|Home | About | Journals | Submit | Contact Us | Français|
Salinity limits crop productivity, in part by decreasing shoot concentrations of the growth-promoting and senescence-delaying hormones cytokinins. Since constitutive cytokinin overproduction may have pleiotropic effects on plant development, two approaches assessed whether specific root-localized transgenic IPT (a key enzyme for cytokinin biosynthesis) gene expression could substantially improve tomato plant growth and yield under salinity: transient root IPT induction (HSP70::IPT) and grafting wild-type (WT) shoots onto a constitutive IPT-expressing rootstock (WT/35S::IPT). Transient root IPT induction increased root, xylem sap, and leaf bioactive cytokinin concentrations 2- to 3-fold without shoot IPT gene expression. Although IPT induction reduced root biomass (by 15%) in control (non-salinized) plants, in salinized plants (100mM NaCl for 22d), increased cytokinin concentrations delayed stomatal closure and leaf senescence and almost doubled shoot growth (compared with WT plants), with concomitant increases in the essential nutrient K+ (20%) and decreases in the toxic ion Na+ (by 30%) and abscisic acid (by 20–40%) concentrations in transpiring mature leaves. Similarly, WT/35S::IPT plants (scion/rootstock) grown with 75mM NaCl for 90d had higher fruit trans-zeatin concentrations (1.5- to 2-fold) and yielded 30% more than WT/non-transformed plants. Enhancing root cytokinin synthesis modified both shoot hormonal and ionic status, thus ameliorating salinity-induced decreases in growth and yield.
More than 800 million ha of land worldwide are affected by salinity (Qadir et al., 2007), thus improving crop salt tolerance is a key global agricultural goal. Salt decreases crop yields by reducing growth and inducing leaf senescence due to both an osmotic stress (plant water deficit) and an ionic stress provoked by nutritional imbalances and the accumulation of toxic ions such as sodium (Munns and Tester, 2008). Transgenic overexpression of certain ion transporters can improve salt tolerance by decreasing the uptake of toxic ions and/or altering their compartmentation within the plant (Zhang and Blumwald, 2001; Shi et al., 2002; Chen et al., 2007). However, since this technology has yet to be exploited in commercial varieties, alternative approaches may also be valuable.
Among the plant hormones acting as ‘master regulators’ of multiple physiological processes (Davies, 2005), cytokinins (CKs) are especially important in regulating cell division and expansion (Kurakawa et al., 2007) and delaying senescence (Gan and Amasino, 1995; Guo and Gan, 2007). Since both water (Kudoyarova et al., 2007; Havlová et al., 2008) and salt (Albacete et al., 2008; Ghanem et al., 2008) stress decreased plant CK status, attenuating this decrease transgenically [via overexpression of CK biosynthesis genes such as isopentenyltransferase (IPT)] might improve crop stress tolerance. However, massively increased (up to 150-fold) CK concentrations of some constitutive [e.g. under the cauliflower mosaic virus (CaMV) 35S promoter] IPT transformants decreased root growth and induced water stress (Smigocki and Owens, 1989; Synková et al., 1999; Pospíšilová, 2003). Tightly regulated transgenic IPT expression offers one approach to improving plant stress tolerance. Expression driven by a senescence-specific promoter (SAG12; Gan and Amasino, 1995) or maturation-induced promoter (SARK; Rivero et al., 2007) delayed the age-dependent decline in photosynthesis (SAG12::IPT; Wingler et al., 1998), thus minimizing yield loss of plants grown under heat stress (Xu et al., 2009) and with limiting water supplies (SARK::IPT; Rivero et al., 2007, 2009).
Despite intense interest in root-to-shoot signalling of environmental stresses (Davies and Zhang, 1991; Jackson, 1993; Dodd, 2005), whether root CK biosynthesis or delivery of CKs from root to shoot via the xylem (Aloni et al., 2005) can regulate shoot CK concentrations and thence development and senescence, especially when the root system is exposed to environmental stress, remains controversial (Itai and Vaadia, 1971; Jackson, 1993; Faiss et al., 1997; McKenzie et al., 1998; Dodd and Beveridge, 2006). Historically, CKs have been regarded as root synthesized (Letham, 1994; Keiber, 2002; Aloni et al., 2005), although analyses of spatial expression patterns of native IPT genes in Arabidopsis using their promoter::reporter constructs suggests that some of these genes are also expressed in aerial organs (Miyawaki et al., 2004; Takei et al., 2004) and influence root CK status via basipetal phloem CK transport (Matsumoto-Kitano et al., 2008). Although young shoot tissues can synthesize CKs (Taylor et al., 1990; Schmülling, 2002; Nordström et al., 2004), several studies have investigated the influence of root-localized CK production on shoot responses.
Root-synthesized CKs may have little physiological impact, since wild-type (WT) tobacco plants grafted on a 35S::IPT rootstock showed similar lateral shoot growth, senescence, and shoot CK concentrations to WT self-grafts (Faiss et al., 1997). However, McKenzie et al. (1998) suggested that a prominent gall at the graft union of those plants blocked xylem CK transport from the roots. In contrast, putatively root-specific IPT gene expression under a copper-inducible promoter (Mett et al., 1996) increased shoot CK concentrations, released apical dominance, and delayed leaf senescence in tobacco under optimal growing conditions (McKenzie et al., 1998). Although the reporter gene β-glucuronidase was massively increased in the root after addition of 50μM CuSO4 to the nutrient solution (Mett et al., 1996), xylem Cu transport to the shoot (Pich and Scholz, 1996) may also have stimulated shoot CK synthesis. Recent reciprocal grafting studies of WT and a quadruple IPT-defective Arabidopsis mutant (with decreased levels of both iP-type and tZ-type CKs) demonstrated that both WT roots and shoots could produce, export, and recover normal CK levels and growth in mutant tissue (Matsumoto-Kitano et al., 2008), thus highlighting (mutant/WT; scion/rootstock) or downgrading (WT/mutant) the physiological significance of root-synthesized CKs. The sometimes conflicting data of studies of root-localized IPT expression, within a more general recent context that root hormone supply has little physiological impact on the shoot (Holbrook et al., 2002; Christmann et al., 2007; Dodd et al., 2009), require a re-evaluation of the role of root-synthesized CKs in regulating shoot responses, especially when root systems are exposed to stressful conditions likely to down-regulate root CK production.
