Ethylene stimulates tracheary element differentiation in Zinnia elegans cell cultures


Author for correspondence:
Hannele Tuominen
Tel: +46 (0)90 786 9693


  • The exact role of ethylene in xylogenesis remains unclear, but the Zinnia elegans cell culture system provides an excellent model with which to study its role during the differentiation of tracheary elements (TEs) in vitro.
  • Here, we analysed ethylene homeostasis and function during Z. elegans TE differentiation using biochemical, molecular and pharmacological methods.
  • Ethylene evolution was confined to specific stages of TE differentiation. It was found to peak at the time of TE maturation and to correlate with the activity of the ethylene biosynthetic 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase. The ethylene precursor ACC was exported and accumulated to high concentrations in the extracellular medium, which also displayed a high capacity to convert ACC into ethylene. The effects of adding inhibitors of the ethylene biosynthetic ACC synthase and ACC oxidase enzymes to the TE cultures demonstrated for the first time strict dependence of TE differentiation on ethylene biosynthesis and a stimulatory effect of ethylene on the rate of TE differentiation.
  • In a whole-plant context, our results suggest that ethylene synthesis occurs in the apoplast of the xylem elements and that ethylene participates, in a paracrine manner, in the control of the cambial stem cell pool size during secondary xylem formation.


The water-conducting elements of the xylem, commonly called tracheary elements (TEs), are responsible for distribution of raw sap throughout the plant body. TEs are formed from the meristematic cells of the (pro)cambium according to a well-defined differentiation program that sequentially includes the following phases: cell expansion, secondary cell wall formation, programmed cell death, autolysis of the cellular contents, and terminal perforation of the end wall (Turner et al., 2007). This results in the formation of dead, hollow cells with an accessible lumen and a reinforced cell wall, which are ideal for raw sap conduction. In vitro plant cell cultures can be very useful for studying TE differentiation because the process can be readily induced in them and specific features of particular stages of the process can then be explored with little interference or potentially confounding signals from other cell types. Cell cultures are also amenable to pharmacological studies. Among such cell cultures, the Zinnia elegans TE differentiating system is the most widely used. In this system, mesophyll cells are mechanically isolated from the first pair of leaves of Z. elegans plantlets and cultivated in a specific medium supplemented with auxin and cytokinin (Fukuda & Komamine, 1980). The sequential differentiation events in this system include dedifferentiation of the mesophyll cells into cambial-like cells, followed by acquirement of the xylogenic potential and finally the formation of TEs (Fukuda, 1997; Demura et al., 2002). This system has many advantages, including high efficiency (c. 30% of all cells differentiate into TEs) and sequential and semi-synchronous progress of TE differentiation normally within 3 d after the initial hormonal stimulus. Differentiation of TEs in the Z. elegans system is analogous to TE differentiation in planta with regard to cell morphology, progression of TE differentiation, and expression of marker genes for cambial cell identity and the different stages of TE differentiation (Demura & Fukuda, 1993; Ye & Varner, 1994; Pesquet et al., 2003, 2004).

Since the 1980s, the Z. elegans system has provided valuable information about the diverse morphological, biochemical and molecular events that occur during TE formation, including the hormonal regulation of TE differentiation. Classical hormones such as auxins, cytokinins (Fukuda & Komamine, 1980), brassinosteroids (Yamamoto et al., 1997) and gibberellins (Tokunaga et al., 2006), nitric oxide (Gabaldon et al., 2005), and peptidic signals including tracheary element differentiation inhibitory factor (TDIF; Ito et al., 2006), xylogen (Motose et al., 2004) and phytosulfokine (Matsubayashi et al., 1999), have all been shown to participate in different phases of TE formation in the Z. elegans system. However, the roles of other growth regulators, such as ethylene, have not been previously elucidated. Ethylene is a gaseous plant growth regulator that has been implicated in fruit ripening, senescence, flowering, gravitropism, shoot apical meristem activity and a wide variety of processes involving some form of external stimulus, such as abiotic, mechanical or biotic stress (Abeles et al., 1992; Ruonala et al., 2006; Tsuchisaka et al., 2009; Davies, 2010). Ethylene is synthesized from S-adenosyl methionine via 1-aminocyclopropane-1-carboxylic acid in reactions catalyzed by 1-aminocyclopropane-1-carboxylic acid (ACC) synthase (ACS) and ACC oxidase (ACO), respectively. Both of these enzymes are encoded by genes belonging to multigene families, whose expression is highly dependent on the tissue type, developmental stage, and both abiotic and biotic signals (Lin et al., 2009). Although the biosynthetic pathway is well known and the regulation of this pathway has been extensively examined, the exact location of ethylene biosynthesis and its physiological significance are still unknown.

