Plant Molecular Biology

, Volume 80, Issue 3, pp 255–272

Over-expression of a cytosolic isoform of the HbCuZnSOD gene in Hevea brasiliensis changes its response to a water deficit

Authors

  • J. Leclercq
    • CIRAD, UMR AGAP
  • F. Martin
    • CIRAD, UMR AGAP
  • C. Sanier
    • CIRAD, UMR AGAP
  • A. Clément-Vidal
    • CIRAD, UMR AGAP
  • D. Fabre
    • CIRAD, UMR AGAP
  • G. Oliver
    • CIRAD, UMR AGAP
  • L. Lardet
    • CIRAD, UMR AGAP
  • A. Ayar
    • CIRAD, UMR AGAP
  • M. Peyramard
    • CIRAD, UMR AGAP
    • CIRAD, UMR AGAP
Article

DOI: 10.1007/s11103-012-9942-x

Cite this article as:
Leclercq, J., Martin, F., Sanier, C. et al. Plant Mol Biol (2012) 80: 255. doi:10.1007/s11103-012-9942-x

Abstract

Hevea brasiliensis is the main commercial source of natural rubber. Reactive oxygen species (ROS) scavenging systems are involved in various biotic and abiotic stresses. Genetic engineering was undertaken to study the strengthening of plant defences by antioxidants. To that end, Hevea transgenic plant lines over-expressing a Hevea brasiliensis cytosolic HbCuZnSOD gene were successfully established and regenerated. Over-expression of the HbCuZnSOD gene was not clearly related to an increase in SOD activity in plant leaves. The impact of HbCuZnSOD gene over-expression in somatic embryogenesis and in plant development are presented and discussed. The water deficit tolerance of two HbCuZnSOD over-expressing lines was evaluated. The physiological parameters of transgenic plantlets subjected to a water deficit suggested that plants from line TS4T8An displayed lower stomatal conductance and a higher proline content. Over-expression of the HbCuZnSOD gene and activation of all ROS-scavenging enzymes also suggested that protection against ROS was more efficient in the TS4T8An transgenic line.

Keywords

Abiotic stressDehydrationDevelopmentDroughtHevea brasiliensisGenetic transformationOxidative stressPlant regenerationSomatic embryogenesis

Introduction

Reactive oxygen species (ROS) are generated by biotic and abiotic stresses such as drought, salinity, strong light, extreme temperatures, heavy metals, UV radiation, atmospheric contamination, mechanical wounding, nutrient starvation and pathogen attacks (Jaspers and Kangasjarvi 2010). The antioxidant metabolism protects cells from oxidative damage caused by ROS, such as peroxidation of membrane compounds, polysaccharide degradation, enzyme denaturation and DNA lesions (Scandalios 2005). Several enzymes act jointly to maintain redox status homeostasis. ROS-scavenging systems play an important role in cell functioning because ROS are extremely cytotoxic. Many antioxidant enzymes catalyse redox reactions, such as the ascorbate–glutathione cycle whose steps rely on electrons supplied by reducing compounds of low molecular weight, such as ascorbate and glutathione (for a review see Noctor and Foyer 1998). ROS-scavenging systems have complicated regulation involving feed-back control of enzymes (Noctor and Foyer 1998). Of the antioxidant enzymes, superoxide dismutase (SOD) activity constitutes the first line of defence against ROS by converting O2·− to H2O2. Depending on the metal co-factor used by the enzyme, SODs are classed in three groups: copper–zinc SOD (CuZnSOD), manganese SOD (MnSOD) and iron SOD (FeSOD). In addition, the sub-cellular locations of SODs are different: FeSODs are chloroplastic isoforms, MnSODs are mitochondrial or peroxisomal isoforms and CuZnSODs are cytosolic, chloroplastic or peroxisomal isoforms (Alscher et al. 2002). SODs are encoded by small multigene families. The Arabidopsis thaliana genome harbours three FeSOD, two MnSOD and three CuZnSOD genes (http://www.uniprot.org).

Preventing oxidative stress plays an important role in tolerance to abiotic stresses (Gill and Tuteja 2010; Mittler 2002). However, enhancing tolerance to oxidative stress depends on the nature and location of the imposed stress, the induced isoform and its level of expression, intracellular targeting, plant age and growth conditions (Noctor and Foyer 1998). In the literature, both successful and unsuccessful genetically engineered stress-resistant plants have been obtained (Prashanth et al. 2008; Wang et al. 2005; Gao et al. 2003; Faize et al. 2011; Tepperman and Dunsmuir 1990; Slooten et al. 1995; Pitcher and Zilinskas 1996). For instance, over-expression of the Tamarix androsowiMnSOD gene protects poplar plants from NaCl stress (Wang et al. 2010).

ROS are involved in the coagulation of rubber particles that dramatically reduces natural rubber production. Hevea brasiliensis, a tropical perennial species native to the Amazon basin, is the sole commercial source of natural rubber. In 2010, 92 % of the 10 million tons of natural rubber produced in the world came from South-East Asian countries. The worldwide increase in demand for natural rubber calls for rapid improvement of the rubber tree. Natural rubber, or more precisely cis-1,4 polyisoprene, is polymerized in the rubber particles of latex cells. Latex cells are specialized cells periodically emitted from the cambium and then anastomosed to form latificer mantels (De Faÿ and Jacob 1989). Latex is the cytoplasm of laticifers which contain 30–50 % of rubber particles. Latex is expelled from laticifers after tapping, through the effects of high turgor pressure maintained in the soft bark tissues. An ethylene generator, 2-chloroethylphosphonic acid (ethephon), is applied to the tapping panel in order to stimulate latex production by increasing both the duration of latex flow after tapping and its regeneration between two tappings (Lacote et al. 2010a). Over-tapping or excessive ethephon stimulation provokes in situ coagulation of rubber particles. High ROS production in latex cells triggers oxidative stress (Chrestin 1989; Chrestin et al. 1984). ROS damage cellular membranes and especially lutoids, which contain coagulant factors involved in the aggregation of rubber particles (Chrestin et al. 1984). In extreme cases, ROS induce senescence, which leads to Brown Bast in bark tissues. These two symptoms are called Tapping Panel Dryness (TPD).

Four Hevea brasiliensisSOD genes have been reported in the public database: two HbMnSOD genes (Miao and Gaynor 1993), and two HbCuZnSOD genes, of which one is cytosolic (GI:27449245) and the other a partial sequence of a chloroplastic (GI:186920322). Expression of the HbMnSOD genes has been found to be 50-fold higher in embryogenic callus than in leaves (Miao and Gaynor 1993), weak in the latex of untapped trees and slightly stimulated upon ethylene stimulation (Sookmark et al. 2005). Interestingly, the TPD-susceptible clone RRIM 600 showed a lower level of expression of the cytosolic HbCuZnSOD gene compared to clones PB 235 and GT1, clones that are less susceptible to TPD (Sookmark et al. 2005). In the case of chilling injury, the Hevea clone RRIM 600 has been described as being the most tolerant, with higher SOD activity and rapid stomatal closure upon cold stress (Mai et al. 2010). By contrast, no change in SOD activity was observed for clone PB260, a clone susceptible to chilling. Transgenic Hevea plants over-expressing the mitochondrial HbMnSOD gene have been regenerated, but further characterization under stress conditions has yet to be reported (Jayashree et al. 2003).