However, transgenic IPT expression may produce multiple hormonal phenotypes, potentially obscuring a positive role for root-synthesized CKs. Although IPT gene expression regulated by the photosynthetic small subunit promoter (PSSU::IPT) both increased shoot CK concentrations and decreased shoot abscisic acid (ABA) concentrations (Synková et al., 1999), leaf ABA concentrations of WT and SAG12::IPT (Cowan et al., 2005) and WT and SARK::IPT transgenic tobacco plants (Rivero et al., 2007) did not differ irrespective of soil moisture. Taking into account the intensive cross-talk between CKs and ABA and the importance of their ratio in regulation of several physiological processes (Dodd, 2003), analysis of ABA levels was performed in IPT transformants.
To determine whether root CK biosynthesis is important in mediating the relationship between decreased shoot CK status and salt-induced changes in growth, senescence, and fruit yield (Albacete et al., 2008, 2009, 2010; Ghanem et al., 2008), root IPT expression was up-regulated under control of a heat shock-inducible promoter (HSP70::IPT; Smigocki, 1991) by transient exposure to elevated root zone temperature (RZT) in short-term (3 weeks) experiments examining vegetative growth, and fruit yield was assessed in long-term (3 months) experiments where WT shoots were grafted onto a rootstock constitutively expressing IPT (WT/35S::IPT; Smigocki et al., 2000). Augmenting root-to-shoot CK transport improved vegetative growth and ion homeostasis, delayed leaf senescence (induced HSP70::IPT plants), and increased fruit yield (WT/35S::IPT plants) of salinized tomato, potentially providing a novel strategy to attenuate salt-induced limitations to crop productivity.
An XbaI/XmnI DNA fragment of 456bp from the Drosophila melanogaster HSP70 gene, containing the promoter and 199 nucleotides of untranslated leader sequence, was fused through its 5′-untranslated region to the coding region of the Agrobacterium tumefaciens IPT gene from pTiB6S3 (Smigocki, 1991). The resultant HSP70::IPT constructs were transferred into A. tumefaciens LBA4404 bacteria and used to transform Solanum lycopersicum L. cv. UC82-B plants as previously described (Smigocki et al., 2000).
Seeds of WT tomato (S. lycopersicum L. cv. UC82-B), transgenic T3 homozygous HSP70::IPT tomato, and internal UC82-B (denoted as 116-9) control (transformed with an empty vector) plants were sown in trays filled with a perlite–vermiculite mix (1:3, v/v) moistened regularly with half-strength Hoagland nutrient solution. Fourteen days after sowing, the substrate was gently washed from the roots and seedlings were placed in PVC plates floating on 52.0l tanks containing aerated half-strength Hoagland nutrient solution (eight plants per tank). Solutions were refilled every 2d and renewed every week. The pH was continuously maintained at 5.6–6.0. Plants were grown in a growth chamber under a 16h photoperiod. Air temperature and relative humidity during the day were 25–28°C and 70±5%, respectively, and during the night 17–18°C and 50±5%, respectively. Light intensity at the top of the canopy was ~250μmol m−2 s−1 (PPFD). Eighteen days after sowing, seedlings were exposed to 0 (control) or 100mM NaCl added to the nutrient solution in one step. A total of 240 plants were evaluated in each experiment: 30 plants per genotype, salt, and RZT treatment. Treatments were randomly distributed across available tanks within the growth chamber, and re-randomized for each experiment to avoid positional effects on growth. Although the experiment was repeated three times with similar results (Table 1), the data set presented (Figs 11–7) corresponds to Experiment 1.
To induce root IPT gene expression (regulated by a heat-shock promoter with an optimum induction temperature between 37°C and 42°C), on days 1, 8, and 22 (after salinization), half the WT and HSP70::IPT seedlings from each treatment were temporarily transferred to heated (42°C) aerated half-strength Hoagland nutrient solutions (by carefully lifting the PVC plates containing the plants from their original solution, and gently placing them into the heated solution). After 2h treatment (Smigocki, 1991), plants were returned to their normal growing conditions (nutrient solution at room temperature of 20–25°C). Trays of plants kept at normal RZT were carefully removed and replaced as described above.
RNA was isolated, and plants sampled, 2h and 48h, respectively, after each exposure to elevated RZT. Shoots and roots were separated to determine fresh weight (FW). An actively growing leaf, present when salt was applied (leaf 3, numbering from the base of the plant), was tagged for leaf area determination (using a leaf area meter: Model AM 300, ADC BioScientific Ltd, Herts, UK), and ion and hormone analysis. The two youngest unfolded upper leaves (identified as ‘young leaves’) were also collected for hormone and ion analysis. Xylem sap was obtained after severing the shoot ~2–3cm above the root, and applying pressure (–0.5MPa for control plants and about –0.9MPa for salinized plants) with nitrogen to the root system with a Scholander pressure chamber (Pérez-Alfocea et al., 2000).
