extraction and activity measurement of enzymes

The extraction of enzymes was carried out according to the method of Zrenner et al. (1995) with minor modifications. Instead of phenylmethylsulfonyl fluoride (PMSF), 0.1 mm pefabloc phosphatase inhibitor was used and bovine serum albumin (BSA) was omitted.

The activity of cytosolic and vacuolar invertases was assayed as described by Zrenner et al. (1996). Cell wall-bound acid invertase was measured using the remaining pellet after enzyme extraction. Pellets were washed two times with a buffer containing 50 mm Tris-HCl (pH 6.8) and 5 mm MgCl2. Incubation was carried out in a buffer containing 50 mm sodium acetate (pH 5.2), 0.5 m sucrose and complete pellets of protein extract in a final volume of 100 µl at 37°C for 180 min and neutralized by adding 10 µl of 1 m Tris-HCl (pH 8.0). The reaction was stopped at 95°C for 5 min. Blank samples were prepared without plant material. Samples were centrifuged for 1 min at 10 000 g and glucose produced was measured in the supernatant as described in Hajirezaei et al. (2000). Sucrose synthase was assayed as described by Zrenner et al. (1995) with some modifications. The determination of UDP-glucose produced was carried out by liquid chromatography–mass spectrometry (LC-MS) according to Chen et al. (2005). Activity of cytosolic glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was measured as described by Hajirezaei et al. (2006). Activity of glucose-6-phosphate dehydrogenase was determined in a buffer containing 100 mm glycylglycine (pH 8.0), 150 mm MgSO4 and 60 mm NADP. The reaction was started by the addition of 15 mm glucose-6-phosphate and the absorption of NADP was followed at a wavelength of 340 nm. The activity of phosphofructokinase (PFK) and pyruvate kinase (PK) were performed according to Hajirezaei et al. (1994) and for aldolase according to Haake et al. (1998). The activity of cytosolic fructose-1,6-bisphosphatase (FBPase) was measured according to Kelly et al. (1982). Phosphoenolpyruvate (PEP) carboxylase and malate dehydrogenase activities were determined as described by Rolletschek et al. (2004) and Jenner et al. (2001), respectively. Protein content of enzyme extracts was determined according to Bradford (1976).

Targeted metabolite profiling

To measure the levels of important metabolites involved in primary metabolism, two frozen cutting bases with a fresh weight of c. 50 mg were used and processed as described by Chen et al. (2005).

Soluble amino acids were determined as described by Rolletschek et al. (2002). Quantification of the content of individual amino acid was carried out by using Empower Pro software (Waters, Milford, MA, USA).

The ATP and ADP-glucose contents were determined using a highly sensitive fluorescence method according to Haink & Deussen (2003). Chromatograms were integrated in all cases using the Empower Pro software (Waters).

Measurement of carbon to nitrogen ratio

The relative content of total carbon and total nitrogen was measured according to Rolletschek et al. (2002) by an elemental analyser (Vario EL; Elementaranalysen-systeme GmbH, Hanau, Germany). Each value is represented by the mean of six independent replicates ± SE. Each biological sample was measured in duplicate.

Recovery experiment

A recovery experiment was carried out to evaluate the stability of the appropriate metabolites during the extraction procedure. For different intermediates, a standard mix of desired metabolites was prepared. The mixture was added either to stem cuttings or to the extraction buffer before the extraction (each n = 5). The percentage of found metabolites was calculated after analysis by LC-MS: UDP-glucose 125 ± 4%, citrate 87 ± 10%, malate 72 ± 8%, hexose phosphate 82 ± 7%, 3PGA 62 ± 2%, pyruvate 70 ± 6%, ADPGlc 106 ± 9% and ATP 118 ± 0.3%.


Establishment of a stable rooting system and anatomy of ARF in P. hybrida

To avoid artificial in vitro conditions using, for example, agar as substrate and exogenous hormones, which might lead to a misinterpretation of the pathways that are modulated during ARF, an ex vitro experimental system was established: shoot tip cuttings of P. hybrida of similar size with four to five leaves were transferred without any external additives to Perlite as neutral substrate. Roots emerged after 8 d from the first 5 mm of the cutting stem base. All analyses were therefore carried out within the first 8 d using the rooting zone.

