Introduction

Harpin proteins, produced by Gram-negative plant pathogenic bacteria, belong to an important class of type-III proteins, which elicit multiple plant responses when secreted by the pathogenic bacteria during infection (Alfano and Collmer 2004), artificially applied to plants (Dong et al. 1999; Peng et al. 2003; Pandey et al. 2005; Clarke et al. 2005; Liu et al. 2006) or expressed in transgenic plants (Peng et al. 2004a, b; Jang et al. 2006). Harpins are acidic proteins possessing the nature of glycine abundance, protease sensitivity, and heat stability (Kim and Beer 2000). The characters were recognized early for HrpNEa encoded by the hrpN Ea gene in Erwinia amylovora (Wei et al. 1992) and similar proteins found later in other bacteria (Kim and Beer 2000; Alfano and Collmer 2004; Liu et al. 2006). The common features are believed to be an element that harpins from different bacteria cause a similar battery of pleiotropic effects in a variety of plant species (Dong et al. 1999; Peng et al. 2003; Ren et al. 2006a, b; Reboutier et al. 2007; Wu et al. 2007).

The versatility of harpin application is attributable to the activation of distinct signal transduction pathways. When plants grow under normal moisture conditions, the application of HrpNEa stimulates salicylic acid (SA) to induce pathogen defense (Dong et al. 1999; Peng et al. 2003) and the ethylene signaling pathway to regulate plant growth enhanced concomitantly with insect resistance (Dong et al. 2004). The SA and ethylene signaling processes are independent (Dong et al. 1999, 2004). In Arabidopsis, sensing of ethylene is essential for the induction of plant growth and insect resistance because both effects are arrested in the ethylene-resistant etr1-1 mutant or by inhibiting wild-type (WT) plants to sense ethylene. The ethylene signaling components EIN2 and EIN5 are required because insect resistance and promotion of plant growth are impaired in the ethylene-insensitive ein2-1 and ein5-1 mutants, respectively. When SA-mediated pathogen defense and ethylene signaling for plant growth enhancement and insect resistance are activated, jasmonic acid (JA) signaling is blocked (Dong et al. 2004). When Arabidopsis plants are growing under drought stress conditions, HrpNEa stimulates abscisic acid (ABA) to mediate drought tolerance (Dong et al. 2005). In WT plants, HrpNEa treatment promotes stomatal closure and other adaptive responses, reducing plant drought severity. The process is nullified in the ABA-insensitive abi2-1 mutant, rather than abi1-1, suggesting a requirement for ABI2, instead of ABI1. Nevertheless, either of abi1-1 and abi2-1 compromises a set of ABA-dependent responses in plants growing with water deficit, salinity, and other environmental cues, which often activate both ABI1 and ABI2 (Allen et al. 1999; Wu et al. 2003). Therefore, HrpNEa treatment is distinct from other stimuli in affecting components and pathways of plant signal transduction.

The signaling pathways may be synergistic or antagonistic depending on biological processes and exogenous stimuli that activate the processes in plants (Brocard-Gifford et al. 2003; Traw and Bergelson 2003; Bostock 2005). JA often synergizes ethylene to mediate insect resistance developed in response to insect feeding or environmental hazardous agents (Reymond et al. 2000; Moran and Thompson 2001). In Arabidopsis, JA and ethylene antagonize ABA in signaling of pathogen defense (Anderson et al. 2004; Mauch-Mani and Mauch 2005), and ethylene antagonizes ABA in regulating hyponastic growth (Benschop et al. 2007). However, ABA, ethylene, and several other signals cooperate to regulate seed and root development (Beaudoin et al. 2000; Lu and Hill 2002; Brocard-Gifford et al. 2003; Chiwocha et al. 2005; Iwama et al. 2007). Whereas ethylene, JA, and SA independently respond to HrpNEa in regulating defense and growth in plants growing with normal moisture (Dong et al. 1999, 2004), whether other signals, besides ABA (Dong et al. 2005), are affected in plants growing with drought stress is unclear.