Given the importance of the location of ROS-scavenging systems for improving tolerance to oxidative stress by genetic engineering (Noctor and Foyer 1998), we suggest that the cytosolic isoform of the HbCuZnSOD gene would be the most appropriate for overcoming abiotic stress in Hevea, and in particular in latex cells. Based on the GFP-based transformation procedure (Leclercq et al. 2010), we successfully regenerated transgenic Hevea plants modified for the homologous HbCuZnSOD gene. The regenerated transgenic lines showed a range of HbCuZnSOD gene expression and enzymatic activity. Over-expression of the HbCuZnSOD gene was related to changes in somatic embryogenesis and plant development. In order to check the effect of this modification on abiotic stresses, biochemical and eco-physiological parameters were studied in response to a water deficit treatment for two regenerated fast-growing transgenic plantlet lines. One of them displayed greater osmoprotection with reduced stomatal conductance and better ROS protection, with higher antioxidant enzyme activity.

Materials and methods

Plant materials

The friable callus line CI05519 of Hevea clone PB260 was established from integument calli (Lardet et al. 2009). The line was subcultured every 2 weeks on a maintenance culture medium (MM) containing macro-elements (20 mM NH4NO3, 20 mM KNO3, 3 mM MgSO4.7H2O, 2 mM NaH2PO4.H2O, 9 mM CaCl2), micro-elements (150.08 μM H3BO3, 100 μM MnSO4. H2O, 40 μM ZnSO4.7H2O, 1.48 μM CuSO4.5H2O, 0.99 μM Na2MoO4.2H2O, 5 μM KI, 1.01 μM CoCl2.6H2O), vitamins (300 μM inositol, 20 μM nicotinic acid, 3 μM pyridoxine–HCl, 2 μM thiamine-HCl, 0.2 μM biotin, 1 μM D-calcium pantothenate, 1 μM ascorbic acid, 0.1 μM choline chloride, 60 μM l-cysteine-HCl, 5 μM glycine, 1 μM riboflavin), 1.35 μM benzylaminopurine (BAP), 1.35 μM 3,4-dichlrophenoxyacetic acid (3,4-D), 100 μM FeSO4, 100 μM Na2EDTA, 30 μM AgNO3, 234 mM sucrose, 0.5 μM abscisic acid (ABA) and 2.3 g L−1 Phytagel (Carron et al. 1995); (Lardet et al. 2007). The pH of the whole medium was adjusted to 5.8 prior to autoclaving. Callus cultures were grown in the dark at 27 °C. Prior to Agrobacterium inoculation, the callus line was precultured for 15 days in glass tubes on preculture medium (PM), namely a CaCl2-free MM medium supplemented with 1.35 μM BAP and 3,4-D (Montoro et al. 2003).

Binary vectors and Agrobacterium strain

The two binary vectors had a pCAMBIA 2300 backbone, a pPZP-based small binary vector (Hajdukiewicz et al. 1994), with the NPTII gene conferring resistance to neomycin under the CaMV35S promoter. The first binary vector (pCAMBIA2301-GFP-SOD) was constructed with 2 reporter genes, the GFP S65T gene containing an StLS1 intron 2 from pCAMBIA30063 and the uidA gene containing a catalase intron (Vancanneyt et al. 1990) and the CuZnSOD gene ORF (gi:27449245), all under the CaMV35S promoter. The second binary vector (pCAMBIA2300-GFP-SOD) was identical to the first one without the uidA gene expression cassette.

The binary vectors were introduced into Agrobacterium tumefaciens strain EHA 105 by electroporation. For inoculation, bacteria were grown in liquid Lysogeny Broth medium (Duchefa, Haarlem, Netherlands) supplemented with 50 mg L−1 kanamycin and 100 μM acetosyringone at 28 °C up to OD600nm = 0.6. After centrifugation at 1,000 g for 10 min, the pellet was dissolved to OD600nm at 0.06 in liquid MM from which Fe-EDTA, CaCl2 and growth regulators were eliminated and to which 100 μM acetosyringone was added (Montoro et al. 2000; Rattana et al. 2001).

Inoculation and selection of transgenic calli

Inoculation was performed as described previously (Blanc et al. 2006). Briefly, forty glass tubes containing precultured embryogenic calli from clone PB260 were used (Montoro et al. 2000). Calli were immersed directly in the tube for 1 s in the Agrobacterium suspension prepared as described above. Two coculture durations (4 and 5 days) were tested at 20 °C (Blanc et al. 2006). Six hundred small aggregates per treatment were then placed in 20 Petri dishes containing a decontamination medium (DM), a MM containing 500 mg L−1 ticarcillin (Sigma, Saint-Louis, USA), to prevent Agrobacterium growth (Fig. 1a).
https://static-content.springer.com/image/art%3A10.1007%2Fs11103-012-9942-x/MediaObjects/11103_2012_9942_Fig1_HTML.gif
Fig. 1

Magnification is indicated on each photo. a GFP fluorescence in callus 3 weeks after inoculation; b GFP fluorescence after 3 sub-cultures, arrow 1: non-transformed part; arrow 2: fluorescent part; c fully fluorescent calli; d GFP fluorescence in a somatic embryo; e GFP fluorescence in the leaves and stem of a transgenic plantlet; f GFP fluorescence in the root of a transgenic plantlet; g Southern-blot analysis with a radio-labelled uidA probe, lane 1: line TS5T3Af, lane 2: line TS4T8Ah, lane 3: line TS4T8Aj, lane 4: line TS4T8An, arrows indicate 10,000 base pairs; h Southern-blot analysis with a radio-labelled NPTII probe, lane 1: line TS4T2A44, lane 2: line TS4T2A23, lane 3: line TS4T8Ah, lane 4: line TS4T8Aj, lane 5: line TS4T8An, arrows indicate 10,000 base pairs. i Necrotic WT callus with embryos in the RITA® system. j Healthy and non-embryogenic callus over-expressing the HbCuZnSOD gene

To isolate transgenic callus lines, GFP-positive aggregates were successively subcultured every 3 weeks on DM and then several times on DM with increasing concentrations of paromomycin from 50 to 150 mg L−1 (Fig. 1b; Rattana et al. 2001). Finally, transgenic callus lines were established from sub-aggregates showing full GFP activity (Fig. 1c; Leclercq et al. 2010). These callus lines were then subjected to molecular characterization, plant regeneration and/or cryopreservation according to the protocol described previously (Lardet et al. 2007).

GFP assay

GFP visualisation was performed on callus at the end of each subculture under a fluorescence stereomicroscope and macroscope (MZ FLIII, Leica Microsystems, Wetzlar, Germany) using the GFP2 filter (480 nm excitation filter/510 nm barrier filter).

Genomic DNA extraction from calli and Southern–blot hybridization

DNA from transgenic callus lines was isolated as described in (Leclercq et al. 2010). One gram of tissues was ground in liquid nitrogen and then mixed with 6 mL of MATAB buffer (100 mM Tris–HCl pH 7.5, 2 % w/v MTAB, 0.4 % w/v sodium sulphite, 1 % PEG 6000, 1.4 M NaCl, 20 mM EDTA). Extracts were maintained at 74 °C for 20 min, and proteins removed using an equal volume of 24:1 chloroform-isoamyl alcohol (CIAA) followed by centrifugation at 6,220 g for 10 min. Supernatants were transferred to clean tubes and DNA precipitated with 5 mL of isopropanol followed by centrifugation at 13,000×g for 15 min. DNA pellets were washed in 70 % ethanol. Air-dried pellets were re-suspended in 300 μL of TE buffer.