Representative cross-sections of stems were taken and showed that first new meristematic cells (dense cytoplasm, large nucleus) appeared at 72 hpe (hours post excision, Fig. 1c), well-developed young root meristems appeared at 96 hpe (Fig. 1d) and root primordia with apical meristem and differentiation of the root body appeared at 144 hpe (Fig. 1e). Roots with first cells characteristic for the elongation zone were regularly visible at 192 hpe (Fig. 1f). To check whether the first cytological signs of new meristematic cells are also detectable at molecular level, expression of CycB1 encoding a mitotic B1 cyclin and serving as marker for mitotic cells was monitored (Porceddu et al., 1999). The RNA from this gene started to accumulate 48 hpe, the amounts increased up to 144 hpe and declined again at 192 hpe (see the Supporting Information, Fig. S1).

Figure 1.

Anatomy of adventitious root formation (a–f) and starch accumulation (g–i) in the stem base of petunia (Petunia hybrida) cuttings. In all micrographs, cross-sections from c. 1–4 mm above the excision site are shown. (a,b) 0 hours post excision (hpe), typical stem anatomy consisting of the cortex (co), the pith parenchyma (pi) and a ring of vessels (r) with outer phloem (oph), the cambium (ca), the xylem (xy) and the inner phloem (iph); (c) 72 hpe, first meristematic cells (mc) of developing root meristems, that is, small cells with a dense cytoplasm and a large nucleus; (d) 96 hpe, first root meristems (me); (e) 144 hpe, first differentiating root primordia with an organized meristem and a backward differentiation of cells of the root body containing root cortex (ro) and vascular bundle (v); (f) 192 hpe, first roots with vascular bundles (v) in the center surrounded by elongated cells (ec) of the elongation zone; (g–i) histochemical staining of starch granules at 0 hpe (g), 24 hpe (h) and 72 hpe (i). Arrows show the presence of starch. Bars, (a), 500 µm; (b–i), 100 µm.

Levels of JA, OPDA and transcripts during ARF

Because excising of cuttings from stock plants involves mechanical wounding, the contents of the wound responding compounds JA and OPDA were monitored. The JA content was raised about 12-fold within 0.5 hpe, but returned to basal levels at 6 hpe (Fig. 2a), while OPDA levels increased about four-fold within 0.5 hpe, remained at this level till 6 hpe and dropped at 12 hpe to basal levels (Fig. 2b). There was no significant change in levels of JA and OPDA during further development of adventitious roots. Both metabolites showed similar kinetics in leaves and other stem parts of the cuttings with lower maximum values pointing to a systemic response in these organs (data not shown).

Figure 2.

Accumulation of jasmonic acid (JA), 12-oxophytodienoic acid (OPDA), allene oxide cyclase (AOC) transcripts and protein as well as AOC activity in stem bases of Petunia hybrida cuttings. Contents of (a) JA and (b) OPDA, both given as pmol g−1 FW; (c) relative PhAOC transcript accumulation, ΔΔCt values are shown; (d) AOC enzyme activity throughout different developmental stages given as nmol g−1 FW of enzymatically formed OPDA and AOC protein accumulation detected by immunoblot (inset). Each value is represented by the mean of three independent replicates ± SE. Hours post excision (hpe) of the cuttings is given on the x-axis.

To check whether the increase in JA levels by wounding was accompanied by altered expression of JA biosynthesis genes, a cDNA coding for PhAOC (EU652410) was cloned. Allene oxide cyclase in P. hybrida is encoded by a single copy gene (data not shown) as in other solanaceous species (Ziegler et al., 2000). The deduced protein sequence (26.4 kDa) revealed 85% identical amino acids to the tomato AOC (Ziegler et al., 2000). Heterologous expression of the PhAOC cDNA without the predicted chloroplast-targeting signal (identified by ChloroP and TargetP), was performed in Escherichia coli. Enzymatic assays for AOC activity with extracts of bacteria carrying the PhAOC-containing vector resulted in the exclusive formation of cis-(+)-OPDA (data not shown), which is indicative of AOC activity (Ziegler et al., 1999, 2000).