Total RNA was prepared from leaves and roots using the RNAgents Total RNA isolation system of Promega (Promega Benelux b.v). First-strand cDNA was synthesized using Superscript™ III first strand synthesis for reverse transcription-PCR (RT-PCR; Invitrogen, Life Technologies, The Netherlands) according to the manufacturer's instructions. The amplification procedure for IPT and tomato actin (used as a constitutive control) genes consisted of 35 cycles; 1μl of first strand was used as template. Internal oligonucleotides, IPT forward (5'-CATCTAATTTTCGGTCCAACTTGCA-3'), IPT reverse (5'-CGATATCCATCGATCTT-3'), actin forward (5'-ATTCCCTGACTGTTTGCTAGT-3'), and actin reverse (5'-TCCAACACAATACCGGTGGT-3'), were designed, and standard RT-PCR performed. After an initial denaturation step at 94°C for 4min, each cycle consisted of 30s at 94°C, 30s at 55°C of annealing temperature, 1min extension at 72°C, followed by a final extension of 7min at 72°C. Amplifications were conducted using the GoTaq DNA polymerase (Promega Benelux b.v). PCR products were analysed by 1.5% (w/v) Tris/acetic acid/EDTA/agarose electrophoresis. The PCR products of each gene were sequenced using an automatic sequencer (Genetic Analyser 3100, Applied Biosystems, Belgium) and the ‘BigDye Terminator v1.1 cycle sequencing kit’. The resulting nucleotide sequences were aligned and compared using ClustalX.
Phytohormones (CKs and ABA) were extracted as previously reported in Dobrev and Kaminek (2002) and adapted in Novák et al. (2003) and Ghanem et al. (2008). Plant material (1g FW of leaf, root, or fruit) was homogenized in liquid nitrogen and placed in 5ml of a cold (–20°C) extraction mixture of methanol/water/formic acid (15:4:1, v/v/v, pH 2.5). After overnight extraction at –20°C, solids were separated by centrifugation (20000g, 15min) and re-extracted for 30min in an additional 5ml of the same extraction solution. Pooled supernatants were passed through a Sep-Pak Plus †C18 cartridge (SepPak Plus, Waters, USA) to remove interfering lipids and plant pigments and evaporated to dryness. The residue was dissolved in 5ml of 1M formic acid and loaded on an Oasis MCX mixed mode (cation-exchange and reverse phase) column (150mg, Waters, USA) pre-conditioned with 5ml of methanol followed by 5ml of 1M formic acid. To separate different CK forms (nucleotides, bases, ribosides, and glucosides) from ABA, the column was washed and eluted stepwise with different appropriate solutions as indicated (Dobrev and Kaminek, 2002). After each solvent passed through the columns, they were purged briefly with air. Solvents were evaporated at 40°C under vacuum. Samples then dissolved in a water/acetonitrile/formic acid (94.9:5:0.1, v/v/v) mixture for high-performance liquid chromatography/mass spectrometry (HPLC/MS) analysis. Xylem sap samples were filtered through 13mm diameter Millex filters with 0.22μm pore size and nylon membrane (Millipore, Bedford, MA, USA), and injected directly onto the LC-MS/MS system.
LC-MS/MS analysis of leaf, root and xylem sap was performed as previously described (Dobrev et al., 2002; Novák et al., 2008) using an Ion Trap Mass Spectrometer Finnigan MAT LCQ-MSn (Thermo Fisher, Waltham, MA, USA) equipped with an electrospray interface. Detection and quantification were carried out using a Finnigan LCQ MS ion trap LC-MS operated in the positive ion, full-scan MS/MS mode using a multilevel calibration graph with deuterated CKs and ABA ([2H6]cis-trans-abscisic acid, Olchemin Ltd, Olomouc, Czech Republic) as internal standards. For analyses of endogenous CKs, 50pmol of each of the following 15 deuterium-labelled standards were added: [2H5]Z, [2H5]ZR, [2H5]Z9G, [2H5]ZOG, [2H5]ZROG, [2H6]iP, [2H6]iPR, [2H6]iP9G, [2H3]DHZ, [2H3]DHZR, [2H3]DHZ9G, [2H7]DHZOG, [2H5]ZMP, [2H3]DHZMP, and [2H6]iPR5'MP (Apex Organics, Honiton, UK). Recovery percentages ranged between 92% and 95%. cis-Zeatin derivatives were determined using the retention times and MS spectra of unlabelled standards and the response ratio of their trans-zeatin counterparts. Of 47 CK derivatives of interest, the concentrations of 21 were above the detection limit: tZ, tZR, tZOG, tZROG, tZ9G, tZR5'MP, cZ, cZR, cZOG, cZROG, cZ9G, cZR5'MP, DHZ, DHZR, DHZOG, DHZROG, DHZ9G, iP, iPR, iP9G, and iPR5'MP (Supplementary Fig. S1 available at JXB online). The detection limit was calculated for each compound as 3.3 σ/S, where α is the standard deviation of the response and S is the slope of the calibration curve. Each sample was injected at least twice. Fruit trans-zeatin concentrations were determined according to the LC-MS/MS method previously described (Albacete et al., 2008; Ghanem et al., 2008). Treatment differences in tissue hormone status were consistent irrespective of whether concentrations were calculated on an FW or dry weight (DW) basis (data not shown).
Leaf and root tissues were oven-dried at 70°C for 48h and 50mg DW was digested in 35% (v/v) HNO3 at 80°C. Ions were dissolved in 0.1M HCl and concentrations determined (in triplicate for each sample) using an inductively coupled argon plasma emission spectrophotometer (YJ48; Jobin Yvon, Edison, NJ, USA) calibrated with certified standard solutions.