We have sought to characterize interactions between the distinct pathways activated by harpin proteins, beginning with dissection of the relationship between ABA and ethylene signaling activated by HrpNEa. Here, we describe genetic and pharmacological binary analyses combined with molecular and cytological studies, which demonstrate the dispensability of ethylene signaling to the ABA-mediated drought tolerance induced by HrpNEa in Arabidopsis plants growing under drought stress conditions. In response to HrpNEa, drought tolerance is induced even when biosynthesis and sensing of ethylene are disrupted. Thus, ABA acts alone in response to HrpNEa depending on the immediate requirement for plant growth toward drought stress, although multiple pathways can be stimulated in other circumstances (Dong et al. 1999, 2004, 2005; Liu et al. 2006; Ren et al. 2006a; Wu et al. 2007).

Materials and Methods

Plant Growth

Seeds of Arabidopsis ecotypes Columbia (Col-0; seed stock no. CS20) and Landsberg erecta (Ler-0; CS1092), the Col-0 mutant etr1-1 (CS237), and the Ler-0 mutant abi2-1 (CS23) were obtained from Arabidopsis Biological Research Center, Columbus, Ohio http://arabidopsis.org ). Plants used in physical drought assays were grown in 60-ml pots containing a mixture of sand, vermiculite, and potting soil (1:1:1) for 20 days before use. To prepare plants used in physiological drought tests, 20-day seedlings transferred from pots were incubated for another 10 days in 10-cm plastic bottles containing the nutrient solution. Plants were all incubated in controlled environment chambers with a 14-h-day (200 μE/m2 s at 24°C) and 10-h-night (20°C) cycle (Dong and Beer 2000).

HrpNEa Preparation and Plant Treatment

HrpNEa was produced in Escherichia coli strain DH5α harboring pCPP2139, which was made by cloning the hrpN gene in the vector pCPP50 (Bauer et al. 1995; Dong et al. 1999). The empty vector preparation (EVP) that contains inactive proteins but not HrpNEa was produced by the bacteria harboring pCPP50 only (Dong et al. 1999). Concentrations of HrpNEa and proteins in EVP were determined as described (Dong et al. 1999). Based on different responses of Arabidopsis genotypes to various doses of HrpNEa (Peng et al. 2003; Dong et al. 2005), EVP and HrpNEa were used in aqueous solutions at 15 μg/ml, except if specified elsewhere, in the presence of the surfactant Silwet-77 (0.03%), applied by spraying plant tops. In treatments to determine gene expression, plants were sprayed when top buds and two young leaves were protected carefully by an overlay with facial tissue and plastic wrapper.

Chemical treatments included ABA and ethylene biosynthesis or perception inhibitors: amonooxyacetic acid (AOA) (Sigma-Aldrich Inc., Nanjing Agency, Nanjing, China); AgNO3, fluridone, and nordihydroguaiaretic acid (NDGA) (Sigma-Fluka Inc., Nanjing Agency); 1-methylcyclopropene (1-MCP) (Lytone Enterprise Inc., Nanjing Agency). Each inhibitor was applied together with HrpNEa or EVP in aqueous solutions by spraying plant tops. Concentrations were determined based on previous studies (Ghassemian et al. 2000; Hall et al. 2000; Dong et al. 2004). Aqueous solution of 20 μM AgNO3 was prepared freshly before use. Stocks of 10 mM AOA, 0.1 mM fluridone, and 0.1 mM NDGA were made in 95% ethanol, maintained at 4°C, and diluted in water to 1 mM, 30 μM, and 20 μM as final concentrations. Use of water-volatilizable 1-MCP tablets was according to the vender’s protocol. Immediately before treatment, tablets were resolved in water in a small beaker to release gassy 1-MCP into plants growing in pots. The pots were placed together with the beaker in a 12-cm3 glass box sealed immediately. The 1-MCP gas was adjusted to a final concentration of 0.22 μl/l by using proper amounts of the tablets. Plants were incubated in the sealed box for 6 h.