Ten micrograms of genomic DNA was fragmented with EcoRI restriction enzyme and fractionated by electrophoresis in a 0.8 % agarose gel in TAE 1X buffer. After transfer onto a Hybond N+ nylon membrane (Amersham Biosciences, England), hybridization was performed as described in (Sambrook et al. 1989), using random primed 32P radio-labelled probes corresponding to uidA and NPTII genes amplified with the following primers:
  • uidA-F: 5′- GTGGGTCAATAATCAGGAAG-3′

  • uidA-R: 5′- CCAATCCAGTCCATTAATGC-3′

  • NPTII-F: 5′- CCGGCTACCTGCCCATTCGA-3′

  • NPTII-R: 5′-GCGATAGAAGGCGATGCG-3′.

The number of bands reflected the number of T-DNA insertions.

Plant regeneration

The production of embryos and their conversion into plantlets were carried out as described in (Lardet et al. 2007) (Fig. 1d–f). Somatic embryogenesis was initiated for 4 weeks by subculturing 1 gram of callus showing full GFP activity in 250 mL flasks containing 50 mL of a semi-solid embryogenesis expression medium (EXP), which was a modified MM medium supplemented with 58.5 mM sucrose, 175.5 mM maltose, 0.44 μM BAP and 0.44 μM 3,4-D. Pro-embryo development was then carried out in a temporary immersion system (RITA®, CIRAD, France) for two subcultures of 4 weeks each with 1 min of immersion per day in the liquid development medium (DEV), which was a MM containing 234 mM sucrose and 3 mM CaCl2, without any growth regulator. Each RITA was considered as an experimental replication. Conversion of mature embryos was carried out according to (Lardet et al. 1999): well-shaped mature embryos were collected and transferred to glass tubes on a semi-solid germination medium (DEV3), which consisted of the MM medium supplemented with 1.5 mM CaCl2 solidified with 7 g L−1 Agar (Sigma, Saint-Louis, USA). Embryos were incubated under a light intensity of 60 μmol m−2 s−1 and a 12 h day/dark photoperiod up to the full conversion of embryos into plants. Plantlets were then acclimatized in the greenhouse at 28 °C with 60 % relative humidity.

To compare the regeneration ability of wild-type and transgenic callus lines, wild-type callus line CI05519 was cultured over the length of the transformation experiment and regenerated. Once enough transgenic callus was produced, plant regeneration was initiated. For both non–transformed and transgenic callus lines, the regeneration replication number, the number of total embryos per gram of callus (T), the number of well-shaped embryos per gram of callus (WS), the number of plantlets per gram of callus (P) and the conversion percentage (P/WS) were recorded.

RNA extraction from callus

RNA from transgenic callus lines was isolated with a modified protocol described in (Leclercq et al. 2010). One gram of tissues was ground in liquid nitrogen and then mixed with 6 mL of MATAB buffer (100 mM Tris–HCl pH 7.5, 2 % w/v MTAB, 0.4 % w/v sodium sulphite, 1 % PEG 6000, 1.4 M NaCl, 20 mM EDTA). Extracts were maintained at 74 °C for 20 min, and proteins removed using an equal volume of 24:24:1 phenol–chloroform-isoamyl alcohol (PCIAA) followed by centrifugation at 13,000×g for 10 min. Supernatants were again mixed with an equal volume of 24:1 chloroform-isoamyl alcohol (CIAA). Nucleic acids contained in the supernatant were then precipitated with 5 mL of isopropanol followed by centrifugation at 13,000×g for 30 min. Pellets were dissolved in 600 μL of sterile milliQ water and RNA was differentially precipitated with 200 μL of 8 M LiCl overnight at 4 °C. After centrifugation at 13,000×g for 30 min, pellets were washed in 70 % ethanol, air-dried and dissolved in 100 μL TE buffer.

RNA extraction from leaves

Four 1-year-old plants per line were used for gene expression analysis. The RNA extraction procedure used has been previously described in Duan et al. (2010). Briefly, 1 g of plant material was ground in liquid nitrogen and 30 mL of extraction buffer (4 M guanidium isothiocyanate, 1 % sarcosine, 1 % PVP, and 1 % ß-mercapto-ethanol) was added to the powder. After mixing and centrifugation at 13,000×g at 4 °C for 30 min, the supernatant was loaded on 8 mL of 5.7 M CsCl. Ultracentrifugation at 89,705×g, at 20 °C for 20 h was carried out in a swinging bucket. After discarding the supernatant and the cesium cushion, the RNA pellet was washed with 70 % ethanol, air dried and dissolved in 200 μL of sterile water. Total RNAs were quantified with Nanoquant (Tecan, Männedorf, Switzerland) and conserved at −80 °C.

Complementary DNA (cDNA) synthesis

Before cDNA synthesis, the absence of contaminating genomic DNA was checked on all RNA samples by performing a PCR reaction with HbActin primers. If genomic DNA was detected, a DNAse treatment was performed using TurboDNAse (Ambion, Texas, USA) following the manufacturer’s instructions. Four micrograms of DNA-free RNAs was used for cDNA in a 40 μL reaction mixture using a RevertAid™ M-MuLV reverse transcriptase following the manufacturer’s instructions (MBI, Fermentas, Canada).

Real-time RT-PCR analysis

Primers were designed for the HbActin reference gene (internal control) and the HbCuZnSOD target gene: HbActin-F: 5′-AGTGTGATGTGGATATCAGG-3′, HbActin-R: 5′-GGGATGCAAGGATAGATC-3′, HbCuZnSOD-F: 5′-AGACACAACAAATGGCTGC-3′ and HbCuZnSOD-R: 5′-TGAGTGAAGGTCTTGTAAC-3′. Real-time RT-PCR analysis was carried out using a Light Cycler 480 (Roche, Basel, Switzerland) as described in Duan et al. (2010). The HbActin gene and the target genes were amplified in parallel allowing calculation of the relative gene expression ratio takking into account primers efficiencies (HbCuZnSOD primers pair: E = 1.90 and HbActin primers pair: E = 1.95; Putranto et al. 2012). All the expression data were automatically calculated by Light Cycler Software version 1.5.0 provided by the manufacturer.

Application of a drought treatment

The experiment was performed in a greenhouse cell, under controlled conditions with a mean temperature of 28.4 °C and 43.6 % relative humidity. The daylight period in the cell was 12 h, and the photosynthetic active radiation flux was an average of 600 μmol m−² s−1 above the canopy. Three-year-old plants were cut and placed in pots with the same weight of soil (EGO 140 substrate, Tref group, Netherlands). After the emission of two growth units, seven plants from the transgenic control line (TS3T4Ab), five plants from line TS5T3Af and four plants from line TS4T8An were subjected to a controlled water deficit.

Drought stress was imposed by withholding water from the pots. At the onset of soil dry-down, the surfaces of the pots were sealed with cellophane to prevent soil evaporation. In this way, it was possible to calculate both the dynamics of soil water depletion and plant transpiration from gravimetric observations. Soil water status was monitored using the fraction of transpirable soil water (FTSW) (Luquet et al. 2008). In order to estimate the FTSW value of each pot, full watering of all the pots the day before the start of measurements was followed by one night of drainage. On the next morning, the initial pot water capacity was determined by weighing all the pots. FTSW was estimated as the ratio of actual transpirable soil water (ATSW) to total transpirable soil water (TTSW), ATSW being the mass difference between daily and final pot weight. TTSW was calculated as the difference between initial pot capacity and the final pot weight after soil desiccation. The experiment ended when the transpiration rate of each stressed pot was less than 10 % of that of the fully watered pots (Sinclair and Ludlow 1986). Its value matched 1 when the plant was well watered. Drought stress continued up to FTSW = 0.1.