Modulated chlorophyll fluorescence was measured in dark-adapted (30min) leaves using a pulse-modulated chlorophyll fluorometer FMS-2 (Hansatech Instruments Ltd, Norfolk, UK) as reported previously (Ghanem et al., 2008). The maximum quantum yield of open photosystem II (PSII) (Fv/Fm) was calculated as (Fm–F0)/Fm.
To test whether root-localized CK production could improve fruit yield of salinized plants grown under commercial greenhouse conditions, seedlings from the semi-determinate cultivar P-73 were grafted as previously described (Albacete et al., 2009) onto a determinate commercial cultivar UC-82B (P-73/UC-82B, denoted as WT/WT) and on a transformant overexpressing the IPT gene as rootstocks under the control of the constitutive CaMV 35S promoter (P-73/35S::IPT; denoted as WT/35S::IPT) (Smigocki et al., 2000). The particular 35S::IPT transformant selected had a moderate 30% higher leaf trans-zeatin concentration (compared with azygous plants) after 50d of salinization without apparent phenotypic alterations in the whole plant (data not shown). Four plants per combination and salinity level were transferred to a polyethylene greenhouse after the grafts had established, using perlite as substrate. Plants were distributed in a planting pattern of 2m between rows and 0.5m between plants within rows, and cultivated as one stem by eliminating all axillary buds weekly. A standard fertilization mixture for tomato was applied by a drip irrigation system. Salt treatment started 10d after transfer and continued for 3 months. Electrical conductivities of the fertilization solutions were 1.2dS m−1 (control) and 8.7dS m−1 (75mM NaCl). Fruit yield, number, and weight were determined, and actively growing (20d and 35d after anthesis) and ripe fruits were collected for CK analysis as described above.
Data from transient root IPT induction experiments were subjected to an analysis of variance (ANOVA II) using the SAS software (SAS System for Windows, version 8.02) with mean discrimination achieved using the Student–Newman–Keuls test at the 5% level. Since WT and internal control (denoted as line 116-9 and transformed with an empty vector) plants showed similar growth responses in all experiments (Fig. 5), only data from WT plants are presented. Data of grafting experiments were subjected to two-way ANOVA (main effects of graft combination and salinity) with a t-test for small sample size based on the t-distribution using the SPSS software (version 17.0, SPSS Inc., Chicago, IL, USA) used to discriminate means within a salinity treatment.
Elevated RZT (exposure to 42°C nutrient solution for 2h to induce the promoter region controlling IPT gene expression) caused IPT gene expression in roots (Fig. 1, lanes 1 and 4) but not leaves (lanes 2 and 5) or stems (lanes 3 and 6) of HSP70::IPT plants in both control (0mM NaCl) and salt-stressed (100mM NaCl) conditions. No IPT expression was detected in roots or leaves of WT plants exposed to elevated RZT (Fig. 1, lanes 7–12), or control WT or HSP70::IPT plants maintained at normal RZT throughout the experiment (Fig. 1, lanes 13–24). Thus transient exposure of HSP70::IPT plants to elevated RZT successfully established root-localized IPT gene induction.
Individual CK metabolites were assigned to groups based on their structure and function as previously described (Havlova et al., 2008): bioactive CKs (tZ, tZR, iP, and iPR), CK deactivation forms (N-glucosides: tZ9G and iP9G), CK storage forms (O-glucosides: tZOG, tZROG, cZOG, and cZROG), and CK phosphates (CK biosynthesis intermediates: tZR5'MP, cZR5'MP, and iPR5'MP). As cis-zeatin (cZ) and its riboside (cZR) are recognized by some CK receptors, but lack physiological activity in most CK bioassays (Havlova et al., 2008), cis-zeatin derivatives cZ, cZR, and cZ9G were considered separately, while cZMP was included in other phosphates and cZOG and cZROG were included in CK storage forms. Total CK concentrations, classified as above, are available (Supplementary Fig. S1 at JXB online). Treatment differences in spatial patterns of bioactive CK metabolites are presented (Fig. 2).
Under both control and salt treatments, no genotypic differences in total bioactive CK concentrations were found at either 8d or 22d under normal RZT in young leaves (Fig. 2a, ,e),e), the more mature leaf 3 (Fig. 2b, ,f),f), xylem sap (Fig. 2c, ,g),g), or roots (Fig. 2d, h).
Elevated RZT increased bioactive CK concentrations (by 60–130%) in all compartments of HSP70::IPT plants on both measurement occasions, independently of salinization. In contrast, in WT plants, elevated RZT increased bioactive CK concentrations (by 10–40%) essentially in young leaves (days 8 and 22; Fig. 2a, ,e),e), and leaf 3 and xylem sap of control plants (day 22; Fig 2f, ,g).g). In salinized WT plants, elevated RZT only increased bioactive CKs in leaf 3 (days 8 and 22; Fig. 2b, f) and roots (day 8, Fig. 2d).
Generally, salinization significantly decreased (by 20–50%) bioactive CK concentrations in all measured plant compartments, to a similar degree in WT and HSP70::IPT plants grown under normal RZT, and in WT plants exposed to elevated RZT, especially for tZ and tZR (Fig. 2). However, this salinity-induced decrease was counteracted in HSP70::IPT plants exposed to elevated RZT; compared with salinized WT plants exposed to elevated RZT, induced HSP70::IPT plants showed higher bioactive CKs concentrations not only in roots (3-fold; Fig. 2d, h), but also in leaves and xylem sap (2-fold; Fig. 2a–c, e–g), mainly due to greatly increased tZ and tZR. Although salinity decreased CK concentrations (by 20–35%) in leaves of HSP70::IPT plants exposed to elevated RZT (compared with control HSP70::IPT plants exposed to elevated RZT), this effect was not detected in xylem sap (Fig 2c, g) or was the opposite in roots (Fig. 2d, h). Taken together, these data indicate that the CK accumulation in WT plants exposed to elevated RZT (mostly restricted to the leaves under control conditions) was much less than in HSP70::IPT plants.