Ethylene has been implicated in the control of xylem formation since the discovery that ethylene production increases during active phases of wood formation in conifer trees (reviewed in Little & Pharis, 1995). In support of this hypothesis, expression of ethylene biosynthetic ACS- and ACO-encoding genes (Andersson-Gunnerås et al., 2003; Tsuchisaka & Theologis, 2004; Pesquet et al., 2005; Love et al., 2009), the abundance of the corresponding proteins and enzyme activities (Plomion et al., 2000; Andersson-Gunnerås et al., 2003) all increase during active phases of xylem formation. Exogenous ACC or ethylene also reportedly stimulates xylem formation in stems or cuttings of both conifers and angiosperm trees (Little & Pharis, 1995; Eklund & Little, 1998; Eklund & Tiltu, 1999; Love et al., 2009). However, ethylene evolution does not always reflect cambial activity, and conflicting results have been found following modifications of the ethylene balance by exogenous treatments (Yamamoto & Kozlowski, 1987; Eklund & Little, 1995; Little & Eklund, 1999). A preferred approach is to utilize mutants or transgenic lines altered in ethylene biosynthesis and/or signaling. Indeed, a clear role for ethylene was demonstrated in the stimulation of asymmetric xylem expansion in response to mechanical stimulus of the stem using transgenic, ethylene-insensitive Populus tremula × tremuloides trees (Love et al., 2009). Ethylene-insensitive Arabidopsis mutants are quite normal in appearance, suggesting that ethylene does not play an important role during normal growth of the plant. This view, however, was challenged in a recent paper where suppression of ethylene biosynthesis was shown to result in severe alterations in plant growth and vitality (Tsuchisaka et al., 2009), and it seems therefore possible that the ethylene-insensitive mutants appear normal because of incomplete suppression of ethylene signaling or the presence of parallel ethylene signaling pathways (Stepanova & Alonso, 2009).

It is possible that much of the controversy regarding the role of ethylene in xylem formation is a result of the complexity of whole-plant systems and the pleiotropic effects of ethylene modification. We therefore aimed to elucidate the role of ethylene in the simple and well-defined Z. elegans cell culture system. Our approach consisted of: analysis of ethylene synthesis from gene expression to protein activities; examination of the effects of pharmacological modulation of ethylene synthesis in specific phases of TE formation; and identification of ethylene biosynthetic and receptor genes expressed in differentiating Z. elegans TEs. For the first time we have been able to demonstrate strict dependence of TE differentiation on ethylene biosynthesis and a stimulatory effect of ethylene on the rate of TE differentiation in vitro. On the basis of these results, we propose a model for the function of ethylene in xylem formation in planta.

Materials and Methods

Plant material and xylogenic cell cultures

Mesophyll cells were isolated from the first pair of leaves of 14-d-old Zinnia elegans L. cv Envy (Hem Zaden BV, Venhuizen, Holland) seedlings, to generate xylogenic cell suspension cultures, by the method of Fukuda & Komamine (1980). The cells were cultured at 25°C and agitated at 120 rpm in a volume of either 12 ml in 50-ml Erlenmeyer flasks or 1 ml in 12-well plates. The initial cell density was c. 2 × 105 cells ml−1. The cells were cultured in an induction medium (CA+C) containing 0.1 mg l−1 of the auxin alpha-naphthylacetic acid and 0.2 mg l−1 of the cytokinin benzyladenine (Sigma-Aldrich). Cells were also cultured in control media lacking auxin (CC), or cytokinin (CA), or both hormones (C0). In the following experiments, 200–400 cells were analysed at each time-point in each replicate experiment.

Pharmacological treatments

Pharmacological treatments of the cell cultures included addition of the following inhibitors of ethylene production: the ACS inhibitor aminoethoxyvinylglycine (AVG; at 0–12 μM); the ACO inhibitor cobalt chloride (CoCl2; at 0–900 μM); and the ACO inhibitor 1-aminoisobutyric acid (AIB; at 5–15 μM). In addition, other cultures were treated with: the ethylene biosynthesis precursor ACC (at 0–200 μM) and the ethylene-releasing compound 2-chloroethyl-phosphonic acid (ethephon; at 0–200 μM). As ethephon decomposition releases ethylene and phosphoric acid, further control experiments were conducted to verify that equivalent amounts of phosphoric acid had no effect on TE differentiation (data not shown). The pharmacological substances were added to sets of cell cultures when they were initiated. AVG was also added to other sets of cultures at successive 12-h intervals (i.e. at 12 h after initiation to one set, and at 24 h after initiation to another set, etc.). Similarly, sets of initially AVG-treated cell cultures were washed every 12 h with fresh medium. The H+/ATPase inhibitor sodium orthovanadate (1 μM) was added to 60-h-old induced cell cultures, and the culture medium was collected by centrifugation at 150 g after 0, 1, 2 and 4 h. All chemicals were purchased from Sigma-Aldrich (Sweden) and added from aqueous stock solutions. Each treatment was replicated at least three times, and gave similar results.