Transpiration, stomatal conductance, and chlorophyll fluorescence measurements

Chlorophyll fluorescence measurements were performed with a Handy-PEA® chlorophyll fluorometer (Handy-Plant Efficiency Analyser, Hansatech Instruments, King’s Lynn, Norfolk, UK) on stressed plants at the beginning of water stress and at several FTSW values during the dehydration treatment, always in the morning, at the same time as stomatal conductance measurements. The transients were induced by 1-s illumination with an array of six light-emitting diodes providing a maximum light intensity of 3,000 μmol (photons) m−2 s−1 and uniform irradiation over a 4-mm diameter leaf area. Fast fluorescence kinetics (F0 to FM) were recorded from 10 μs to 1 s. The fluorescence intensity at 50 μs was considered as F0 (Strasser et al. 1995). Readings were taken on the abaxial side of mature leaves, dark adapted with a lightweight plastic leafclip for 30 min before measurement. The performance index (PIabs) plant vitality indicator (Strasser et al. 2000; Strauss et al. 2006), which comprises light energy absorption, excitation energy trapping, and conversion of excitation energy into electron flow, was also measured to quantify photosystem II integrity (PSII) (Strasser et al. 1995). Each measurement was performed on apparently healthy leaves.

Stomatal conductance was measured at the same time with an SC-1 Decagon Devices leaf porometer (Pullman, USA). All stomatal conductance measurements were carried out and compared under the same environmental conditions.

Colorimetric measurement of antioxidant and protective compounds

For all measurements of antioxidant and protective compounds, a total extract was obtained from 0.5 to 0.9 g of leaf ground in liquid nitrogen in 10 mL of 3 % sulphosalicylic acid at 4 °C. Samples were centrifuged twice at 11,000×g for 15 min at 4 °C.

Ascorbic acid content was measured with the help of a standard curve established with known ASA concentration in 50 μL (0, 20, 40, 60, 80, 100 nmol). For ASA content determination in the leaf, 50 μL of total extract was mixed with 1.8 mL of freshly prepared buffer A (10 mL of 0.15 M potassium phosphate buffer pH = 7.4, 25 mL of 10 % TCA, 15 mL H2O mQ, 20 mL of 50 % phosphoric acid, and 20 mL of 4 % 2.2′dipyridyl dissolved in absolute ethanol). Then 200 μL of buffer B was added (3 % FeCl3). After 30 min, absorbance measurements were taken with a spectrophotometer at 525 nm.

Thiol content was measured with the help of a standard curve established with known GSH concentration in 1 mL (0, 20, 40, 60, 80, 100 nmol). For GSH content determination in the leaf, 500 μL of total extract was mixed with 500 μL of 3 % sulphosalicylic acid and 50 μL of DTNB (10 mM DTNB and 20 mM EDTA) and 1 mL of 0.5 M Tris. In less than 30 min, absorbance measurements were taken with a spectrophotometer at 412 nm.

Proline content was measured with the help of a standard curve established with known proline concentration in 1 mL (0, 50, 100, 150, 200, 250 nmol). For proline content determination in the leaf, 1 mL of total extract was mixed with 1 mL of ninhydrin (1.25 g of ninhydrin dissolved in 30 mL of glacial acetic acid and 20 mL of 6 M phosphoric acid) and 1 mL of glacial acetic acid. After heating for 1 h at 100 °C, the samples were placed on ice and 2 mL of toluene was added. Absorbance was measured on the upper phase at 520 nm.

Antioxidant enzyme activities

For the detection of SOD activity (EC1.15.1.1), 200 mg of fresh matter from leaves was ground in liquid nitrogen. Total proteins were then extracted from the leaf powder with 50 mM of potassium phosphate buffer, pH 7.6, 0.1 % Triton X-100 and 1 % PVP. Supernatants were collected after two centrifugations for 20 min at 15,000×g at 4 °C. Proteins were then precipitated sequentially with 25 % ammonium sulphate and 80 % ammonium sulphate. After centrifugation, the pellet was dissolved in the buffer for SOD activity (50 mM of potassium phosphate buffer, pH 7.6 and 0.1 mM EDTA) and the protein concentration was determined by the Bradford method. SOD activity was performed as described in (Misra and Fridovich 1977; Madon 2001). The reaction was performed in 1 mL in SOD activity buffer with 0.25 mM of xanthine (Sigma, Saint-Louis, USA), 0.02 U of xanthine oxidase (Roche, Basel, Switzerland) and 10 μL of protein extract. After 5 min at 25 °C, the reaction was initiated by adding 50 μL of 10 mM nitroblue tetrazolium (NBT). One unit of SOD activity was defined as the amount of enzyme extract required to cause 50 % inhibition of the rate of NBT reduction at 550 nm.

For the following enzyme activities, an extract from leaves frozen in nitrogen was prepared in 0.1 M potassium phosphate buffer, pH 7, with EDTA (0.1 mM), ascorbate (0.1 mM) and PVP (10 % w/w). Protein concentration was determined by the Bradford method. Ascorbate peroxidase, glutathione reductase, peroxidase and catalase activities were assayed according to the protocol modified by (Joët et al. 2002). Total peroxidase activity was measured in a reaction performed in 1 mL of 50 mM potassium phosphate buffer, pH 7, containing 5 mM of guaiacol and 0.25 mM of H2O2. After adding 50 μL of total extract, the kinetic evolution of absorbency at 470 nm was measured for 1 min. Peroxidase activity was calculated using the extinction coefficient (26.6 mM−1 cm−1 at 470 nm). Ascorbate peroxidase activity was measured in a reaction performed in 1 mL of 50 mM phosphate buffer, pH 7, containing 0.5 mM of ascorbate and 0.25 mM of H2O2. After adding 50 μL of total extract, the kinetic evolution of absorbency at 290 nm was measured for 1 min. Ascorbate peroxidase activity was calculated using the extinction coefficient (2.8 mM−1 cm−1 at 290 nm). Glutathione reductase activity was measured in a reaction performed in 1 mL of 100 mM phosphate buffer, pH 7.5, containing 1 mM of oxidized glutathione, 1 mM of EDTA and 0.15 mM of NADPH. After adding 50 μL of total extract, the kinetic evolution of absorbency at 340 nm was measured for 1 min. Glutathione reductase activity was calculated using the extinction coefficient (6.22 mM−1 cm−1 at 340 nm). All enzyme activities were calculated by IU/mg of protein.

Database and statistical analysis

The Biotekva database (Microsoft Access) (Leclercq et al. 2010) was designed to gather, store and manage data arising from all the experimental steps, from callus to plant production. The data were normalised prior to statistical analysis using XLSTAT (Addinsoft, Paris, France). The number of total embryos.g−1 callus, the conversion percentage and the number of plantlets.g−1 callus were normalised using the square root (X) function. The number of well-shaped embryos.g−1 callus was normalised with the Log10 (X + 1) function. Stem height was normalised with the Log10(X) function. An ANOVA analysis followed by a two-tailed Student Neuman-Keuls test were used in the statistical analyses (p < 0.05). For the expression study by real-time PCR, the expression ratio was normalised using the Log10(X) function. For each gene, an ANOVA analysis followed by a two-tailed Student Neuman-Keuls test were used in the statistical analyses (p < 0.05).