Under normal RZT and control conditions, ABA concentrations of HSP70::IPT and WT plants were similar in young leaves, leaf 3, and xylem sap on both measurement occasions (Fig. 3a–c, e–g). Roots of HSP70::IPT plants had higher (1.5- to 5-fold) ABA concentrations under these conditions (cf. columns 1 and 5 of Fig. 3d, h).
Under control conditions, the effects of elevated RZT varied according to genotype and plant compartment. Elevated RZT decreased ABA concentrations (by 15–30% averaged over both genotypes and measurement occasions) in young leaves (Fig. 3a, e), had no effect in mature leaves (Fig. 3b, f), increased ABA concentrations 1.8-fold in xylem sap of HSP70::IPT plants only on day 22 (Fig. 3g), and consistently decreased (by 30–40%) root ABA concentrations of HSP70::IPT, but not WT, plants (Fig. 3d, h).
Although effects of salinization were minimal in young leaves, ABA concentrations increased significantly in mature leaves (1.5-fold), xylem sap (4- to 5-fold), and roots (5- to 25-fold) of HSP70::IPT and WT plants grown under normal RZT, and WT plants grown with elevated RZT (Fig. 3b–d, f–h). However, exposure of HSP70::IPT plants to elevated RZT attenuated this salinity-induced increase in ABA in the mature leaf, xylem sap, and roots, hence these plants had 2- to 3-fold lower ABA concentrations in those organs than salinized WT plants. Taken together, these data indicate that root IPT gene induction attenuated salinity-induced ABA accumulation.
At the end of the experiment, under normal RZT, control WT and HSP70::IPT plants had similar Na+ concentrations in young (Fig. 4a, d) and mature (Fig. 4b, e) leaves and similar K+ concentrations in roots (Fig. 4i, l). However, WT plants had higher Na+ concentrations in roots (cf. Fig. 4c, f) and higher K+ concentrations in leaves (cf. Fig. 4g, h, j, k) than HSP70::IPT plants.
Transiently elevating the RZT of control plants had no effect on Na+ concentrations in either genotype (cf. first two columns in each panel of Fig. 4a–f), but significantly increased K+ concentrations in all analysed organs of HSP70::IPT plants (Fig. 4g–i) but not WT plants (Fig. 4j–l).
Salinity significantly increased Na+ concentrations throughout the plant (Fig. 4a–f), but exposure of HSP70::IPT plants to elevated RZT attenuated the salinity-induced increases in Na+ in all plant compartments by 20mM (roots) to 40mM (young leaves) (Fig. 4a–c). Furthermore, salinity significantly decreased K+ concentrations (>50%) in mature leaves and roots (Fig. 4h, k, i, l) and either increased (HSP70::IPT plants; Fig. 4g) or had no effect (WT plants; Fig. 4j) on K+ concentrations in young leaves. However, exposure of HSP70::IPT plants) to elevated RZT ameliorated (by 20%) the decrease in K+ concentrations in leaf 3 (Fig. 4h) and roots (Fig. 4i), and actually increased K+ concentrations of young leaves (Fig. 4g), but had no significant effects in WT plants (Fig 4j–l). Taken together, these data indicate that the effects of salinity on plant ion status were modified by IPT gene induction.
Since WT and 116-9 (transformed with empty vector) plants showed similar growth irrespective of treatments (Fig. 5e–j), only WT data are described. Irrespective of RZT treatment, WT and HSP70::IPT plants were phenotypically normal (Fig. 5a), although HSP70::IPT plants tended to have a 10% lower biomass (P=0.38) than WT plants prior to imposition of salinity and RZT treatments (18d after germination). Under control conditions at normal RZT, WT plants had a greater leaf 3 area (Fig. 5b, e), and a higher shoot (1.1-fold; cf. Fig. 5c, f) and root FW (1.3-fold; cf. Fig. 5d, g) than HSP70::IPT plants. Elevated RZT abolished these genotypic differences in the shoot but not in the roots, decreased leaf area (Fig. 5e) and shoot FW (Fig. 5f) of WT plants by 20%, but had no effect on leaf area (Fig. 5b) and shoot FW (Fig. 5c) of HSP70::IPT plants. Elevated RZT inhibited root growth of WT and HSP70::IPT plants by ~15%. Similar results were obtained when data from three independent experiments were pooled (Table 1).
Under normal RZT, salinization decreased leaf 3 area, and shoot and root FW of both WT and HSP70::IPT plants. In WT plants, elevated RZT magnified the salinity-induced growth reduction of leaf 3 area, but had no additional impact on shoot and root FW. In HSP70::IPT plants, elevated RZT did not affect the salinity-induced growth reduction in root biomass, but doubled leaf 3 area and shoot FW. Thus shoot FW of salinized HSP70::IPT plants exposed to elevated RZT was only 30% less than that of control plants, while that of salinized WT plants was 60% less (Fig. 5c, f). On average, the magnitude of the increased shoot relative growth rate of salinized HSP70::IPT plants exposed to elevated RZT was equivalent to the magnitude of the decrease in leaf Na+ concentration, compared with the other treatments (Fig. 6). Taken together, these data indicate that root IPT gene induction prevented or attenuated negative impacts of elevated RZT or salinity in the shoot.