PhAOC mRNA levels were altered in the stem base upon generation of cuttings and showed a significant accumulation with a transient maximum between 1 and 2 hpe (Fig. 2c). As for JA, similar kinetics of PhAOC transcript accumulation for leaves and other stem parts of cuttings and no significant increase at later stages was recorded (data not shown). This does not correlate with the enzymatic activities, which remained constant over the whole period of ARF (Fig. 2d).

Transcript accumulation of genes encoding enzymes involved in primary carbohydrate metabolism

Focusing on key enzymes of carbohydrate metabolism, RNA accumulation analyses over the time-course of ARF were performed using potato and petunia fragments as probes (Fig. 3). An increased accumulation of transcripts coding for cell wall invertase and the apoplastic localized monosaccharide transporter STP4 occurred at 2 and 4 hpe. While the transcript level of cell wall invertase already decreased 6 hpe, there was still a high transcript level detectable for STP4 up to 48 hpe. Additionally, the transcript level of vacuolar invertase was high throughout the first 5 d after excision followed by a decrease to a low level. No significant changes of the transcript level of sucrose synthase (SuSy) could be detected at all monitored stages of ARF (Fig. 3).

Figure 3.

Transcript accumulation of genes coding for enzymes involved in carbohydrate metabolism.

Activities of enzymes involved in primary carbohydrate metabolism

To assess whether carbohydrate metabolism was changing during ARF, activity of enzymes involved in various pathways was analysed (Fig. 4). To this end, several enzymes were chosen that might play an important role in regulation during the transition of the carbon source sucrose to energy production and protein synthesis. For all determinations of enzymatic activities it must be taken into consideration that the respective values represent their total cellular activity without distinguishing their possible functions in different subcellular compartments. A general decrease was observed for sucrose metabolism, with particular differences among the enzymes. The activities of vacuolar (Fig. 4b) and cytosolic invertase (Fig. 4c) remained unchanged during the first 6 hpe and then decreased continuously until 192 hpe to 20% and 25% of the initial value, respectively. The activity of sucrose synthase (SuSy) (Fig. 4d) decreased to 40% at 4 hpe and increased again until 12 hpe to the initial value followed by a continuous decrease to 20% at 192 hpe. Unlike the soluble invertases, cell wall invertase activity increased at the beginning threefold to reach a maximum at 6 hpe and reduced again approximately to its basal activity before roots emerged (Fig. 4a).

Figure 4.

Changes of the activity of key enzymes in carbohydrate metabolism during adventitious root formation (ARF) in petunia (Petunia hybrida). (a) Cell wall invertase, (b) vacuolar invertase, (c) cytosolic invertase, (d) sucrose synthase, (e) phosphofructokinase (PFK), (f) cytosolic glyceraldehyde-3-phosphate dehydrogenase (GAPDH), (g) cytosolic aldolase, (h) cytosolic fructose-1,6-bisphosphatase (FBPase), (i) glucose-6-phosphate dehydrogenase (Glc6PDH), (j) pyruvate kinase (PK). Data are shown as nmol min−1 mg−1 protein. Each value is represented by the mean of seven independent replicates ± SE. This figure was prepared using the Software vanted, version 1.5 developed by Junker et al. (2006).

Among the enzymes involved in glycolysis and gluconeogenesis, the activity of PFK increased constantly up to fourfold (Fig. 4e). An increase of twofold was observed for the cytosolic GAPDH until 6 hpe, but it decreased again to the basal level (Fig. 4f). Cytosolic aldolase activity showed no significant change up to 144 hpe. Subsequently, a twofold increase was observed followed by a decrease until 192 hpe to the basal level (Fig. 4g). The activity of PK exhibited fluctuation up to 12 hpe and revealed a slight but continuous increase thereafter (Fig. 4j), whereas cytosolic FBPase remained unchanged during the first period but was reduced to 40% in the later course of ARF (Fig. 4h). Furthermore, the activity of glucose-6-phosphate dehydrogenase, the key enzyme in pentose phosphate pathway, exhibited a threefold increase beginning at 12 hpe (Fig. 4i).