Artificial Drought and Tests of Plant Phenotypic and Cellular Responses

Physiological drought was exerted on hydroponic 30-day plants by adding polyethylene glycol (PEG) 600 (20%, w/v) into the nutrient solution. Drought symptoms were observed every day in 5 days. Physical drought stress was applied to plants growing in pots by withholding water during the experiment. Before that, plants were watered uniformly to 45% soil moisture and 85% moisture content in plant tissues. Subsequently, cell turgidity and water potential in leaves were studied. Drought tolerance was quantified as percent decrease in wilted leaves, relative to controls. Previously described methods (Dong et al. 2005) were used in the assays. Cellular adaptive responses were investigated. Stomata on the lower epidermis of leaves were observed using a microscope equipped with a blue filter. Stomatal apertures were measured with a calibrated optical micrometer. To observe changes in the cell, micro-sectioned leaves were examined with a transmission electron microscope (Hitachi Model E570, Tokyo).

ABA and Ethylene Determination

Endogenous ABA and ethylene levels were tested. Amounts of ethylene gas release from plants were determined by gas chromatography (Guzmán and Ecker 1990) using specific method (Dong et al. 2004). Gas was collected at scheduled time from the environment of seedlings growing in pots, in sealed glass boxes. Ethylene release was quantified as ng/h g fresh plant. ABA was extracted from leaves; its concentration was quantified as described (Li et al. 2006; Dong et al. 2005). For each treatment, 0.5 g leaf sample collected from the fourth to sixth leaves of five plants were ground with liquid nitrogen and homogenized with 3 ml ethanol (80%, v/v), in the presence of 250 ng D3-ABA used as internal standard. Leaf homogenates were maintained overnight at 4°C, followed by centrifuge (10,000×g, 4°C, 10 min). ABA in the supernatants was purified with the ISOLUTE NH2 Solid-Phase-Extraction column (Argonaut Technologies, Inc., Hengoed, UK) and quantified by gas chromatography-mass spectrometry (Li et al. 2006; Peng et al. 2006). ABA level in leaves was given as ng/g fresh weight.

Gene Expression Analysis

RNA was isolated from tops and the two youngest leaves of treated plants as described (Clark 1997; Dong and Beer 2000). Reverse transcriptional polymerase chain reaction (RT-PCR) was done with the M-MLV RT-PCR Kit (GenScript Corp., China Branch, Nanjing, China) used according to the provided protocol. The EF1α gene constitutively expressed in eukaryotes (Gallie et al. 1998) was used as a standard. Primers specific for genes studied and size (bp) of RT-PCR products are as follows: ETR1, 5′-GGAATTCCATATGGAAGTCTGCAATTGTATTGAACC-3′, 5′-CGGGATCCTTACTCTCTAAATAATGTATGAAGATTGA-3′, 1284; EIN2, 5′-GATTCACTGAAGCAGCAGAGGAC-3′, 5′-CTGTGGCAAACTGTAGGCATCTC-3′, 766; CRK5, 5′-CACAATTACAAGACCCTACTTACGT-3′, 5′-ACACAAATAAGAACTGAGATAGCGA-3′, 821; CRK6, 5′-GGGTTCACAAGTTTTGTTTCCTCCAC-3′, 5′-AAGATGATATCGGAGGAGTAGACACCAAC-3′, 355; ABI2, 5′-CCATTAGTGACTCGACCATCAAG-3′, 5′-GTTCTTGTTCTGGCGACGGAGC-3′, 322; PP2CA, 5′-AACGGCAGAAGCGTGAGACAGT-3′, 5′-GCGTGACAACCGATACAACAGC-3′, 534; RD29B, 5′-GTGAAGATGACTATCTCGGTGGTC-3′, 5′-GCCTAACTCTCCGGTGTAACCTAG-3′, 687; EF1α, 5′-AGACCACCAAGTACTACTGCAC-3′, 5′-CCACCAATCTTGTACACATCC-3′, 495. RT-PCR protocols were optimized similarly as described (Peng et al. 2003; Dong et al. 2004). The first-strand cDNA was synthesized with 2 μg RNA that had been treated with RNase-free DNase. An equal volume of cDNA was amplified with specific primers by 25 to 30 cycles, depending on genes. RT-PCR products were cloned, and sequences were confirmed by sequencing and comparison using the Blast Search program. They were visualized by staining with ethidium bromide in agarose gels following electrophoresis.

Sequence data from this article have been deposited with the GenBank data libraries under accession numbers L24119 (ETR1), AF141203 (EIN2), NM_179094 (CRK5), NM_118443 (CRK6), NM_125087 (ABI2), NM_111974 (PP2CA), T04323 (PDF1.2), D13044 (RD29B), AJ223969, AF120093, AF181492, and X97131 (EF1α).