Under control conditions, elevated RZT increased the maximum quantum efficiency of PSII (Fv/Fm) in HSP70::IPT plants compared with those plants maintained at normal RZT (Fig. 7a). In contrast, RZT treatment did not change Fv/Fm in WT plants, irrespective of salt treatment (Fig. 7b). Salinization for 17d sharply decreased Fv/Fm values of WT plants independently of RZT (Fig. 7b), while this salinity-induced decrease in Fv/Fm was delayed by 1 week in HSP70::IPT plants exposed to elevated RZT (Fig. 7a).
Elevated RZT increased stomatal conductance (gs) of control HSP70::IPT plants by 20%, but had no effect on WT plants (Fig. 7c, d). From day 9 until the end of the experiment, salinity decreased gs of WT plants by 16% independently of RZT (Fig. 7d). Salinized HSP70::IPT plants at normal RZT also showed this decreased gs, but exposure to elevated RZT maintained gs comparable with control plants during the first 2 weeks of salinization. However, this positive impact disappeared by the end of the experiment.
Under control conditions, the rootstock had no significant impact on any measured fruit yield-related parameters (Table 2). However, after 3 months growing under moderate salinity, WT/35S::IPT (scion/rootstock) plants had a 30% higher fruit yield than WT/WT plants, mainly due to 25% more fruit and a significantly higher fruit weight (Table 2).
CK analyses of actively growing (20–35d after anthesis) and ripe fruits revealed that fruits of WT/35S::IPT plants had 2-fold higher tZ concentrations than fruits from WT/WT plants under both control and salinized conditions (Table 2). Salinity decreased fruit tZ concentration by 5-fold compared with control plants of both graft combinations (Table 2).
Salinity-induced leaf growth inhibition and premature senescence were correlated with decreased root, xylem sap, and leaf bioactive CK concentrations (Albacete et al., 2008; Ghanem et al., 2008; cf. columns 1, 3 and 5, 7 of Fig. 2). Root-synthesized CKs may regulate these shoot responses, since rootstocks conferring better growth and fruit yield under salinity had increased shoot xylem CK concentrations with minimal impact on xylem Na+ concentrations (Albacete et al., 2009). However, it is not clear whether root-localized IPT expression alters plant growth and development and/or delays leaf senescence of plants exposed to environmental stress, since previous studies purportedly addressing this issue (i) showed conflicting results (cf. Faiss et al., 1997; McKenzie et al., 1998); (ii) did not intentionally impose environmental stress (Matsumoto-Kitano et al., 2008); or (iii) localized IPT expression in the leaves (Gan and Amasino 1995; Synkova et al., 1999; Rivero et al., 2007; Xu et al., 2009). It was hypothesized that transient root IPT induction (HSP70::IPT) and grafting WT shoots onto a constitutive IPT-expressing rootstock with moderately higher CK production (WT/35S::IPT) would augment root CK biosynthesis and xylem transport, thus ameliorating deleterious effects of salinity on shoot growth and fruit yield.
It was first necessary to establish whether IPT gene expression was localized to the roots after exposure to elevated RZT. Since the root zone was insulated from the aerial environment by floating the hydroponically grown plants on PVC plates, and the duration of elevated RZT was relatively short (2h), leaf and stem temperature measurements did not detect any shoot warming (data not shown). Although some heat transfer through xylem sap may have occurred, RT-PCR did not detect any IPT transgene expression in stems or leaves (Fig. 1). Despite this apparent tight spatial and temporal control of IPT expression, leaf 3 area, and shoot and root FW of HSP70::IPT plants grown with normal RZT were 11% less than those of corresponding WT plants (cf. column 1 of Fig. 5b, e, c, f, and d, g), in agreement with the 10% lower biomass of the transgenic population before starting the treatments. In contrast, considerable ‘leakiness’ of heat shock-inducible promoters in tobacco under control conditions, particularly in vitro (Smart et al., 1991; Smigocki, 1991), caused pronounced phenotypic alterations (i.e. shorter, greener, bigger leaves, and delayed senescence) linked to increased CK concentrations in planta. This was not the case here (Fig. 5a); thus HSP70::IPT and WT plants had similar CK concentrations (cf. columns 1, 5 and 3, 7 in Fig. 2a–c, e–g) under normal RZT, and elevated RZT was essential to induce root IPT gene expression (Fig. 1) to increase CK concentrations (Fig. 2) of HSP70::IPT plants.
In all plant compartments measured, elevated RZT caused higher CK concentrations in HSP70::IPT plants than corresponding WT plants (cf. columns 2, 6 and 4, 8 in Fig. 2). However, exposure of WT plants to elevated RZT usually increased the bioactive CK concentration in leaves (cf. columns 5, 6 of Fig. 2a, e, b, f) but not roots (Fig. 2d, h), perhaps due to increased xylem CK delivery, since xylem CK concentration was sometimes increased (Fig. 2g) while stomatal conductance was apparently unaffected by elevated RZT (Fig. 7d). Further detailed measurements of the temporal responses of xylem CK concentration, stomatal behaviour, and native IPT root gene expression in WT plants following exposure to elevated RZT are necessary to substantiate this hypothesis.