Secondary cell wall and viability staining of TE cultures

To simultaneously stain viable cells, cellulose and lignin in secondary cell walls of developing TEs, 100-μl samples of cell culture were first stained with 10 μl of 0.01% calcofluor (Fluorescent brightener 28; Sigma-Aldrich), and then with 2 μl of 0.5% fluorescein diacetate (FDA) (Sigma-Aldrich) in acetone. Cells were mounted on glass slides and observed using a direct Axioplan2 microscope (Zeiss, Oberkochen, Germany) with bright field optics or epifluorescence illumination. FDA staining was observed in the blue excitation range, (excitation filter 450–490 nm; emission filter 515 nm), calcofluor in the UV excitation range (excitation filter 270–380 nm; emission filter 410–580 nm) and lignin autofluorescence in the green excitation range (excitation filter 546 nm; emission filter 590 nm). Images were taken with an Axiocam camera (Zeiss).

Ethylene measurement

Ethylene production in cell cultures was quantified by gas chromatography, as follows. Portions (2 ml) of each culture to be evaluated were incubated in 4-ml air-tight vials for 6 h, under otherwise unchanged conditions. The time of 6 h was selected as the incubation time as differentiation of the individual TEs varied temporally within a 6-h timeframe. One ml of the headspace gas was collected from the vials with a plastic hypodermic syringe and injected into a GC-8A gas chromatograph (Shimadzu, Kyoto, Japan) equipped with a Porapak T column coupled to a flame ionization detector (FID), and analyzed with the following settings: carrier gas nitrogen (flow rate 300 ml min−1), column oven temperature 90°C, detector temperature 220°C. Ethylene was identified based on its retention time and quantified by comparing FID signals from the samples with a standard curve obtained by injecting standards with known concentrations of C2H4 (purity 99.8%). All assays were conducted in triplicate and ethylene production was expressed as nmol ethylene produced per cell number h−1. The analyses were replicated in at least three independent experiments, all of which gave similar results.

ACC measurement

The concentrations of free ACC (intra- and extracellular) and total intracellular ACC released by acid hydrolysis (2 N HCl for 3 h at 120°C) were determined according to the method described by Lizada & Yang (1979). Samples were placed in 4-ml air-tight vials and adjusted to a total volume of 2 ml with ultra-pure water. ACC was converted to ethylene by the addition of 200 μl of 20 mM HgCl2, followed by 200 μl of a 1 : 1 mixture of saturated NaOH : bleach (Klorin; Colgate-Palmolive, Danderyd, Sweden). The vials were immediately capped after the addition of the NaOH : bleach, vortex-mixed and incubated on ice for 10 min. One ml of headspace was removed with a syringe and ethylene was measured as already described. The efficiency of ACC chemical conversion was determined to be 92% after conversion of known amounts of ACC into ethylene. All assays were repeated in triplicate. ACC production was expressed as nmol ACC μg−1 protein for the intracellular amounts and as nmol ACC μg per cell number for the extracellular amounts.

Protein extraction and in vitro ACS and ACO activity measurement

Cell suspensions were harvested by centrifugation for 5 min at 150 g, the cell medium was removed with a Pasteur pipette and frozen in liquid nitrogen, and the cells were homogenized by adding two to three tungsten carbide beads (Qiagen, Solna, Sweden) and vortex-mixing for 20 s in the interval between one freeze–thaw cycle and the next (with three freeze–thaw cycles in all). Proteins were then extracted by suspending the homogenate in 100 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 4 mM dithiothreitol, 2% polyvinylpolypyrrolidone and 10% glycerol adjusted to pH 8.0. After centrifugation at 10 000 g for 15 min at 4°C, the supernatant was removed and applied to a Pierce desalting spin column (Thermo Fisher Scientific Inc., Rockford, IL, USA). The proteins were eluted and quantified using the Bio-Rad protein assay (Bio-Rad, Sundbyberg, Sweden), following the manufacturer’s instructions. Portions of the eluate were used in the ACS and ACO assays, as described below.

To measure ACS activity, extracts containing 50 μg of protein were incubated at 37°C for 2 h in a 2-ml total volume of 50 mM HEPES, 200 μM S-adenosyl-L-methionine and 2.5 mM pyridoxal-5′-phosphate (PLP) adjusted to pH 8.0, in 4-ml air-tight vials agitated on a rotating table. The ACC that was produced was then measured as described above. The results obtained with triplicate preparations for each sample were compared with those obtained for controls, to which no S-adenosyl-L-methionine had been added. The activity was expressed as nmol ACC μg−1 protein h−1.