Results

The HbCuZnSOD gene potentially encodes a functional cytosolic isoform

A full-length sequence of the HbCuZnSOD gene from Hevea brasiliensis was available on NCBI (gi:27449245). The deduced amino acid sequence was compared to three Arabidopsis isoforms known to be located in different cellular compartments: cytosolic AtCDS1, chloroplastic AtCSD2 and peroxisomal AtCSD3. The deduced HbCuZnSOD protein sequence shared 75 % identity with AtCSD1, 34 % with AtCSD2 and 62 % with AtCSD3. Consequently, the HbCuZnSOD gene was predicted to encode a cytosolic isoform. This was also suggested by the absence of both the signal peptide for chloroplast targeting at the N-terminus and the extension at the C-terminus characteristic of a peroxisomal isoform (Fig. 2). Moreover, HbCuZnSOD is potentially a functional protein as the deduced protein sequences contained all the necessary residues for binding the copper ion (His-45, -47, -62, and -119), for binding the zinc ion (His-62, -70, and -79 and Asp-82) and for disulphide bridge formation (Cys-56 and -145) (Shin et al. 2005), as well as for guiding the superoxide anion to the active site (Arg-142) (Li et al. 2010).
https://static-content.springer.com/image/art%3A10.1007%2Fs11103-012-9942-x/MediaObjects/11103_2012_9942_Fig2_HTML.gif
Fig. 2

Protein sequence alignment of HbCuZnSOD with the three AtCuZnSOD isoforms from Arabidopsis (AtCSD1, 2, 3). (asterisk symbols), amino acids involved in copper and zinc ion binding, (plus symbols) cystein residues involved in the disulphide bridge, (dot symbols) arginin residue guiding the superoxide anion to the active site

Establishment and characterization of the HbCuZnSOD transgenic callus lines

The wild-type embryogenic callus line CI05519 was transformed using two binary vectors harbouring the nptII, GFP and HbCuZnSOD genes, and one additional uidA gene. Seventy-two 35S::SOD transgenic callus lines were established using GFP selection (Fig. 1a–f). Southern-blot hybridizations were carried out on fully-fluorescent GFP callus lines. Sixty-two lines harboured a single copy of T-DNA, six lines had two copies and four lines had truncated insertions (Fig. 1g, h and Supplementary Table 1). No obvious difference was detected between the two vectors used for genetic transformation, either for the number of transgenic lines obtained or for the integrity of the T-DNA insertion. The relative transcript abundance of the HbCuZnSOD gene was attempted by real-time RT-PCR on twelve 35S::GUS-35S::GFP control lines (Leclercq et al. 2010) and forty-nine 35S::HbCuZnSOD lines harbouring a single copy of T-DNA (Supplementary Table 2). Of the forty-nine lines, only the thirteen 35S::HbCuZnSOD lines had a relative transcript abundance of the HbCuZnSOD gene exceeding 10−6 (Supplementary Table 2). Further characterizations were conducted on twenty lines harbouring one copy of T-DNA (see Supplementary Tables 3 to 6), and we detailed in this paper on the plant regeneration, physiological, biochemical and molecular parameters of two fast-growing HbCuZnSOD transgenic lines (TS4T8An and TS5T3Af).

Plant regeneration from HbCuZnSOD transgenic callus lines

Plant regeneration was carried out on one non-transformed control line (CI05519), four GFP/GUS transformed control lines (TS3T4Ab, TS3T4Ac, TS3T4A22, TS3T4A24) and twenty HbCuZnSOD transformed lines (TS4 and TS5) with one single copy of T-DNA (Table 1 and Supplementary Table 3). The twenty HbCuZnSOD lines were selected based on a relative transcript abundance of the HbCuZnSOD gene ranging from 8.43 × 10−7 to 1.00 × 10−2. The number of RITA® per line depended on the friable callus proliferation rate. Nine HbCuZnSOD lines showed a very low level of proliferation allowing embryogenesis induction from only 1 gram of callus corresponding to one RITA® unit. The largest number of total and well-shaped embryos, and plantlets per gram of callus fresh matter, was recorded for the non-transformed line CI05519 with 582.2 total embryos/gFM, 88.3 well-shaped embryos/gFM and 13.9 plants/gFM, respectively. However, these best well-shaped embryo and plant production rates in the non-transformed control were not significantly different with those obtained for the transformed control line TS3T4Ab and two HbCuZnSOD transformed lines TS4T8An and TS5T3Af. As usual, all callus lines regenerating embryos and plantlets turned brown after embryogenesis induction (Fig. 1i). By contrast, some HbCuZnSOD transgenic callus lines remained yellow and proliferating (Fig. 1j). These lines were affected at various steps of the embryogenesis process: three lines did not produce any embryos, seven lines did not produce any well-shaped embryos and eleven lines did not produce any plantlets (Supplementary Table 3). No obvious correlation was detected between the relative transcript abundance of the HbCuZnSOD gene and the number of total embryos or well-shaped embryos counted as well as the nature of the vector used for genetic transformation
Table 1

Regeneration ability of the wild-type CI05519 line and independent transgenic callus lines over-expressing HbCuZnSOD derived from wild-type line CI05519 harbouring a single copy of T-DNA

Lines

T-DNA

Relative transcript abundance (callus)

Replication (nb of RITAs)

Total embryos (nb.g−1 FM)

Well-shaped embryos (WS) (nb.g−1 FM)

Plantlets (P) (nb.g−1 FM)

Conversion (% P/WS)

CI05519

nd

8

582.2a

88.3a

13.9a

28.1b

TS3T4Ab

GUS, GFP

9.60 × 10−07

17

256ab

15bc

5c

33ab

TS4T8An

GFP, GUS, SOD

2.23 × 10−04

1

170.1ab

32.5a

8.4ab

26.0ab

TS5T3Af

GFP, GUS, SOD

3.80 × 10−06

5

410.1a

32.2a

9.5ab

24.6bc

The data were analysed with XLSTAT software after normalisation. The number of total embryos.g−1 callus, the conversion percentage and the number of plantlets.g−1 callus were normalised using the square root (X) function. The number of well-shaped embryos.g−1 callus was normalised with the Log10 (X + 1) function. Statistical analysis was performed with an ANOVA followed by the Newman-Keuls test. Values with the same letter are not significantly different at the 0.05 probability level

nd not determined

Relation between the relative transcript abundance of the HbCuZnSOD gene and total SOD activity

The relative transcript abundance of the HbCuZnSOD gene was monitored by real-time RT-PCR on callus and leaves of plants (Table 2 and Supplementary Table 4). The relative transcript abundance of the native HbCuZnSOD gene was of 9.60 × 10−7 in the callus of line TS3T4Ab, and between 1.68 × 10−3 and 2.24 × 10−3 in leaves of plants from wild-type and transgenic control line TS3T4Ab. Consequently, no obvious correlation was found for the relative transcript abundance of the HbCuZnSOD gene between callus and leaves. The relative transcript abundance of the HbCuZnSOD gene in HbCuZnSOD over-expressing lines was greatly enhanced by 4 to 232-fold in callus and between 94 and 258-fold in leaves of plant from lines TS4T8Af and TS5T3An compared to the level in leaves of plant from line TS3T4Ab (Table 2). The accumulation of the HbCuZnSOD transcripts was not in line with the increase in total SOD activity for plant lines TS4T3Af and TS4T8An (Table 2). The highest SOD activity was observed for line TS5T3Af with 6.01 IU/mg of protein. Although the accumulation of HbCuZnSOD transcripts for line TS4T8An was much higher than that for lines TS5T3Af (2.7-fold) and TS3T4Ab (258-fold), its total SOD activity amounted to 40 and 72 % of those lines, respectively.
Table 2

Level of relative transcript abundance and SOD activity

Lines

T-DNA

Relative transcript abundance

SOD activity (IU/mg protein)