Finally, two enzymes involved in citric acid cycle were monitored (Fig. S2). While there was a nearly threefold increase in PEP carboxylase activity after 48 hpe (Fig. S2a), the activity of malate dehydrogenase did not change significantly up to 96 hpe, increased two times at 144 hpe and fell again to the basal value upon root emergence (Fig. S2b).

Levels of metabolites during ARF

The levels of soluble sugars including glucose, fructose and sucrose were low in the first hours post excision, but a continuous increase was observed after 24 hpe and reached the highest level before the emergence of roots (Fig. 5a,b). The content of starch began to increase after 72 hpe to a 16-fold higher level at 192 hpe (Fig. 5c). To find out whether starch accumulation was reflected by a visible starch granule aggregation, a detailed microscopic investigation was carried out. At early stages, starch granules were only detectable in the starch parenchyma cells adjacent to the inner cortex (Fig. 1g,h), but at 72 hpe starch granules accumulated in the starch parenchyma cells and the pith (Fig. 1i).

Figure 5.

Alterations in the content of sugars and metabolites involved in glycolysis, citric acid cycle and respiration during adventitious root formation (ARF) in petunia (Petunia hybrida). (a) Glucose (closed circles) and fructose (open circles), (b) sucrose, (c) starch, (d) UDP-glucose, (e) ADP-glucose, (f) 3PGA, (g) hexose-P, (h) ATP (i) isocitrate, (j) citrate, (k) α-ketoglutarate and (l) malate. All data are shown as either µmol g−1 FW (a–c) or nmol g−1 FW (d–l). Each value is represented by the mean of five independent replicates ± SE.

To assess the levels of other intermediates during ARF in petunia cuttings, a metabolite profiling was performed using ion chromatography–mass spectrometry (ICMS) and fluorescence HPLC. The level of UDP-glucose showed a slight increase between 4 hpe and 8 hpe (Fig. 5d), whereas the level of hexose phosphate was reduced to 50% immediately after excision, increased slightly during the next 12 h and decreased again (Fig. 5g). 3-Phosphoglycerate level decreased slightly during ARF (Fig. 5f). Pyruvate amounts increased fourfold at 48 hpe and decreased again at later stages (data not shown). After 12 hpe, a continuous increase in ADP-glucose was observed, reaching 16-fold higher levels until root emergence (Fig. 5e) which was associated with a similar rise in ATP level up to fourfold (Fig. 5h).

An overall measurement of the level of intermediates involved in citric acid cycle revealed that the content of all metabolites detected increased starting between 24 hpe and 48 hpe. For citrate, a strong continuous increase up to 125-fold higher levels was observed, which was most pronounced between 72 hpe and 96 hpe (Fig. 5j). There was also a tenfold and an eightfold increase in α-ketoglutarate and isocitrate concentrations, respectively (Fig. 5i,k). The concentrations of malate elevated nearly fourfold after 144 hpe and decreased to the half before roots emerged (Fig. 5l).

The basic concentration of total amino acids was strongly depleted between 6 hpe and 48 hpe (Fig. 6a). Thereafter, a recovery was observed reaching c. 50 % of the initial value at 192 hpe. The single amino acids followed this general pattern with glutamine and asparagine (Fig. 6c,e) as the predominant amino acids, followed by glutamate, and aspartate (Fig. 6b,d). Total protein level increased steadily to fourfold levels during the whole time-course of ARF (Fig. 6f).

Figure 6.

Content of (a) total amino acids, (b) glutamic acid, (c) glutamine, (d) aspartic acid, (e) asparagine and (f) total protein during adventitious root formation (ARF) in petunia (Petunia hybrida). Each value is represented by the mean of five independent replicates ± SE.