To measure ACO activity, extracts containing 50 μg of protein or an equivalent volume of cell medium were incubated at 37°C for 2 h in 2-ml total volumes of 50 mM HEPES, 0.1 mM FeSO4, 20 mM NaHCO3, 30 mM ascorbate and 0.1 mM ACC adjusted to pH 8.0 in 4-ml air-tight vials agitated on a rotating table. The ethylene produced was then measured as already described. The results obtained with triplicate preparations for each sample were compared with those obtained for controls, to which no ACC had been added. The activity was expressed as nmol ethylene produced μg−1 protein h−1.


Total RNA was prepared from cell cultures after separating the cells from the culture medium by centrifugation for 5 min at 150 g. Culture medium was removed with a Pasteur pipette and frozen in liquid nitrogen. The cells were homogenized by adding two to three tungsten carbide beads (Qiagen) and vortex-mixing for 20 s in the interval between one freeze–thaw cycle and the next (with three freeze–thaw cycles in all). One ml of TRIreagent solution (MRC, Cincinnati, OH, USA) was then added to the homogenized cells, and total RNA was isolated according to the manufacturer’s instructions and subjected to DNA digestion with 5 U RNase-free DNase I (Promega, Nacka, Sweden) for 1 h at 37°C. A second round of RNA extraction was carried out as described above and the extracts were pooled, and then RNA was quantified using a Nanodrop spectrophotometer (Eppendorf, Hamburg, Germany), and visualized after electrophoresis on 1.5% agarose gels. The lack of DNA contamination was confirmed by PCR amplification of the RNA preparations with 18S rRNA PCR primers at 55°C annealing temperature for 40 cycles.

cDNA was made using the 18-nt dT primer and 1 μg of RNA as a template using superscript II reverse transcriptase (Invitrogen). PCR reactions consisted of the following steps: initial denaturation at 95°C for 10 min, followed by 35 cycles of denaturation at 95°C for 30 s, primer annealing at 60°C for 30 s, and extension at 72°C for 45 s. The PCR primer combinations for each gene were as follows: ZeACO1a (GenBank accession CAT03225), forward GGTTAACTTGGGTGATCAGCTG and reverse TTGTTCTTTCTCTTTTGGCTCT; ZeACO1b (CAT03226), forward GGTTAACTTGGGTGATCAGCTG and reverse CACATTATCCTGCTGTTCCTTG; ER362 (AU285412), forward GAGCAAGTCAAGCTAGAAGC and reverse AGTCATCCATTACGTC; ER3654 (AU289181), forward GGGTGTTTCAGAGGACTAGC and reverse AAACTGCAGCAGTCAAGACC; ER6734 (AU292016), forward CAACCGGGAGCTGTGGCTGC and reverse AGCGCTGATGCCCAATCGT; ER12021 (AU303262), forward CGCGGGGCGCAATGCTATT and reverse TCATCTCAGGGGTCCCACCT; ER14856 (AU305526), forward CACTATTCTAGTCGAGCTTT and reverse CTTGACTCGCCATCATGGCG; TRACHEARY ELEMENT DIFFERENTIATION 4 (TED4) (D30802), forward AAAGGCGTATGTTTCGTCTC and reverse ACTGCAAATTTATTCTAGCA; and SERINE PROTEASE (ZeSP) (DV017578), forward CTGGGTTGGTTGTGAAGGTT and reverse TGAGACCTAATTGACAGTAT. All primers were synthesized by Invitrogen. The 18S rRNA gene (AB089282) was amplified, as a reference gene, using a combination of the reverse primer TGTCACTACCTCCCCGTGTC and the forward primer TGCTACTCGGATAACCGTAG.


Ethylene biosynthesis during TE differentiation

To assess the significance of ethylene production during TE differentiation, we quantified ethylene, its precursor ACC, and corresponding biosynthetic enzyme activities (ACS and ACO) during the course of Z. elegans TE differentiation in both TE-inductive medium (CA+C; supplied with both auxin and cytokinin) and noninductive medium (CA, CC and C0, representing media with only auxin, only cytokinin and without either auxin or cytokinin, respectively). In inductive conditions, the differentiation process can be divided into three stages according to Fukuda (1997). In stage I, during the first day of culture, the mesophyll cells dedifferentiate after hormonal induction; they then obtain xylogenic potential during stage II and undergo TE differentiation during stage III. In our conditions, the secondary cell wall thickenings, defining the start of stage III, were first apparent 48 h after hormonal induction, as detected by cell wall birefringence, cellulose staining and lignin autofluorescence of the differentiating TEs (Fig. 1a–l). FDA staining of the cells revealed that after 96 h almost all TEs were dead (Fig. 1m–q). The three stages of TE differentiation were confirmed by monitoring expression of the stage II-specific marker gene TED4 (Demura & Fukuda, 1993) and the stage III-specific marker gene ZeSP (Pesquet et al., 2005) (Fig. 2e).