Data Treatment

Results were presented when they were similar in replicate experiments. Numbers of replicates and numbers of plants tested in a replicate are described in figure legends. Quantitative data were analyzed by analysis of variance (ANOVA) test at P = 0.05 to determine significance in differences between treatments (Hoyle 1999).

Results

The Induction of Plant Response to Physiological Drought Stress

Previously, HrpNEa-induced drought tolerance in Arabidopsis was characterized with plants growing in pots and subjected to a physical drought stress applied by depriving water (Dong et al. 2005). Physiological drought stress can induce similar plant responses, which, however, occur faster than responses to physical drought stress (Kocheva et al. 2004). We determined the effect of HrpNEa, vs the inactive protein preparation EVP, on physiological drought of hydroponic plants caused by polyethylene glycol (PEG) supplied to the nutrient solution (Fig. 1a). Compared to the control, PEG-600 applied at 20 μM quickly wilted plants; wilt signs appeared in 24 h posttreatment (hpt) and became aggravated thereafter through 5 days posttreatment (dpt). The application of HrpNEa by spraying plant tops immediately before PEG supplement retarded wilt symptoms to occur 3 days later and markedly reduced wilt severity as well. Thus, tolerance was induced against the physiological drought in HrpNEa-treated hydroponic plants.

Fig. 1
figure 1

Chemical analysis for HrpNEa effects on Arabidopsis plant response to physiological drought stress and production of ethylene and ABA. A. Physiological drought induction. WT (Col-0) plants were growing in a nutrient solution in trays covered with a foam board that supported plants. At 30 days old, plants were sprayed with a solution of the inactive protein preparation EVP, water, or a HrpNEa solution. EVP-treated plants remained having regular growth shown in the panel “EVP”. Immediately after protein application, PEG-600 was added at 20% (w/v) into the nutrient solution to cause physiological drought in plants that had been treated with water and HrpNEa, shown as “PEG” and “HrpNEa+PEG”, respectively. Drought response was monitored at the indicated times. Photos represent 45 plants tested similarly in three replicates. b and c Ethylene and ABA levels. Plants growing similarly as in a were sprayed separately with solutions containing EVP, HrpNEa, the ethylene biosynthesis inhibitor amonooxyacetic acid (AOA), the ethylene perception inhibitor 1-methylcyclopropene (1-MCP), the ABA biosynthesis inhibitor nordihydroguaiaretic acid (NDGA), and combinations of HrpNEa with the inhibitors (HrpNEa + AOA, HrpNEa + 1-MCP, and HrpNEa + NDGA). At the indicated times, ethylene release from plants and ABA content in leaves were determined vs fresh weight (FW) of plant tissues. To collect ethylene, a pot with seedlings had been placed for 6 h in a sealed 12-cm cubic glass box. Curves represent means ± SD of results from three replicates; 15 plants were tested in a replicate

Full size image

ABA Level Increases but Ethylene Changes Little During the Induction of Plant Response to Drought Stress

Ethylene and ABA levels were determined with equivalent plants stressed by PEG and treated differently. We found that ethylene release changed little around the basal level (ca 58 ng/h g) during 36 hpt with EVP or HrpNEa (Fig. 1b). During the same period, ABA levels increased greatly (Fig. 1c) in coincidence with drought tolerance responses (Fig. 1a). Although PEG stress stimulated ABA production, HrpNEa treatment provided the optimal and successive increase in ABA levels (Fig. 1c). Thus, the production of ABA, instead of ethylene, was induced by HrpNEa.