Callus

Leaves

Leaves

CI05519

nd

1.68 × 10−03b

nd

TS3T4Ab

GUS, GFP

9.60 × 10−07

2.24 × 10−03b

3.39ab

TS4T8An

GFP, GUS, SOD

2.23 × 10−04

5.79 × 10−01a

2.45b

TS5T3Af

GFP, GUS, SOD

3.80 × 10−06

2.11 × 10−01a

6.01a

The HbCuZnSOD gene was monitored by real-time RT-PCR in callus (one biological replication) and leaves (4 biological replications) of WT (CI05519), control transformed lines (TS3 lines) and transformed lines over-expressing HbCuZnSOD (TS4 and TS5 lines). SOD activities were performed on six plants for TS3T4Ab, three plants for TS5T3Af and four plants for TS5T8An. Expression data from leaves were analysed with XLSTAT software after normalisation with the Log10(X) function. All statistical analyses were performed with an ANOVA followed by the Newman-Keuls test. Values with the same letter are not significantly different at the 0.05 probability level

Survival rate and plant growth in HbCuZnSOD over-expressing lines

Although no obvious morphological difference was observed between the wild-type, the transgenic control lines and the HbCuZnSOD over-expressing lines, survival dramatically decreased in most HbCuZnSOD lines (Table 3 and Supplementary Table 5). Data are shown only for the transgenic lines with more than 10 plants at the acclimatization stage. After acclimatization, the survival rate of the transgenic control plants from TS3 lines was over 90 % after 2 and 6 months. After 1 year, the survival rate of those transgenic control lines decreased to reach 93.9 % for line TS3T4Ab. In contrast, the survival rate of HbCuZnSOD over-expressing lines was lower than that of the transgenic control lines. After 2 months, the survival rate reached 68.9 % for line TS5T3Af and then decreased again to about 40.0 % after 6 months and 1 year. However, no significant difference was observed 12 months after acclimatization, except for line TS5T3Af. At the same time as the survival rate was measured, stem heights were recorded at the acclimatization stage and after 2, 6 and 12 months (Table 4 and Supplementary Table 6). At the beginning of acclimatization, no significant difference was seen between the wild-type, the transgenic control lines and the HbCuZnSOD over-expressing lines, with a stem height of between 43 and 59 mm, except for line TS5T3Af, which had smaller plants with a stem height of 37 mm. After 2 months, plants from line TS5T3Af were significantly smaller with a stem height of 43 mm. In contrast, 6 months after acclimatization, plants from line TS5T3Af were significantly taller than the wild-type and transformed control lines. After 1 year of acclimatization, plants from line TS4T8An, and above all, those from line TS5T3Af were significantly taller with a stem height ranging from 819 to 992 mm, respectively and with well-expanded leaflets (data not shown).
Table 3

Survival rate of transgenic control plants (TS3 lines) and transgenic plants over-expressing HbCuZnSOD (TS4 and TS5 lines)

Lines

T-DNA

Acclimatized plants (no)

Survival rate (%)

2-Month-old

6-Month-old

12-Month-old

TS3T4Ab

GUS, GFP

66

100.0a

93.9a

93.9ab

TS4T8An

GFP, GUS, SOD

27

100.0a

77.8ab

77.8abc

TS5T3Af

GFP, GUS, SOD

61

68.9b

41.0b

39.3c

The values are the percentages of living in vitro plantlets after 2, 6 and 12 months. The data were analysed with XLSTAT software with a parametric comparison using the Marascuilo test. Values with the same letter are not significantly different at the 0.05 probability level

Table 4

Height of WT, transgenic control plants (TS3 lines) and transgenic plants over-expressing HbCuZnSOD (TS4 and TS5 lines)

Lines

T-DNA

Plant (no)

Stem height (mm)

Acclimatization

2-month-old

6-month-old

12-month-old

CI05519

150

52ab

56bc

160b

511b

TS3T4Ab

GUS, GFP

66

44ab

94a

124b

550b

TS4T8An

GFP, GUS, SOD

27

59a

83a

132b

819ab

TS5T3Af

GFP, GUS, SOD

61

37b

43c

207a

992a

Stem height was measured for each in vitro plantlet from the collar. The values are the means of total in vitro plantlets. The data were analysed with XLSTAT software after normalisation with the Log10(X) function. Statistical analysis was performed with an ANOVA followed by the Newman-Keuls test. Values with the same letter are not significantly different at the 0.05 probability level

Effect of water deficit on plant development and biochemical parameters

Plants from the two HbCuZnSOD lines TS5T3Af and TS4T8An and the transgenic control line TS3T4Ab were dehydrated at FTSW levels varying from 1 to 0.1. On the first day, all the plants were well watered and had dark green leaves (Fig. 3a–c). The stem height for the two HbCuZnSOD lines was taller than for the transformed control line. After 21 days without watering, the transformed control and TS5T3Af plants displayed leaf symptoms associated with a water deficit: leaf curling (Fig. 3d) followed by leaf yellowing, highlighting the hydathodes (Fig. 3e). In contrast, plants from line TS4T8An were still healthy (data not shown). All the plants started to defoliate when FTSW was below 0.1 (data not shown). After re-watering, all the plants were able to produce a new and healthy growth unit (Fig. 3g), including plants with necrotic apical meristem that had emitted a new shoot from an auxiliary bud.
https://static-content.springer.com/image/art%3A10.1007%2Fs11103-012-9942-x/MediaObjects/11103_2012_9942_Fig3_HTML.jpg
Fig. 3

Application of drought stress under controlled conditions and leaf symptomatology upon water deficit treatment a Day 0: TS4T8An plants, b Day 0: TS5T3Af plants, c Day 0: TS3T4Ab control plants, d leaf curling of TS3T4Ab control plant and TS5T3Af plants, e yellowing of the leaves of TS3T4Ab control plant and TS5T3Af plants highlighting the hydathodes, f after re-watering, appearance of a new growth unit with dark green leaves, g complete refoliation 2 months after re-watering, from left to right: TS5T3Af, TS4T8An and TS3 control plants

Biochemical parameters showed that ROS were generated under the water deficit treatment. Indeed, the redox status of all the plants was modified under the water deficit treatment, with a drastic reduction in the level of thiol and ascorbate contents, without any difference between the plant lines (Fig. 4a, b). The proline content significantly increased under the water deficit (Fig. 4c). However, a discrepancy was observed between the plant lines. Line TS4T8An displayed the highest level of proline (15.75 μmol/g DM) compared to the transformed control line (4.00 μmol/g DM) and the other CuZnSOD line TS5T3Af (5.47 μmol/g DM). Enzymatic activities were also monitored, such as total SOD, total peroxidase, ascorbate peroxidase and glutathione reductase activities. Even though no significant difference in total SOD activity was observed between all plant lines under the water deficit treatment (Table 5), the relative ratio of activity revealed that line TS5T8An displayed a significant increase in SOD activity (2.6-fold) under a water deficit treatment compared with the significant decrease in SOD activity observed for TS3T4Ab and TS5T3Af (1.01 and 0.89, respectively) (Table 5). In addition, total peroxidase activity dramatically decreased after the water deficit treatment in all transgenic plants (Fig. 4d). As regards ascorbate peroxidase and glutathione reductase activities (Fig. 4e, f), no significant difference was noticed.
https://static-content.springer.com/image/art%3A10.1007%2Fs11103-012-9942-x/MediaObjects/11103_2012_9942_Fig4_HTML.gif
Fig. 4

Biochemical data monitored in well-watered plants (black) and upon a water deficit treatment FTSW of 0.1 (grey). The data are the mean and the standard error for seven TS3T4Ab control plants, five TS5T3Af plants and four TS4T8An plants a thiol content, b ascorbate content, c total peroxidase activity, d proline content, e ascorbate peroxidase activity, f glutathione reductase activity. The data were analysed with XLSTAT software. A statistical analysis was performed with an ANOVA followed by the Newman-Keuls test. Values with the same letter are not significantly different at the 0.05 probability level