Figure 1.

 Cytological characterization of in vitro Zinnia elegans tracheary element (TE) differentiation. (a–d) Overall TE morphology. (e–h) Cellulosic secondary walls after staining with calcofluor. (i–l) Lignified secondary walls visualized using lignin autofluorescence. (m–p) Cell viability after staining with fluorescein diacetate. The cells were analysed 48 h (a, e, i, m), 60 h (b, f, j, n), 72 h (c, g, k, o) and 96 h (d, h, l, p) after TE differentiation induction by hormonal stimuli, and visualized using bright field microscopy (a–d) or fluorescence microscopy (e–p). Bars, 8 μm. (q) Rate of TE differentiation (% of the total number of cells; black line) and a typical relationship between the presence of living (nonlignified; green bars) and lignified (red bars) TEs in differentiating TE cell cultures as a function of time after TE induction. The error bars indicate ± SD.

Figure 2.

 Ethylene biosynthesis during in vitro Zinnia elegans tracheary element (TE) differentiation. (a) Ethylene evolution during the cell culture time course. (b) Free and conjugated 1-aminocyclopropane-1-carboxylic acid (ACC) concentrations during the cell culture time course. (c) ACC synthase activity. (d) ACC oxidase activity. The cells were cultured in TE-inductive medium with auxin and cytokinin (CA+C) or noninductive control media with auxin only (CA), cytokinin only (CC) or without auxin or cytokinin (C0). (e) RT-PCR expression analysis of Z. elegans ACC OXIDASES 1a and 1b, TRACHEARY ELEMENT DIFFERENTIATION 4 (TED4), SERINE PROTEASE (ZeSP) and the 18S rRNA control in TE-inductive conditions during the cell culture time course. In our conditions, stage I, in which mesophyll cells lose their photosynthetic capacity, occurred between 0 and 24 h. Stage II, when cells acquire xylogenic potency, occurred between 24 and 48 h, and finally stage III, when the TEs mature, occurred after 48 h of culture. The error bars indicate ± SD.

The Z. elegans mesophyll cell cultures produced relatively low amounts of ethylene. No ethylene detectable by our instrumental set-up was produced by the cells in response to mechanical isolation or wounding (data not shown), or by cells cultured in hormone-free medium (Fig. 2a). However, ethylene evolution increased markedly in TE-inductive conditions, and lower concentrations of ethylene were also produced in non-TE-inductive conditions when the medium was supplemented with either auxin or cytokinin (Fig. 2a). Ethylene production progressed in the TE-inductive conditions in a biphasic manner; the first peak occurring during stage II after 36 h and the second (highest) peak during stage III after 60 h (Fig. 2a). The ethylene precursor, ACC, also accumulated during the course of TE differentiation. In the TE-inductive conditions, concentrations of free ACC began to increase steadily after 36 h, while concentrations of conjugated ACC increased only after 84 h. By contrast, in the auxin-only medium, there was no significant rise in ACC (Fig. 2b).

Associated enzymatic activities (ACS and ACO) were measured in vitro during the differentiation time course in the CA+C and CA cell cultures. ACS activity gradually increased during TE differentiation (Fig. 2c), while ACO activity oscillated during the differentiation time course in a similar manner to ethylene evolution (Fig. 2d). Similarly, the expression of the two ACOs (ZeACO1a and ZeACO1b) that were cloned from differentiating TE cultures showed an oscillating expression profile during TE formation (Fig. 2e). These two Z. elegans ACOs are most similar to the Arabidopsis gene At1g05010, which is commonly called the ethylene-forming enzyme (Supporting Information Fig. S1a). In particular, the expression of ZeACO1a coincided with, or rather preceded, the rise in ACO enzymatic activity, as well as ethylene accumulation, confirming the oscillatory pattern of ethylene biosynthesis during the TE differentiation time course. The only known Z. elegans ACS (DV017373) did not show any oscillatory expression pattern during the TE time course (data not shown). We also analyzed the expression pattern of five genes showing high homology to ethylene receptors (Fig. S1b), which were identified from the Z. elegans EST database ( Only one of these genes (EST-362; a homolog of Arabidopsis ethylene receptor ETR2) was expressed during TE differentiation (Fig. S1c), but not in an oscillatory manner. Both the gene expression and enzyme activity assays indicate that ACO correlates with the level of ethylene evolution, and hence regulates ethylene production in this system.