To confirm the different effects of HrpNEa on the production of ethylene and ABA, pharmacological studies were done with AOA and 1-MCP, which inhibit plants to synthesize (Beaudoin et al. 2000) and sense ethylene (Mainardi et al. 2006), respectively, and the ABA-biosynthesis inhibitors fluridone (Ghassemian et al. 2000) and NDGA (Han et al. 2004). We tested effects of the four inhibitors on the production of both hormones in PEG-stressed plants treated with EVP or HrpNEa. Ethylene was affected differently depending on types of inhibitors (Fig. 1b; data not shown). Treatment with 1-MCP, NDGA or fluridone did not change ethylene level; when any of the three inhibitors was present in HrpNEa treatment, ethylene level increased similarly as induced by only HrpNEa. In contrast, the application of AOA caused ethylene decline to the basal level. ABA also was affected differently by the inhibitors (Fig. 1c). Treating plants with only NDGA markedly reduced ABA; applying NDGA together with HrpNEa cancelled a significant part of HrpNEa-induced increase of ABA concentrations (ANOVA test, P < 0.05). Inversely, neither of AOA and 1-MCP affected ABA production; ABA levels were close in control and treatment with AOA or 1-MCP, and close in treatments with only HrpNEa and combinations of HrpNEa and the inhibitors (Fig. 1c; AOA data not shown). Therefore, the inhibition of ABA and ethylene biosynthesis did not mutually affect levels of the hormones. The evident increase in ABA level was caused by HrpNEa treatment.

Expression of most Ethylene and ABA Signaling Genes is Repressed During the Induction of Plant Response to Drought Stress

The genes that are pivotal to ABA and ethylene signal transduction were studied for expression in PEG-stressed plants responding to HrpNEa. In a signal transduction circuit, EIN2 is active in the presence of an ethylene signal (Wang et al. 2002; Guo and Ecker 2004) and is expressed coordinately with ETR1 in HrpNEa-treated Arabidopsis plants (Dong et al. 2004). In this study, however, expression levels of both genes in PEG-stressed plants were evidently depressed with similar patterns after HrpNEa treatment, relative to control (Fig. 2a). The CRK5 and CRK6 genes have been shown to regulate activities of 2C phosphatases involved in plant HCD response to infection by pathogens (Chen et al. 2004). Both genes exhibited constitutive expression but decreased expression levels in response to HrpNEa treatment (Fig. 2a). Then, expression of the genes encoding 2C phospohatases was determined. ABI1, ABI2, and PP2CA are 2C phospohatases involved in several ABA-induced responses crucial to stomatal closures and drought tolerance (Leung et al. 1997; Allen et al. 1999; Wu et al. 2003; Dong et al. 2005; Kuhn et al. 2006; Yoshida et al. 2006). The expression of ABI2 was greatly increased but that of ABI1 was not induced in drought-stressed plants during the course of time after treatment with HpNEa (Fig. 2a; data not shown), consistent with our previous finding that ABI1 did not respond to HrpNEa treatment (Dong et al. 2005). During the same period, PP2CA was expressed at a great level similar as constitutive expression, suggesting that its transcription was not affected by HrpNEa. Of the six genes tested, only ABI2 was upregulated coincidently with ABA elevation and drought tolerance induction by HrpNEa (Fig. 2a).

Fig. 2
figure 2

Expression of the genes critically involved in ethylene and ABA signal transduction in drought-stressed and HrpNEa-treated plants. a Expression of the signaling regulatory genes in Col-0 WT plants. Plants were grown and treated similarly as in Fig. 1a. RT-PCR was done using EF1α as a standard to analyze expression of ETR1 and EIN2 genes, which critically regulate ethylene signaling, and CRK5, CRK6, ABI2, and PP2CA, which pivotally regulate ABA signaling. RNA was isolated from untreated buds and two top leaves of plants at 48 h after treatment with EVP (CK) or at the indicated times of treatment with HrpNEa. Experiments were done three times with similar results. b Comparison of induced expression of the drought stress response gene RD29B in the specific genotypes of the plant. Experiments were similar to those in Fig. 1a. Expression of the gene in untreated top leaves of plants growing with physiological drought stress was determined by RT-PCR at 2 days after treatment with EVP (CK), zero, and 2 days after treatment with HrpNEa

Full size image

Induced Expression of the Drought Stress Response Gene RD29B Depends on Sensing of ABA but not Ethylene