Although elevated RZT increased shoot CK concentrations of control WT plants, it decreased leaf 3 area, and shoot and (slightly) root biomass of control WT plants (cf. columns 1, 2 of Fig. 5e, f, g), and also leaf 3 area of salinized WT plants (cf. columns 3, 4 of Fig. 5e). This shoot growth inhibition was not mediated by increased leaf ABA concentration (cf. columns 5, 6 of Fig. 3a, b, e, f); instead ABA concentrations decreased (up to 30%) in young leaves (Fig. 3a, b) consistent with a role for ABA in maintaining tomato shoot growth (Sharp et al., 2000; Dodd et al., 2009). Alternative explanations for shoot growth inhibition by elevated RZT include decreased shoot water status (He et al., 2001) and/or increased production of the growth-inhibitory plant hormone ethylene (Qin et al., 2007). Thus it was desirable to minimize plant exposure to elevated RZT. However, elevated RZT did not inhibit leaf and shoot growth (Fig. 5b, c) of control HSP70::IPT plants, probably due to increased CK concentrations (Fig. 2a, b, e, f) stimulating cell division and expansion, and thus compensating negative impacts of elevated RZT. Instead, inhibitory effects of elevated RZT on HSP70::IPT plants were restricted to the roots of non-salinized plants (cf. columns 1, 2 of Fig. 5d) which had 15% less biomass, consistent with previously reported root growth inhibition by exogenous (Bertell and Eliasson, 1992) and endogenous CKs (Werner et al., 2001). In contrast, shoot-localized IPT expression stimulated root biomass (SARK::IPT; Rivero et al., 2007), length, and number (SAG12::IPT; Xu et al., 2009) by up to 3-fold under drought and heat stress, respectively. This is unlikely to be a direct promotion of root growth by additional root CKs, but rather an indirect effect of greater photoassimilate availability in the roots due to delayed canopy senescence.
The stimulation of CK synthesis of HSP70::IPT plants by elevated RZT resulted in concentrations similar to those reported in WT tomato xylem sap (Kudorayova et al., 2007) and IPT-transformed tomato shoots (Groot et al., 1995), but higher (~10-fold) than reported in WT tomato leaves (Walker and Dumbroff, 1981; Rahayu et al., 2005; Kudorayova et al., 2007). Absolute CK concentrations vary according to plant developmental stage, tissue sampled, and plant nutrition. The magnitude of CK accumulation in HSP70::IPT plants depended on salinity treatment (cf. columns 2, 4 in Fig. 2): it was greater in leaves under control than saline conditions (Fig. 2a, b, e, f), similar in both control and saline conditions in xylem sap (Fig. 2c, g), and greater in roots under saline than control conditions due principally to an accumulation of CK ribosides (Fig. 2d, h). Paradoxically, the higher root CK concentrations of salinized HSP70::IPT plants after elevated RZT did not cause additional root growth inhibition (Fig. 5d), perhaps due to the reduced levels of the free forms tZ and iP (Fig. 2d, h). In spite of an apparent doubling of shoot CK concentration (cf. columns 1, 2 of Fig 2a, b, e, f) in non-salinized HSP70::IPT plants, shoot growth was independent of RZT treatment (Fig. 5b, c) either since absolute CK concentrations were not growth limiting or due to negative impacts of elevated RZT. However, additional shoot CKs increased leaf expansion 1.5-fold (Fig. 5b), shoot growth 1.9-fold (Fig. 5c), and maximum efficiency of PSII (Fv/Fm) (Fig. 7a, b) in salinized plants, while only Fv/Fm increased in control plants (Fig. 7a). Indeed, the impacts of root-localized IPT expression on shoot CK concentrations under stress were of a similar magnitude to the improved growth rate. Thus the average 2- to 2.5-fold increase in shoot biomass and relative growth rate (Figs 5c, ,6)6) was associated with a 2- to 2.6-fold increase in leaf CK concentration (cf. columns 3, 4 of Fig. 2a, b, e, f), while the 4.5-fold increase in biomass of droughted tobacco was associated with up to a 5-fold increase in leaf CK concentration (SARK::IPT; Rivero et al., 2007). However, depending on the promoter used and the growing conditions, it generally seems that the physiological impacts of overcoming decreased CK status (via transgenic IPT expression) were most pronounced in plants experiencing abiotic stress.
The attenuated growth inhibition of salinized HSP70::IPT plants exposed to elevated RZT may be via the relative maintenance of cytokinin-mediated cell division and cell wall extensibility (Rayle et al., 1982) and/or improved carbon status due to both delayed stomatal closure and leaf senescence (Fig. 7a–d). However, other hormonal or ionic changes may also be involved. Salinized HSP70::IPT plants exposed to elevated RZT also had lower ABA concentrations in mature leaves, xylem, and roots (cf. columns 3, 4 of Fig. 3b–d, f–h), which may directly delay stomatal closure (Fig. 7c, d) and/or decrease antagonism of CK-dependent stomatal opening (Dodd, 2003), allowing more photosynthesis. A lower ABA concentrations in photosynthetically active leaves may also indirectly delay senescence, photoinhibition, and photooxidation during the osmotic phase of salinity (Ghanem et al., 2008). However, decreased ABA concentration, perhaps mediated by decreased foliar Na+ concentrations (Montero et al., 1998), is unlikely to explain increased growth directly, since ABA-deficient mutants grew less than WT plants when salinized (Mäkelä et al., 2003; Mulholland et al., 2003). Furthermore, leaf ABA concentrations showed limited (up to 1.5-fold) variation with salinity treatment in young (actively growing) HSP70::IPT leaves (Fig. 3a, e), in contrast to the larger variations (up to 3-fold) in leaf CK concentrations, which seem directly responsible for the improved shoot growth of induced HSP70::IPT plants under salinity.