Interestingly, high concentrations of ACC were detected in the extracellular medium (the culture medium) under the TE-inductive conditions, 100–200 times the measured intracellular amounts, while no ACC was detected in the control (CA; auxin only) medium (Fig. 3a). Vanadate, which inhibits plasma membrane H+/ATPases and therefore active transport, significantly reduced accumulation of the extracellular ACC, which suggests that ACC is actively exported to the apoplast (Fig. 3b). Surprisingly, the in vitro ACO activity was also much higher in the extracellular medium than intracellularly (Fig. 3c). This suggests that the extracellular space is the main site of ethylene synthesis in the TEs. A smaller proportion of this ACO activity was loosely bound to the cell wall, as demonstrated by NaCl washing of the cells (Fig. 3c). Adding ACC did not increase the maximal ACO activity detected in the extracellular compartment (Fig. 3c). Therefore, our analyses of the Z. elegans system show that the extracellular space plays a key role in the ethylene biosynthesis of differentiating TEs, as both the precursor of ethylene, ACC, and the capacity to oxidase ACC largely reside in this compartment.

Figure 3.

 Localization of ethylene synthesis. (a) Extracellular 1-aminocyclopropane-1-carboxylic acid (ACC) concentrations in tracheary element (TE) inductive (CA+C; closed bars) and control (CA; open bars) conditions during the cell culture time course. (b) Extracellular ACC concentrations in 60-h-old induced cell cultures treated (open bars) or not treated (closed bars) with 1 μM vanadate (a H+/ATPase inhibitor used to inhibit active transport). (c) ACC oxidase activity in different cellular compartments of 60-h-old induced cell cultures after autocatalytic activation of ethylene biosynthesis by 50 μM ethephon treatment (closed bars, +ACC; open bars, −ACC). The assays were performed using samples of induced cell cultures containing both the cells and the extracellular medium (‘Total activity’), only the extracellular medium, and cells separated from the extracellular medium both before and after washes with 150 mM NaCl (‘Cells before NaCl’ and ‘Cells after NaCl’, respectively). Note that the inclusion of 100 μM ACC did not have any significant effect on the ACC oxidase activity, suggesting that the availability of ACC does not limit its activity. The vertical bars indicate ± SD.

Pharmacological inhibition of ethylene production during TE differentiation

To determine the function of ethylene during TE formation, the effects of adding ethylene biosynthesis inhibitors during in vitro TE differentiation were explored. Aminoethoxyvinylglycine (AVG), an ACS inhibitor, dose-dependently decreased TE differentiation efficiency, and completely inhibited TE formation at its highest concentration (Fig. 4a). This inhibition correlated with a reduction in ethylene evolution during the TE differentiation time course (Fig. 4b). In order to determine the timeframe and the target of ethylene action during TE differentiation, AVG was added at different time-points during the culture (every 12 h) (Fig. 4c). Conversely, washing experiments were also performed, in which AVG was added at 0 h and washed out of the cultures every 12 h to determine when its action began (Fig. 4d). These experiments confirmed the requirement for ethylene for cells to acquire xylogenic potential in phase II, as TE formation was blocked when AVG was added during the first 36 h, and restored if it was removed during phase I (Fig. 4c,d).

Figure 4.

 Modulation of 1-aminocyclopropane-1-carboxylic acid (ACC) synthase activity during Zinnia elegans tracheary element (TE) differentiation. (a) Effect of aminoethoxyvinylglycine (AVG) treatment on TE differentiation efficiency (expressed as the percentage of TEs of the total number of cells) and TE lignification (expressed as the percentage of lignified TEs of all TEs) after 120 h of culture. (b) Effect of AVG treatment on ethylene evolution during the differentiation time course when the treatment was performed at the initiation of the culture. Black bars, 0 μM AVG; gray bars, 3 μM AVG; white bars, 6 μM AVG. (c) Effects of AVG additions on TE differentiation efficiency and TE lignification when applied at 12-h intervals along the time course. (d) Effect of AVG removal on TE differentiation efficiency and TE lignification when cells that were initially treated with AVG were washed at 12-h intervals along the time course with fresh medium. The concentration of AVG was 12 μM in (c) and (d). The controls in (c) and (d) were not treated with AVG. The vertical bars indicate ± SD.

It was difficult to investigate any potential reversal of the effect of AVG on TE differentiation through the addition of ACC because exogenous ACC did not increase ethylene production in this system, but instead had the opposite effect (Fig. S2b) and only partially rescued ethylene evolution in AVG-treated cells to the levels observed in noninducing cell culture conditions (Fig. S2c). Exogenous ethylene could not be used for this purpose either, because of its gaseous nature and incompatibility with the liquid cell cultures. However, inhibitors of ACO activity, cobalt chloride and 1-aminoisobutyric acid, reduced the TE differentiation rate in a similar manner to AVG, except at the lowest tested concentrations, at which both of these inhibitors had a slight stimulatory effect on TE differentiation (Fig. 5a–d). Taken together, these results confirm that ethylene is a prerequisite for TE differentiation and that ethylene enhances the rate of TE differentiation. The slight increase in the rate of TE differentiation in response to a moderate decrease in ethylene biosynthesis (induced by ACO inhibitors) suggests that the level of ethylene evolution in the inductive TE conditions is too high for maximal TE differentiation.