Table 5

SOD activities and relative SOD activity in well-watered (FTSW = 1) and stressed plants (FTSW = 0.1) in the three transgenic lines tested

Line (name)

SOD activity (IU/mg)

FTSW = 1

FTSW = 0.1

Ratio (FTSW = 0.1/FTSW = 1)

TS3T4Ab

3,39 ± 0.65ab

3.45 ± 0.63a

1.01 ± 0.28b

TS5T3Af

6,01 ± 1.35a

4.27 ± 1.79a

0.89 ± 0.41b

TS4T8An

2,46 ± 0.65b

6.85 ± 1.81a

2.6 ± 0.16a

The data are the mean and the standard error for six TS3T4Ab control plants, three TS5T3Af plants and four TS4T8An plants. Data were analysed with XLSTAT software. A statistical analysis was performed with an ANOVA followed by the Newman-Keuls test. Values with the same letter are not significantly different at the 0.05 probability level

Physiological analysis in HbCuZnSOD over-expressing plantlets under a water deficit treatement

Several physiological parameters, such as stomatal conductance (gs), the time needed to reach FTSW = 0.1 and the performance index (PIabs), were monitored as the FTSW was decreased. In fact, without any treatment, the initial stomatal conductance of the TS4T8An plants was significantly lower (85 mmol.m² s−1) than in the plants from the TS5T3Af and TS3T4Ab control lines (150 and 170 mmol.m² s−1, respectively). Stomatal conductance was not statistically correlated to the relative level of HbCuZnSOD expression (data not shown). When FTSW reached 0.1, the level was about 10 mmol.m² s−1 for all the plant lines (Fig. 5a). There was a distinction between the TS4T8An plants and the TS5T3Af and control plants for the time taken to reach a FTSW of 0.1 (Fig. 5b). Plants from line TS4T8An took 25 days without any watering to reach a FTSW of 0.1, compared with the 14 days taken by plants of the other two lines (Fig. 5B). This additional period was probably due to a lower transpiration rate in the plants of line TS4T8An. The PIabs evolved differently between the control plants and those of the HbCuZnSOD lines during dehydration (Fig. 5c). In control plants, the PIabs = 9 in well-watered plants increased to 14.1 between FTSW = 0.4 and FTSW = 0.2 and finally decreased to reach PI = 11.6 at FTSW = 0.1. Although quite stable for the plants of line TS5T3Af (PI = 8.5), the PI of the plants of line TS4T8An varied between 6 and 9 with a peak for a FTSW of 0.2. A decline in PIabs was observed for all lines after FTSW = 0.2.
https://static-content.springer.com/image/art%3A10.1007%2Fs11103-012-9942-x/MediaObjects/11103_2012_9942_Fig5_HTML.gif
Fig. 5

Ecophysiological data monitored in well-watered plants. The data are the mean and the standard error for seven TS3T4Ab control plants (diamond, grey), five TS5T3Af plants (black square, black dottedline) and four TS4T8An plants (black triangle, black solid line) a stomatal conductance measurement depending on the FTSW, b days without watering needed to reach FTSW = 0.1, c PI abs measurement depending on the FTSW

Discussion

The work presented here described the successful establishment of transgenic lines carrying an additional copy of the Hevea CuZnSOD gene and characterization of the development of transgenic plants over 1 year following acclimatization, and the biochemical and physiological response to a water deficit treatment. A genetic modification effect was found on somatic embryogenesis and plant growth. Plants from two HbCuZnSOD over-expressing lines subjected to water deficit displayed changes in several physiological and biochemical parameters related to ROS-scavenging systems.

Transgenic lines carrying an additional copy of the Hevea CuZnSOD gene are affected in the somatic embryogenesis process

In our system, plants were regenerated from transgenic embryogenic callus lines (Blanc et al. 2006; Leclercq et al. 2010). The relative abundance of HbCuZnSOD transcripts varied from 2.6 × 10−7 to 2.09 × 10−9 in transgenic control lines with only the native HbCuZnSOD gene. In transgenic HbCuZnSOD over-expressing lines, transcript abundance varied greatly from 8.43 × 10−7 to 1.00 × 10−2. These lines had one copy of the native gene and one additional copy from the T-DNA insertion. In some cases, genetic modification inhibited callus browning and the regeneration of somatic embryos. The relationship between oxidative stress and somatic embryogenesis is well described in the literature. Oxidative stress enhances somatic embryogenesis in many plant species (Luo et al. 2001; Pasternak et al. 2002; Ganesan and Jayabalan 2004). Indeed, slight oxidative stress is necessary to induce the somatic embryogenesis process but excessive oxidant compounds trigger the oxidation of phenols, leading to callus browning and tissue senescence. ROS were long suggested to be also involved in recalcitrance to tissue culture in some species (Benson 2000). The mastery of somatic embryogenesis by controlling the redox status with exogenous antioxidants has been described for white spruce (Belmonte and Stasolla 2007). Recently, involvement of the redox status in embryogenesis was demonstrated (Stasolla 2010). In both angiosperms and gymnosperms, a controlled environment imposing a reduced glutathione state during the early embryogenic phases promotes cellular proliferation. In contrast, embryo development is better achieved if the glutathione pool is experimentally maintained in an oxidized state (Stasolla 2010). In our case, the proliferation of callus in the RITA® system without necrosis suggested better ROS detoxification and maintenance of a reduced environment throughout the somatic embryogenesis process, which is putatively detrimental to embryo formation and conducive to callus proliferation.

Two HbCuZnSOD over-expressing lines had a high growth rate

Despite the low rate of plant regeneration by somatic embryogenesis for the transgenic lines carrying a HbCuZnSOD gene, about 150 plants were produced and acclimatized in the greenhouse. High heterogeneity in plant height between lines may also have been related to the integration sites of the T-DNA (Mlynarova et al. 1994), gene disruption, and epigenetic phenomena (Vaucheret 2006) that can be triggered by environmental stress, including in vitro culture (Labra et al. 2004). In Hevea brasiliensis, the long in vitro culture phase might affect somatic embryogenesis by triggering somaclonal variation, which might modify embryo and plant development (Lardet et al. 2007). Abnormalities in canopy architecture, branching, leaf colour and shape have been observed in field trials (Carron et al. 2009). No morphological differences except stem heights were noticed between plants of the HbCuZnSOD lines with low or high HbCuZnSOD gene expression. One-year-old plants showed significant differences in stem height and two lines produced fast-growing plants (TS4T8An and TS5T3Af). Physiological data showed HbCuZnSOD over-expression to affect stomatal conductance. Indeed, among the plant lines tested, the lowest stomatal conductance was observed for the line with the highest relative level of HbCuZnSOD expression (line TS4T8An). Further molecular analyses are needed to characterize the integration site of the plantlets produced from line TS4T8An and TS5T3Af.

Modifying the redox status of plants could thus influence the expression of a number of ROS-inducing genes. Indeed, ROS are also signalling molecules involved in plant growth and root differentiation (Gill and Tuteja 2010; Tsukagoshi et al. 2011). In addition, post-transcriptional regulations of cytosolic CuZnSOD by microRNA 398 were reported in Arabidopsis (Beauclair et al. 2010; Sunkar et al. 2006) and could also affect the stability of HbCuZnSOD transgene expression. However, cytosolic HbCuZnSOD is not predicted to be targeted by miR398 (Gébelin et al. 2012). This is in agreement with this work showing concomitant over-expression and SOD activity in transgenic Hevea plants.