However, increasing shoot levels of bioactive CKs also significantly increased leaf and root K+ (control and saline conditions; cf. Fig. 4g–l) and decreased Na+ concentrations (only under salinity, Fig. 4a–f) of HSP70::IPT plants compared with WT plants. To the best of our knowledge, this is the first demonstration that modulating CK biosynthesis alters K+ and Na+ concentrations, contributing to improved crop salt tolerance. Electrophysiological studies with excised barley roots demonstrated that exogenous (kinetin) application directly increased root cell plasmalemma K+ uptake (Shabala et al., 2009), suggesting a potential mechanism for the positive zeatin–K+ correlation seen in rootstock-mediated improvement of tomato salt tolerance (Albacete et al., 2009). The decreased Na+ concentrations of salinized, induced HSP70::IPT plants (30–40mM less than WT plants and HSP70::IPT plants grown at normal RZT) may minimize or delay the toxic Na+ effect (Ghanem et al., 2008; Munns and Tester, 2008), perhaps due to dilution of Na+ ions by additional growth. This idea is supported by the correspondence of shoot growth improvement and decrease in leaf Na+ concentration (Fig. 6). Although xylem Na+ concentrations were not assayed, a higher delivery of Na+ from root to shoot due to enhanced transpiration (Fig. 7) may explain the decreased root Na+ concentrations (Fig. 4c) despite no additional root growth (Fig. 5d). This may also suggest direct (gene expression) or indirect (membrane potential, energy availability) effects on selective ion transport systems. Indeed, decreased Na+ concentration may occur via selective K+/Na+ uptake, as reported in several isolated tissues (Sastry et al., 1973), and/or increasing the exclusion efficiency of the Na+/H+ antiporters by providing more energetic substrates from the shoot (Malagoli et al., 2008). Although rootstock-influenced leaf xylem CK and Na+ concentrations were not correlated across a range of rootstocks (Albacete et al., 2009), further exploration of the role of CKs in controlling root cell ion transporter expression or activity seems warranted.
Another technique for localizing root IPT expression, that avoided dynamic changes in RZT, was used in longer term experiments. Grafting WT plants onto a constitutively expressing IPT rootstock increased fruit yield (by 30%) compared with salinized WT/WT plants, probably due to increased shoot development (as suggested by the 25% increase in fruit number) and the higher tZ concentrations (1.5- to 2-fold) in the actively growing fruits (Table 2) promoting cell division and expansion, thus slightly increasing (by 5%) fruit weight. In contrast, transient elevated fruit CK concentrations (up to 9-fold) very early in tomato fruit development did not increase final fruit size or total fruit yield per plant in the absence of stress (AGPase::IPT; Luo et al., 2005). However, IPT expression in senescing leaves increased individual tomato fruit weight by 20% (SAG12::IPT; Swartzberg et al., 2006), probably due to an increased sink assimilate import because of delayed source leaf senescence. Shoot growth maintenance (induced HSP70::IPT plants) and increased fruit production (WT/35S::IPT plants) under salinity may result not only from an increased source capacity to produce and/or export more assimilates due to both increased leaf area and delayed senescence, but also from an increased import and/or utilization of assimilates in sink organs achieved by an increased cell population attracting photoassimilates and/or an increased capacity to utilize imported sucrose metabolically (Roitsch and González, 2004). Thus topical (i.e. to the surface) application of exogenous CK (kinetin 10−5M) increased sink activity (radiolabelled sucrose transport) of developing salinized fruits (Albacete, 2009). This idea is also supported by functional approaches that decreased CK concentration and shoot growth of tobacco by overexpressing the CK-degrading enzyme cytokinin oxidase; those plants had lower shoot sink activities due to both decreased cell number in leaf meristems (Werner et al., 2001) and sucrolytic activities (Werner et al., 2008). Similarly, growth inhibition of CK receptor- (Nishimura et al., 2004) and CK-activating enzyme- (Kurakawa et al., 2007) defective mutants was associated with lower meristematic activity, confirming a key role for a threshold CK concentration or sensitivity in regulating plant growth, which seems critical under abiotic stress.
Although induced root IPT gene expression could restrict root growth, increased root-to-shoot CK transport improved salt tolerance by increasing vegetative and fruit growth and also delaying leaf senescence and maintaining stomatal conductance and PSII efficiency, thereby avoiding or delaying the accumulation of toxic ions (Ghanem et al., 2008; Munns and Tester, 2008; Albacete et al., 2009). Expressing IPT in the roots also decreased the (fruit) yield penalty caused by salinity, and may allow a more efficient use of saline water (3–6dS m−1 is typical of aquifers used to abstract water for semi-hydroponic tomato culture in southern Spain). While additional tools to regulate root cytokinin production spatially and temporally, alone, or in combination with other hormones, may provide additional crop improvement possibilities, it will be important to determine whether this approach is valuable in ameliorating salinity-induced growth inhibition of other species grown in other cropping systems.
Supplementary data are available at JXB online.
Figure S1. Total cytokinin (CK) levels in young leaves, leaf 3, and roots of HSP70::IPT and WT tomato plants grown in half-strength Hoagland medium in the absence (Ctrl) or presence (Salt) of 100mM NaCl, and transiently exposed to elevated (RZT) or normal (Norm) root zone temperature.
The authors thank Dr Alejandro Torrecillas, Hana Martínková, and Ondřej Novák for technical assistance. This work was supported by the Fundación Séneca de la Región de Murcia, Spain (project 087121/PI/08), the MICINN, Spain (project AGL2008-01733/AGR), by the Fonds de la Recherche Scientifique F.R.S.-FNRS, Belgium (project FRFC-2456308), and by MEYS, Czech Republic (project MSM619859216).