Figure 5.

 Modulation of 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase activity during Zinnia elegans tracheary element (TE) differentiation. (a, c) Effect of exogenous cobalt chloride (CoCl2) (a) and 1-aminoisobutyric acid (AIB) (c) on TE differentiation efficiency (expressed as the percentage of TEs of the total number of cells) and TE lignification (expressed as the percentage of lignified TEs of all TEs) after 120 h of culture. (b, d) Effect of exogenous CoCl2 (b) and AIB (d) on ethylene production along the differentiation time course. The treatments were performed at the start of the culture. The vertical bars indicate ± SD.

Finally, in order to evaluate the effect of pharmacological modification of ethylene concentrations on cell division activity, we analyzed changes in cell density in response to exogenous AVG treatments in TE-inducing conditions. AVG at concentrations ranging from 4 to 12 μM did not have any consistent effect on cell density during the TE differentiation time course (Fig. S3), confirming that the action of ethylene is strictly associated with the TE differentiation process itself and not with cell division.


It has become clear that in order to increase our understanding of ethylene physiology it is crucial to define the exact location of ethylene synthesis and signaling not only at the tissue level but also at the cellular level (Stepanova et al., 2008; Love et al., 2009; Tsuchisaka et al., 2009). The Z. elegans in vitro system allowed us to define the role of ethylene in differentiation of tracheary elements (TEs) at a cellular level. We observed little ethylene evolution at the beginning of TE differentiation, but it increased along with TE differentiation, peaking at the stage of TE maturation (Fig. 2a). Also, the ethylene precursor ACC and the activities of the ethylene biosynthetic enzymes increased during the TE differentiation time course (Fig. 2), which corroborates the observed pattern of ethylene evolution in this system, although the absolute levels of ethylene evolution might be underestimated because of impaired diffusion of ethylene in the aqueous cell culture system. This kind of accumulation pattern is typical for a hormone that, in order to coordinate its involvement in multiple physiological processes, needs to be turned off until the right types of cues are encountered.

Our data suggest that ethylene biosynthesis was regulated by the expression of ACO (Fig. 2d,e) in a manner similar to that found in some earlier studies (English et al., 1995; Vriezen et al., 1999; Andersson-Gunnerås et al., 2003; Wagstaff et al., 2005), but different from that found in the large numbers of reports showing that the reaction catalyzed by ACS is the rate-limiting step. Notably, large quantities of ACC were detected in the extracellular medium of our cultures containing differentiating Z. elegans TEs (Fig. 3a). ACC has been identified previously in xylem sap both following exposure of plants to flooding or high salinity (Bradford & Yang, 1980; Else et al., 1995; Albacete et al., 2009) and in nonstressed conditions (Hall et al., 1993; Else & Jackson, 1998; Andersson-Gunnerås et al., 2003; Dodd et al., 2009). We also obtained evidence that the intracellularly synthesized ACC is actively transported to the extracellular medium (Fig. 3b). The nature of the transport mechanism remains unknown, but there are several proteins with characteristics of amino acid transporters that have no known function and are potential candidates for such transporters. Surprisingly, we also detected a high capacity of the extracellular medium to convert ACC into ethylene (Fig. 3c), suggesting that the ACO in differentiating xylem elements has an apoplastic location. The subcellular localizations of ACOs are not known, although several different locations, including the apoplast, have been suggested (Peck et al., 1992; Reinhardt et al., 1994; Ramassamy et al., 1998; Chung et al., 2002). It is also possible that different isoforms are targeted to different compartments (Bidonde et al., 1998). We cloned two Z. elegans ACO genes that were highly abundant in differentiating TEs and therefore are probably the major genes responsible for ethylene biosynthesis in this system (Fig. S1a). When we compared these ACO genes to the well-characterized family of tomato (Solanum lycopersicum) ACO genes, these Z. elegans genes were found to be most similar to the tomato pTOM13 ACO, which has been assigned an apoplastic localization by immunolocalization using antibodies raised against the pTOM13 peptide (Rombaldi et al., 1994). Therefore, our results strongly suggest that ACC can be converted into ethylene by an apoplastic ACO in differentiating xylem elements.