The HbCuZnSOD line TS4T8An displayed physiological and biochemical characteristics for tolerance to water deficit

The fast-growing plants from line TS4T8An showed some specific traits compared with the other lines. A high proline content and a dramatic increase in enzyme activity (SOD, ascorbate peroxidase and glutathione reductase) were noted in TS4T8An plants after a water deficit. The higher proline accumulation in TS4T8An plantlets was the first phenotypic trait to distinguish them from the control line. The drastic increase in proline content confirmed the physiological response to drought, as previously described for other plants (Chaves et al. 2009). The role of proline in stabilizing protein structure and in regulating cellular redox potential by controlling free radical levels has been described in the literature (Hong et al. 2000; Hare et al. 1999). Plants from line TS4T8An seemed to display better ROS protection, probably either through better protection of the antioxidant proteins by proline, as described in Székely et al. (2008), or through activation of the complete enzymatic detoxification system, as already observed, for example, in Nicotiana tabacum (Gupta et al. 1993). Further physiological analyses are needed for a better understanding of the role of antioxidant systems in line TS4T8An against oxidative stress. Moreover, TS4T8An plants displayed reduced stomatal conductance. Surprisingly, the low stomatal conductance of this line was related to a high growth rate. Saving more water meant that it took 25 days to reach a FTSW of 0.1 when 17 days were necessary for the other lines. The high rate of growth under normal conditions and the low water consumption under water stress suggest the plants of this line have better water use efficiency.

SOD is critical for tolerance to water stresses in Hevea brasiliensis

Hevea is known to be susceptible to flooding in plantations (Compagnon 1986), which was confirmed here on young plantlets too. The curve traced by PIabs in line with the FTSW for the control plants was reproducible and was also observed for wild-type plants (data not shown), suggesting a sensitivity of Hevea to over-watering, with a FTSW of 1 leading to root anoxia. Different studies on rice have also shown that PI is a sensitive parameter for submergence effects (Sarkar et al. 2004) and is positively correlated with CO2 assimilation (Van Heerden et al. 2007). Consequently, Hevea plantlets might be effectively susceptible to water excess. Recovery (increase in PI) seemed to operate for the control plant line at intermediate water stress levels (FTSW of 0.4). Further physiological analyses will be carried out to confirm or refute this hypothesis by combining fluorescence measurements with other non-invasive techniques, such as gas exchanges. Secondly, the response to water deficit was different for HbCuZnSOD over-expressing plantlets and for wild-type and transgenic control plantlets. An excessive water deficit (FTSW = 0.1) led to oxidative stress revealed by the consumption of thiols and ascorbate and a decrease in peroxidase activity for all the plant lines. This decrease in antioxidants under a water deficit was previously described in several plant species (Apel and Hirt 2004; McKersie et al. 1996; Mittler et al. 2004; Smirnoff 2003). Total peroxidase dramatically decreased after a water deficit, whereas changes in ascorbate peroxidase and glutathione reductase activities were not clear. Although the relative CuZnSOD transcript abundance was stable and reproducible, we noted that enzymatic activity was more variable. The relative transcript abundance of HbCuZnSOD was 2.7-fold higher in plants from line TS4T8An than plants from line TS5T3Af. However, the SOD activity measurement did not suggest over-activity in relation with over-expression of the HbCuZnSOD gene. The inhibition of SOD activity in line TS5T8An might be explained by its self-inhibition by H2O2 (Misra and Fridovich 1977, Faize et al. 2011).

As regards the integrity of PSII and plant vitality during the water deficit treatment (Strasser et al. 2000), PIabs also distinguished between the HbCuZnSOD over-expressing plants and the control plants. Indeed, PIabs was stable for TS4T8Af and TS4T5An plants, suggesting an avoidance strategy. The PI parameter is also known to be sensitive to water deficit (Oukarroum et al. 2007; Christen et al. 2007). When FTSW reached 0.2, a decline in PIabs was observed for all lines.

Is SOD critical for tolerance to TPD in Hevea brasiliensis?

TPD is recognized as a complex physiological disorder, as no biotic causal agent has been isolated from TPD-affected trees (Chen et al. 2003). TPD causes annual rubber production losses estimated at 10–40 % after 30 years of rubber cultivation. Susceptibility to TPD is likely to be genotype-dependent. For instance, clones PB 260 and RRIM 600 are the most susceptible of the cultivated clones (Yang and Fan 1995; Jacob 1989). These clones have a high latex metabolism and low antioxidant protection (Gohet et al. 1995; Lacote et al. 2010b). To date, few articles have described the identification of genes or proteins associated with the occurrence of TPD (Sookmark et al. 2002; Dian et al. 1995; Venkatachalam et al. 2007, 2009). Recently, 237 differentially expressed genes were identified and classified in clone RY8-79 (Li et al. 2010). Venkatachalam et al. (2009) proposed a hypothetical model of the cellular transduction pathway during TPD, placing reactive oxygen species (ROS) accumulation and signalling at the starting point of the TPD syndrome. Li et al. (2010) demonstrated at transcriptional level that TPD trees are overwhelmed with ROS production, as genes encoding proteins scavenging ROS are globally down-regulated, which should be correlated with previous data showing a decrease in the SOD isoenzyme (Xi and Xiao 1988) and an increase of NAD(P)H oxidase activity in TPD trees (Chrestin 1989; Chrestin et al. 1984). In Hevea, ROS accumulation in the latex cells leads to membrane peroxidation, the release of coagulant compounds, the flocculation of rubber particles and the plugging of latex vessels (Chrestin 1989; Chrestin et al. 1984), which is one of the symptoms of TPD trees. As regards the complexity of ROS control, transgene targeting in the right cellular compartment, where oxidative stress is the most detrimental, is important. Hevea is known for its susceptibility to over-harvesting, flooding, drought and cold (Compagnon 1986). All theses stresses could lead to TPD beyond a certain threshold. Our data suggested that plantlets are susceptible to both water excess and water deficit. Over-expression of the HbCuZnSOD cytosolic isoform could improve the ROS scavenging system against an oxidative burst whatever its origin. Targeting transgene expression in latex cells could help to improve latex redox homeostasis for better protection of lutoids against oxidative damage, by using the promoter from gene HEV2.1, which encodes the major latex Hevein protein (Montoro et al. 2008).

Conclusion

Even though many transgenic lines over-expressing HbCuZnSOD were obtained in our experiment, few lines produced enough plants for physiological analysis. Many limitations were encountered during the somatic embryogenesis process and counter-selection probably took place. TS4T8An and TS5T3Af were the best plant material obtained but probably not the best lines over-expressing the gene of interest.

Acknowledgments

This work was supported by the Institut Français pour le Caoutchouc. The authors thank Thierry Chapuset for the Biotekva database. The authors would also like to thank V. Pujade-Renaud for critical reading of the manuscript and P. Biggins for revision of the English version.

Supplementary material

11103_2012_9942_MOESM1_ESM.docx (11 kb)
Supplementary material 1 (DOCX 10 kb)
11103_2012_9942_MOESM2_ESM.docx (12 kb)
Supplementary material 2 (DOCX 11 kb)
11103_2012_9942_MOESM3_ESM.docx (16 kb)
Supplementary material 3 (DOCX 15 kb)
11103_2012_9942_MOESM4_ESM.docx (13 kb)
Supplementary material 4 (DOCX 12 kb)
11103_2012_9942_MOESM5_ESM.docx (13 kb)
Supplementary material 5 (DOCX 12 kb)
11103_2012_9942_MOESM6_ESM.docx (13 kb)
Supplementary material 6 (DOCX 13 kb)

Copyright information

© Springer Science+Business Media B.V. 2012