To evaluate whether the accumulation of carbon-containing metabolites was reflected by a change of carbon to nitrogen ratio during ARF, total contents of carbon and nitrogen were measured. This showed that parallel to sugar accumulation, an increase of carbon to nitrogen ratio up to threefold was monitored, beginning 48 hpe (Fig. S3).


A cutting removed from the donor plant normally undergoes various anatomical changes accompanied by changes in metabolic activity and gene expression during the wound response and subsequent rhizogenesis. To evaluate the occurrence and timing of specific anatomical, molecular and biochemical changes during ARF, different developmental stages were investigated using the model P. hybrida.

Anatomical changes during ARF

The root formation phases in petunia were designated as root initiation phase, followed by the root primordium formation phase and the root elongation and/or emergence phase. After severance of the cuttings from the stock plant, it took 72 h to initiate the earliest anatomical event that was unambiguously related to ARF (Fig. 1c). This root initiation phase resulted in the occurrence of small cells with a large nucleus and a dense cytoplasm – both typical signs for meristematic cells. The appearance of such meristemoids marks the transition from the root initiation phase to the root primordium formation phase. This first cytological sign of ARF at 72 hpe was preceded by the RNA accumulation of the cyclin B1 gene at 48 hpe (Fig. S1). Because CycB1 is not expressed in roots (Porceddu et al., 1999), it might serve as a marker gene specific for the root initiation phase of ARF.

During the root primordium formation phase the first well-developed young root meristems became visible at 96 hpe (Fig. 1d). These, at first globular structures, developed into root primordia with the typical dome shape (De Klerk et al., 1999) at 144 hpe (Fig. 1e), which included the meristem and behind it the first cells of the root body. The first roots with elongated cells of the elongation zone appeared at 192 hpe (Fig. 1f). These structures were still inside the stem, but revealed all morphological characteristics of a complete root, except for root hairs. They mark the transition to the root elongation phase, which resulted in the emergence of the earliest roots visible after one additional day.

Excision results in wound response as one of the earliest detectable reaction

The activation of wound-induced responses involves a complex network of signaling cascades, in which jasmonates represent the best-characterized class of signal molecules (Wasternack, 2007). In the work presented here, wounding occurred at the stem base of cuttings during their excision from the donor plant. Because of this process, JA and its precursor OPDA accumulated transiently with a maximum at 0.5 hpe (Fig. 2a,b). This occurred locally within the stem base as well as systemically but at lower levels in other parts of the cutting (data not shown). Further analyses suggest that the protein constitutively present in vascular tissues has constantly high activity and is involved in the increased biosynthesis of JA as already shown for tomato (Hause et al., 2003; Stenzel et al., 2003).

Usually, the wound-induced elevation of jasmonate levels is followed by activation of JA-responsive genes (Wasternack & Hause, 2002). Among those coding for JA biosynthetic enzymes, PhAOC quickly followed the rise in OPDA and JA, exhibiting a transient maximum of RNA accumulation at 1 hpe and 2 hpe (Fig. 2c). In addition to PhAOC, cell wall invertase transcript accumulation seems also to be induced by wound-induced jasmonates in stem base, as also described for other cell wall invertases (Fig. 3) (Godt & Roitsch, 1997; Schaarschmidt et al., 2006). Higher levels of transcripts followed by increased activity of cell wall invertase might then lead to a higher sink status of stem tissue (see below).

Role of carbohydrate metabolism in ARF

Cell division and cell enlargement during ARF require high input of energy and carbon skeletons. A major carbon source is sucrose, which is formed in photosynthetically active tissues and translocated towards sink parts of the plant. Sucrose can be used after cleavage into hexoses as a direct carbon source or is converted to storage compounds such as starch. Moreover, there are increasing indications that sugars have a regulatory role in ARF (Takahashi et al., 2003; Corrêa et al., 2005; Gibson, 2005).