We have shown that the expression of RD29B, a molecular marker of the ABA signal transduction pathway (Hua et al. 2006), is induced by HrpNEa in plants growing under physical drought stress conditions (Dong et al. 2005). To determine if signaling by both ABA and ethylene is engaged in the induction of drought tolerance, we studied HrpNEa effects on RD29B expression in abi1-1, abi2-1, etr1-1, ein2-1, ein5-1, and WT plants growing under drought stress conditions. When WT plants were subjected to the physiological drought stress caused by PEG, RD29B expression was induced by HrpNEa treatment (Fig. 2b, Ler-0 and Col-0) compared to low levels of constitutive expression (Fig. 2b, CK). When plants were treated with HrpNEa and PEG, RD29B expression was conspicuously compromised in abi2-1 compared to WT (Ler-0), as tested at 0 and 2 dpt (Fig. 2a). When abi2-1 and WT plants were growing with physical drought stress and treated with HrpNEa, the gene behaved similarly as it did in response to the physiological stress (data not shown). This result confirms the critical role ABI2 plays in HrpNEa signaling during drought tolerance development (Dong et al. 2005). By contrary, the genetic blocking in ethylene perception and signal transduction, as represented by etr1-1, ein2-1, and ein5-1, did not affect RD29B expression. No matter if plants were growing under conditions of drought stress made physiologically or physically, RD29B behaved similarly in plants of WT (Col-0) and the three mutants (Fig. 2b; ein5-1 data not shown). When determined at 2 dpt with PEG-stressed plants, the gene was strongly expressed in both WT and etr1-1 plants treated with HrpNEa, vs EVP. A great level of RD29B expression also was observed in ein5-1 plants growing with physical or physiological drought stress and after treatment with HrpNEa vs the control (data not shown). This result suggests that the drought stress response gene is activated without requirement for transduction of an ethylene signal perception and transduction, which, however, are required for the induction of plant growth and insect resistance (Dong et al. 2004). Thus, the ABA signaling pathway can be activated by HrpNEa in plants growing under drought stress conditions no matter whether plants have the ability to sense and transduce the ethylene signal.

Cytological Adaptive Responses are Induced in the Ethylene-Insensitive Mutant

As important adaptations in cytology, stomatal closure and maintenance of cellular integrity are an essential part of drought tolerance development (Dong et al. 2005). When hydroponic plants were growing with the physiological drought stress and sprayed with a solution of HrpNEa or EVP, microscopic observations of leaves excised at 24 hpt indicated that stomatal closure was promoted by HrpNEa, vs EVP, on WT and etr1-1. On leaves of both genotypes, stomata were evidently open in EVP treatment but largely closed after the application of HrpNEa; stomatal openings were measured as 2.2 and 0.2 μM in WT and etr1-1, respectively (Fig. 3a, left). Based on stomatal apertures determined at three intervals after the physiological drought stress (Fig. 3a, right), HrpNEa vs EVP caused 67 and 88% more closure of Col-0 stomata in 24 hpt; equivalently, the percentages in etr1-1 were 65 and 85%. Moreover, Fig. 3b shows that drought-caused damage to the cell was alleviated by HrpNEa in etr1-1 similarly as in WT. When observed at 24 hpt, cell membranes were undermined and appeared quite degraded and were not distinguishable from cell walls, which also were destroyed, in control plants. Inversely, membranes and walls appeared intact in HrpNEa-treated plants of WT and etr1-1. In both genotypes, the application of HrpNEa effectively revived leaves in turgidity and water potential when leaf cell membranes and walls of control plants had been dilapidated, as observed at 96 h after drought stress. Similar cytological responses were observed in WT and etr1-1 plants growing under physical drought stress conditions (data not shown). Membranes seemed healthy in 10 days after physical drought stress. Thereafter, until 20 days, membranes were undermined and appeared quite degraded and were not distinguishable from cell walls, which also were destroyed, in control plants. In contrast, cell membranes and walls appeared intact in HrpNEa-treated plants of WT and etr1-1. These plants also kept integrity of organelles like chloroplasts, which, however, disappeared in control plants. Therefore, the defect in sensing of ethylene does not affect the cellular adaptive responses induced by HrpNEa.