Using several different pharmacological inhibitors of ethylene biosynthesis, we demonstrated that ethylene was a prerequisite for TE differentiation (Figs 4, 5). AVG, a pharmacological inhibitor of ACS, inhibited both ethylene evolution and TE differentiation in a dose-dependent manner. This is in good agreement with the recent report of Tsuchisaka et al. (2009), which showed that suppression of ACC synthesis in an ACS null mutant was lethal to the plants and that several physiological processes were unaffected unless ethylene biosynthesis was almost completely abolished. The pharmacological inhibitors of ACO, cobalt chloride and 1-aminoisobutyric acid, did not completely inhibit TE differentiation in the Z. elegans system, probably because of the residual ethylene biosynthesis present after the inhibitor treatments (Fig. 5). The staggered AVG application and washing experiments demonstrated that ethylene mainly acts at the stage at which the cells redifferentiate into cambial-like cells and obtain xylogenic potential, somewhere between 12 and 24 h after the addition of the inductive hormonal stimulus (Fig. 4c,d). This finding suggests that ethylene stimulates TE differentiation by increasing the capacity of the cell cultures to maintain a cambial-like stem cell identity. Accordingly, ethylene has been recently shown to increase the size of the stem cell pool in the root apical meristem of young Arabidopsis seedlings (Ortega-Martinez et al., 2007), and the number of cambium cell layers in Populus alba trees (Junghans et al., 2004). Interestingly, pharmacological modification of ethylene homeostasis did not affect cell division activity (Fig. S3b). Therefore, it is likely that, in planta, ethylene is involved in the maintenance of a sufficiently large pool of stem cells in the cambial meristem in order to stimulate xylem formation in response to secondary growth inducing stimuli.

In planta, xylem precursor cells are derived from the cambium in a process that can be defined as a ‘continuous cambium’ (Fig. 6a). By contrast, the in vitro TE differentiation system, in which all cells progress sequentially towards TE differentiation, can be defined as a ‘discontinuous cambium’. Discontinuous cambium models are not directly comparable to a continuous cambium, but they can reveal characteristics of the individual stages of TE differentiation in a unique manner. We propose a model for ethylene action in TE initiation and differentiation in planta, which is based on our results from the Z. elegans discontinuous cambium model (Fig. 6b). Most importantly, ethylene is mainly produced in the maturing TEs, while its site of action is in the cambial region; this supports a paracrine mode of action for ethylene. The rate of ethylene evolution rapidly increases in the differentiating TEs as a function of their distance from the cambium. This fast ethylene production relies partly on a saturated environment of the ethylene precursor ACC, which is both actively exported to the apoplast and transported in the xylem sap, and partly on the extracellular activity of the ACO. Ethylene can easily diffuse from the maturing TEs all the way to the cambial zone, where it can participate in maintenance of the stem cell pool and thus control the rate of TE formation. We propose that this paracrine function of ethylene serves a signaling function between differentiating TEs and the cambial zone that modulates meristematic activity, based on the progress of TE maturation and signals from the maturing TEs. The extracellular export of ACC and the biosynthesis of ethylene in the apoplast of the maturing TEs, and the long-distance transport of ACC/ethylene in the xylem sap, provide an additional mechanism for the coordination of vascular development in the longitudinal direction (Fig. 6b). This kind of regulation may be critical in situations where cambial growth needs to be rapidly modified throughout the whole stem in response to certain stimuli, such as mechanical disturbance. In order to verify these predictions, which are based on the discontinuous cambium model, more detailed and stage-specific analyses of continuous cambium models than have been conducted to date are required. Multiple mutant combinationsin the xylem-expressed ACSs and ACOs and the multiple gain-of-function ethylene receptor mutants could provide such information. The data presented on the discontinuous cambium suggest interesting hypotheses to be tested in these types of mutant.

Figure 6.

 A model for ethylene action in tracheary element (TE) differentiation in planta. (a) A transverse section of a Zinnia elegans stem vascular bundle. In planta, xylem TE differentiation proceeds from a cambial cell [1] to a TE precursor [2] and finally to a mature TE [3]. All stages of differentiation are present in the same tissue, as opposed to the in vitro cell culture system where at any given time only TEs of the same differentiation stage are present. (b) A proposed model for the function of ethylene in xylogenesis in planta illustrated as a schematic image of a vascular bundle. Based on the results obtained in the Z. elegans in vitro system, ethylene evolution takes place mainly in the maturing tracheary elements [3]. We propose that, in planta, diffusion of ethylene (illustrated as dashed red lines) from the site of synthesis to the younger TE precursors [2] and to the cambial region [1], together with the transport of 1-aminocyclopropane-1-carboxylic acid (ACC) and ethylene in the xylem sap (illustrated as continuous red lines), coordinates cambial growth and vascular development both radially and axially in a paracrine fashion.


The authors would like to thank the Swedish Research Council Formas (through the Research Programme on Wood Material Science and the FuncFiber Centre of Excellence in Wood Science), the Carl Trygger Foundation and the Kempe Foundation for providing financial support for this study. The authors thank Björn Sundberg for fruitful discussions and his comments on the manuscript.