In the stem base of petunia cuttings, the levels of soluble and insoluble sugars start to accumulate at 24 hpe despite high metabolic activity from 12 to 24 hpe onwards. However, the most pronounced increase in sugar levels was found during the later stages of ARF (Fig. 5a–c). A continuous increase in sucrose in source leaves can be observed from 144 hpe onwards, probably owing to an increased photosynthetic capacity (data not shown). This sucrose can be translocated towards the stem base for further metabolism. Even though a regulatory action of sugars cannot yet be excluded, the data support the suggestion that carbohydrates in petunia cuttings may play an important role in root growth rather than in root initiation.

The basipetal translocated sucrose is, however, not only used to deliver energy for differentiation (cell division and cell enlargement). A considerable portion of sucrose is converted to starch, which probably acts as the major carbon source when the adventitious roots grow. As shown for Pinus radiata, sucrose applied to the growing medium leads to higher levels of sugars and starch in rooting regions of hypocotyl cuttings and enhances root formation (Li & Leung, 2000). Considering the only marginal increase of starch at early stages, starch seems not to be involved in root initiation. Interestingly, the present analyses indicate that starch can be synthesized and stored in different cell types to meet their demand of increased metabolic activity when the adventitious roots emerge.

In agreement with the low sucrose content directly after excision, cell wall invertase activity was increased in the first hours after excising and decreased again to the basal level before root formation (Fig. 4a). By contrast, activities of vacuolar and cytosolic invertases decreased continuously to a low level (Fig. 4b,c). Cell wall invertase is not only a key enzyme of the apoplastic phloem unloading of transported sucrose but also links phytohormone action with primary metabolism (Roitsch & González, 2004). Being present in the apoplast, it can establish a sink function of a certain tissue and thus provide a mechanism for flexible and appropriate adjustment to a wide range of internal and external stimuli (Roitsch et al., 2003). Interestingly, a parallel accumulation of transcripts coding for the sugar transporter STP4 (Fig. 3), which is located in the apoplast (Truernit et al., 1996), has been detected. Although enhanced transcript accumulation does not always result in enhanced protein levels, this result suggests that at early stages of ARF basipetal translocated sucrose is degraded via cell wall invertase and the hexoses produced are transported into the cytosol via STP4 for further metabolism. As wounding causes not only induction of cell wall invertase, but also a rapid increase of STP4-driven GUS activity (Truernit et al., 1996), it is no surprise that both genes are induced at early stages of ARF to meet the increased carbohydrate demand within the cells.

Successive increase of both, PFK and Glc6P DH activity (Fig. 4e,i), and the decrease of cytosolic FBPase activity (Fig. 4h) strongly suggest that catabolism of glucose occurs in parallel in the pentose phosphate pathway and by glycolysis. This yields ATP as energy source and amino acids for protein synthesis. Because the activities of cytosolic GAPDH and pyruvate kinase did not change significantly (Fig. 4f,j), the first enzymes of the corresponding pathways seem to control the metabolic activity through which the hexoses from the basipetal transported sucrose can be catabolized.

Tricarboxylic acid cycle and energy metabolism contribute to the maintenance of ARF

Among the organic acids, citrate increased dramatically whereas all others remained unchanged and at low levels. Accumulation and exudation of citrate from roots have been shown to be induced by several stress factors, such as presence of aluminum (Yang et al., 2006) or phosphorous deficiency (Zhu et al., 2005). During ARF, however, citrate accumulated within the tissue and was probably stored within the vacuole. The question now arises why citrate content increased up to 125-fold, whereas the level of other intermediates of citric acid cycle increased at the most fourfold. Here, a model created for developing Brassica napus embryos could hold true: In the presence of inorganic nitrogen and glucose the refill of α-ketoglutarate is reduced to supply glutamine, glutamate, proline and arginine for protein synthesis (Junker et al., 2007). This adjustment is regulated by a balance of the activity of mitochondrial malic enzyme and PEP carboxylase. The surplus of citrate is transported into the cytosol where it may be transported into the vacuole. Our data strongly suggest that accumulated citrate allows a flexible adaptation of flux in response to ARF. Considering the dramatic increase concomitant with the formation of meristems, citrate level may be used as biochemical marker for entering later stages of ARF in petunia.