Fig. 3
figure 3

Comparison of Col-0 WT and etr1-1 plants in cellular adaptive responses and drought symptoms as affected by HrpNEa. a to c Responses to the physiological drought stress. Plants were grown in trays with a nutrient solution. The physiological drought stress was caused by PEG applied similarly as in Fig. 1a. At 24 hpt, 45 plants in three replicates were tested. In a, lower epidermal cells of excised leaves were observed by microscopy. Apertures of 500−600 stomata were measured and given as means ± SD. In b, Cells in leaf sections were observed by electronic microscopy with similar results in three replicates. Cell membranes (CM) and walls (CW) or CM areas (CMA) and CW areas (CWA) are indicated. In c, drought symptoms were observed at 2 days after physiological drought stress. d Comparison of Col-0 WT and etr1-1 plants in response to physical drought stress. Drought stress was applied by withholding water from 10-day plants growing in pots. Ten days later, plants were sprayed with a solution of EVP or HrpNEa and assayed at the indicated times. Photos were taken at 15 days after physical drought stresses. Symptom severity was scored as rate of wilted leaves; histograms depict means ± SD bars of results from three replicates; 15 plants were tested in each replicate

Full size image

Drought Tolerance Phenotype is Induced in the Ethylene-Insensitive Mutant

Drought symptoms of plants were investigated. In 48 h after PEG was added to the nutrient solution, hydroponic plants of both WT and etr1-1 became wilted when plant tops were sprayed with a EVP solution, but equivalent plants were growing vigorously when HrpNEa was applied similarly (Fig. 3c). In plants growing in pots, symptoms appeared as leaf wilting and darkening (Fig. 3d, left) were observed at 15 days after physical stress (Fig. 3d, right). In both WT and etr1-1 plants, symptoms were not evident until 10 dpt with HrpNEa, compared to the severe leaf wilting and darkening of control plants the same day (Fig. 3d, left). By 20 and 25 days of physical drought (10 and 15 dpt), control plants of both WT and etr1-1 had 83% and 96% leaves wilted (Fig. 3d, right). Relatively, the application of HrpNEa to WT and etr1-1 plants resulted in 60 to 65% fewer leaves wilted during the period. Thus, the defect in sensing of ethylene does not impair the induction of drought tolerance by HrpNEa.

Chemical Blocking in Biosynthesis and Sensing of Ethylene does not Affect Drought Tolerance Responses

Pharmacological analysis was applied in parallel to the specific WT and mutant plants growing in pots under physical drought condition. Stomata were observed at 15 days after drought stress; results are presented in Fig. 4a. Treating Col-0 and Ler-0 plants with HrpNEa caused ca 60% more closure of stomata, compared to control. Fluridone and NDGA either greatly reduced the effect but 1-MCP or AgNO3 did not. Consistently, stomata of HrpNEa-treated etr1-1 plants closed to an extent greater than that in control; NDGA but not 1-MCP cancelled the effect. So etr1-1 behaved similarly as did WT in the response. Nevertheless, HrpNEa failed to stimulate closure of abi2-1 stomata, which remained open regardless of treatments. Thus, neither of the genetic and chemical compromises in ethylene perception impaired ABA signaling critical to the cellular adaptive responses. As a result, drought tolerance phenotype varied with genotypes and treatments (Fig. 4b). Treating WT and etr1-1 plants with HrpNEa vs EVP resulted in ca 60% less leaves wilted; NDGA but not 1-MCP nullified the effect. Leaves of abi2-1, however, wilted at lose rates irrespective of treatments. Clearly, blocking in ABA signaling through ABI2, instead of defects in biosynthesis and sensing of ethylene, compromises the induction of drought tolerance.

Fig. 4
figure 4

Chemical and genetic analyses for the effect of HrpNEa on plant response to the physical drought stress. a Stomatal closure qualification. b Percentage of wilted leaves. Plants were grown similarly as in Fig. 3d and sprayed at 10 days after drought stress with solutions of indicated chemicals, prefixed with the symbol “+” when applied together with HrpNEa. Fld refers to fluridone, an inhibitor of ABA biosynthesis; AgNO3 inhibits plants to sense ABA. Stomata apertures of 150−200 stomata per plant were determined at 5 dpt. Wilted leaves was scored at 10 dpt. Histograms represent means ± SD of results from four replicates; 25 plants were tested in each replicate

Full size image

Discussion

With an attempt to test if ABA and ethylene interact in the induction of Arabidopsis drought tolerance by HrpNEa, this study obtained results recalcitrant to our original idea of a synergistic ABA–ethylene cooperation during the process. Genetic and chemical binary analyses combined with molecular and cytological studies have demonstrated that ABA signaling regulates the drought tolerance induction without requirement for an ethylene signal in drought-stressed plants.