Predominant amino acids in cutting base are glutamine, glutamate, asparagine and aspartate

In most plant systems analysed, the most abundant amino acids in vascular tissues are glutamate, glutamine and in some cases asparagine (Urquhart & Joy, 1982; Schobert & Komor, 1989; Lohaus & Moellers, 2000). Analysis of the petunia stems before excision indicates that mainly glutamine and asparagine are synthesized in source tissues and transported downwards to sink tissues. Since glutamine biosynthesis is the key step of nitrate assimilation in plants (Joy, 1988) and supplies nitrogen for amino acid synthesis, a basipetal translocation of glutamine is an important process to meet the demand of the cells during root formation. Levels of glutamine and asparagine remained low at the later stages (Fig. 6c,e). This indicates a high turnover of the translocated amino acids into others that appear to be essential for the accelerated protein synthesis during ARF.

ARF in petunia cuttings is characterized by a transition of apoplastic to symplastic unloading of sucrose

Based on the results a three-phase mechanism is postulated for the metabolic response involved in ARF in petunia (Fig. 7). (1) ‘Sink establishment phase’: wounding leads via JA accumulation to an induction of genes coding for enzymes that degrade sucrose within the apoplast to hexoses. The hexoses are taken up by at least one induced monosaccharide transporter and are then used for production of energy necessary for wound healing and cell division. Therefore, wounding initiates the establishment of a sink tissue in which all the resources are depleted. (2) ‘Recovery phase’, which is characterized by the replenishment of resources and lasts up to 72 hpe ending with the formation of meristemoids. (3) ‘Maintenance phase’ characterized by symplastic transport of sugars translocated from source leaves into the stem base. They are converted either to intermediates directly flowing into root development or to the intermediate storage compounds starch and citrate thereinafter used for later root formation processes.

Figure 7.

Schematic presentation of the metabolic mechanisms involved in adventitious root formation (ARF) in petunia (Petunia hybrida). Assimilates are produced in source tissues and translocated towards the cutting base to establish a sink organ. Wounding induces among others the accumulations of transcripts coding for cell wall invertase leading to cleavage of sucrose by this enzyme and transport of the hexoses into the cell by the monosaccharide transporter STP. Two days after excision the recovery phase starts followed by the maintenance phase characterized by symplastic transport of sugars translocated from source leaves into the stem base and used either for energy production or accumulated in the vacuole.

Taken all together, ARF seems to be modulated by a combination of different pathways, including hormone biosynthesis and primary metabolism whose individual steps might exert a control over ARF. To understand whether there is a causal interaction between hormones and specific genes and how these interactions regulate the process of ARF will be a challenging step in the future. A possible approach would be the use of reverse genetics to unravel the importance of individual genes/proteins during ARF. In addition, taking into account that not all processes occurring during ARF are causally linked to the developing process, further experiments are needed to separate such responses from ARF. These could include, for example, applications of auxins or compounds inhibiting basal auxin transport to modulate the extent of root formation without affecting noncausal processes.


We thank Heike Nierig and Barbara Weinlich for the excellent technical assistance, Sabine Czekalla and Dr Siegfried Zerche for their help to establish the donor plants and the ex vitro rooting system and Andrea Knospe for taking care of tissue culture plants. We are indebted to Enk Geyer and his colleagues for preparation and taking care of greenhouse plants. In addition, we express many thanks to Prof. Norbert Sauer (University of Erlangen-Nuernberg) for his suggestions to design STP4 fragment and for fruitful discussion. We are also very thankful to Dr Björn Junker for his intensive discussions and meaningful suggestions. Furthermore, we thank Christian Klukas for his intensive assistance in preparing the figures created by the program vanted. This work was funded by the Pakt für Forschung und Innovation of the Leibniz-Gemeinschaft, Germany, supported by the States of Saxony-Anhalt and Brandenburg, the Free State of Thuringia and the Federal Republic of Germany.