Several lines of evidence suggest that ethylene is not likely to play a role in the induction of drought tolerance. Physiological and physical droughts trigger a similar set of plant adaptive responses in physiology and cytology. The increase of hydrophilic proline concentration favors osmotic homeostasis in the cytosol (Khedr et al. 2003; Dong et al. 2005; Verslues and Bray 2006). Stomatal closure, which reduces water loss through transpiration, occurs in guard cells in response to Ca2+ oscillation caused by ion fluxes through functional membranes (Assmann 2003). Maintenance of cellular integrity while undermined by water deficiency is a prerequisite for the cellular adaptations to occur as essential parts of drought tolerance development (Dong et al. 2005). As described in this study, the application of HrpNEa to plants of WT and ethylene-insensitive mutants intensifies these responses and retards plant drought symptoms. In drought-stressed WT plants, HrpNEa does not cause evident changes in ethylene contents compared to the basal level. The expression of the ABA and ethylene signaling genes CRK5, CRK6, ETR1, and EIN2 (Alonso et al. 1999; Gamble et al. 2002; Chen et al. 2004, 2005) are all downregulated in contrast to the marked increase in ABI2 expression. However, HrpNEa activates the ABA-signaling effector gene RD29B and cellular adaptive responses equally well in WT and mutants that have defects in sensing and action of ethylene, conferring drought tolerance phenotype to the plant genotypes. This result suggests that the ethylene insensitivity does not impair the induction of drought tolerance by HrpNEa. In both WT, etr1-1, ein2-1, and eni5-1 plants treated with HrpNEa, stomata close and integrity of cells is sustained during the course of drought stress, resulting in drought severity reduced by similar extents. These effects are arrested by abi2-1 instead of other mutants; these effects also are nullified by inhibiting WT and etr1-1 plants to synthesize or sense ABA, rather than ethylene. Clearly, ethylene does not act together with ABA to mediate drought tolerance development in drought-stressed plants responding to HrpNEa.

Behaviors of the ABA and ethylene signaling genes in response to HrpNEa prioritize the role that ABI2 plays during the induction of drought tolerance. In HrpNEa-treated and drought-stressed plants, the expression of RD29B is implicated in a physiological process in which RD29B acts to increase cell hydration and turgidity under water shortage conditions (Hua et al. 2006); part of these reactions have been demonstrated as a function of HrpNEa (Dong et al. 2005). Effector gene expression is regulated by a number of signaling components. Pivotally, several protein kinases have been shown to regulate early ABA signaling (Osakabe et al. 2005; Iwama et al. 2007). Several identified cysteine-rich protein kinases (such as CRK5 and CRK6) and serine/threonine protein kinases (like salt overly sensitive, SOS) can interact with kinase-associated 2C phospohatases (such as ABI1, ABI2, and PP2CA) to differentially regulate plant responses (Meyer et al. 1994; Merlot et al. 2001; Chen et al. 2004; Kuhn et al. 2006; Yoshida et al. 2006; Zhu et al. 2007). SOS2 and SOS3 regulate plant response to salinity stress by interacting with 2C phospohatases (Ohta et al. 2003; Zhu et al. 2007). In particular, it has been shown that SOS2 interaction with ABI1 is weak but the interaction with ABI2 is much stronger (Ohta et al. 2003). These findings may explain the preferential activation of ABI2 over AtPP2CA, CRK5, CRK6, ETR1, and EIN2 in drought-stressed plants while responding to HrpNEa. Whether these genes play any roles in plant responses to HrpNEa and how HrpNEa preferentially recruits ABI2 into drought tolerance development remain to be studied.

We have shown that the induction of ABA-mediated drought tolerance is independent of ethylene signaling in plants growing under drought stress conditions. When plants grow under normal moisture conditions; however, HrpNEa treatment can activate other pathways to regulate plant growth (Dong et al. 2004; Wu et al. 2007) and defenses against pathogens (Dong et al. 1999; Peng et al. 2003; Ren et al. 2006a, b) and insects (Dong et al. 2004). Therefore, plants conform to the real-time requirement for growth by prioritizing a pathway over others in response to HrpNEa.