• Open Access

Increased expression of phospholipase Dα1 in guard cells decreases water loss with improved seed production under drought in Brassica napus

Authors

  • Shaoping Lu,

    1. National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Sung C. Bahn,

    1. Department of Biology, University of Missouri, St. Louis, MO, USA
    2. Donald Danforth Plant Science Center, St. Louis, MO, USA
    Search for more papers by this author
  • Gang Qu,

    1. National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Haiying Qin,

    1. National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Yue Hong,

    1. National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Qingpeng Xu,

    1. National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Yongming Zhou,

    1. National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Yueyun Hong,

    Corresponding author
    • National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author
  • Xuemin Wang

    Corresponding author
    1. Department of Biology, University of Missouri, St. Louis, MO, USA
    2. Donald Danforth Plant Science Center, St. Louis, MO, USA
    • National Key Laboratory of Crop Genetic Improvement, College of Life Sciences and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author

Correspondence (fax (314) 587 1519; email swang@danforthcenter.org; hongyhk@yahoo.com.cn)

Summary

The activation of phospholipase Dα1 (PLDα1) produces lipid messenger phosphatidic acid and promotes stomatal closure in Arabidopsis. To explore the use of the PLDα1-mediated signalling towards decreasing water loss in crop plants, we introduced Arabidopsis PLDα1 under the control of a guard cell–specific promoter AtKatIpro into two canola (Brassica napus) cultivars. Multiple AtKatIpro::PLDα1 lines in each cultivar displayed decreased water loss and improved biomass accumulation under hyperosmotic stress conditions, including drought and high salinity. Moreover, AtKatIpro::PLDα1 plants produced more seeds than did WT plants in fields under drought. The results indicate that the guard cell–specific expression of PLDα1 has the potential to improve crop yield by enhancing drought tolerance.

Introduction

Currently, agriculture is responsible for 70% of total freshwater consumption, the majority of which is used for crop production (Condon et al., 2004; Hamdy et al., 2003; Wang et al., 2002). The percentage of water usage is expected to grow as demands in agricultural production increase, whereas the supply of fresh water decreases. There is a tremendous need and interest in developing plants with increased drought tolerance and water-use efficiency. Torrential plants lose most of water through the aperture of stomata (Schroeder et al., 2001), and a number of mutants that have impaired stomatal closure are reported to lose more water than wild type (Klein et al., 2003; Sang et al., 2001; Sasaki et al., 2010). Thus, reducing transpirational water loss by modulating the movement of stomatal aperture has been one of the strategies to improve plant water-use efficiency and drought tolerance (Davies et al., 2002; Kim et al., 2012; Liu et al., 2005; Wang et al., 2009).

Stomatal aperture is tightly controlled via a suite of effectors, including abscisic acid (ABA), reactive oxygen species (ROS) and lipid mediators. Lipid messengers, such as phosphatidic acid (PA), have been shown to promote stomatal closure in Arabidopsis and other plant species (Hong et al., 2008a; Jacob et al., 1999; Sang et al., 2001). PA binds to ABA-insensitive 1 (ABI1), a protein phosphatase 2C, and tethers ABI to the plasma membrane to inhibit the negative effect of ABI1 on ABA signalling, thus promoting the ABA effect on stomatal closure (Mishra et al., 2006; Zhang et al., 2004). PA also binds to NADPH oxidase to increase the production of ROS or NO (Zhang et al., 2009). Recent results show that PA interacts directly with phytosphingosine kinase to promote the synthesis of phytosphingosine phosphate that promotes stomatal closure (Guo et al., 2011, 2012b). Plants deficient in drought- or ABA-induced production of PA lose more water and exhibit less tolerance to drought (Hong et al., 2008b; Sang et al., 2001). Thus, increasing PA production in plants may be a novel avenue to decrease plant water loss and tolerance to drought.

Phospholipase D (PLD) is the major family of enzymes that hydrolyse phospholipids to produce the signalling lipid PA (Wang et al., 2006). The Arabidopsis genome contains 12 PLD genes that have been grouped: PLDα(1,2,3), PLDβ(1,2), PLDγ(1,2,3), PLDδ, PLDε and PLDζ(1,2) (Qin and Wang, 2002; Wang et al., 2006). Ten of the 12 PLDs have the Ca2+-dependent phospholipid-binding C2 domain, whereas two PLDζs have pleckstrin homology (PH) and phox homology (PX) domains (Qin and Wang, 2002). The different groups of PLDs have different requirements for the effectors of activity, such as Ca2+, free fatty acid and phosphoinositides. PLDs also differ in substrate preference (Hong et al., 2010). These biochemical differences, together with their distinguishable patterns in expression and subcellular localization, may provide a basis for the distinguishable functions for different PLDs in plant response to drought (Sang et al., 2001), salinity (Hong et al., 2008b; Yu et al., 2010), nitrogen deprivation (Hong et al., 2009) and phosphate deficiency (Cruz-Ramírez et al., 2006; Li et al., 2006a,b). Both PLDα1 and PLDδ play a role in mediating plant response to ABA and promote stomatal closure, but they occupy different steps in the signalling pathway (Guo et al., 2012a). PLDδ is involved in ROS response (Zhang et al., 2003), whereas PLDα1 acts upstream of PLDδ and promotes ROS production (Zhang et al., 2009).

Activation of PLDα1 has been shown to promote stomatal closure, and suppression of PLDα1 expression decreases transpirational water loss in Arabidopsis (Mishra et al., 2006; Sang et al., 2001; Wang et al., 2001; Zhang et al., 2004). An attempt has been made to test whether overexpression of PLDα1 driven by the constitutive 35S promoter in tobacco can enhance drought tolerance in crop plants (Hong et al., 2008a). The constitutively increased expression of PLDα1 promotes stomatal closure at earlier stages of drought, but disrupts membranes in prolonged drought stress in tobacco (Hong et al., 2008a). High PLD activity has also been associated with drought susceptibility in other crop plants (Guo et al., 2006; Maarouf et al., 1999). To explore the application of PLDα1 signalling to improve plant drought tolerance without its detrimental effect on membrane degradation by a high PLDα1 activity in all tissues, here we placed PLDα1 under the control of a guard cell–specific promoter AtKatIpro::PLDα1 and introduced the construct into a spring cultivar and a semi-winter cultivar of canola. The Arabidopsis Kat1 encodes a voltage-gated inward-rectifying K+ channel and is expressed primarily in guard cells (Nakamura et al., 1995). Testing of the AtKatIpro::PLDα1 plants shows that plants display reduced water loss and increase drought tolerance in Brassica napus.

Results

Expression of AtPLDα1 in guard cells of Brassica napus

To express PLDα1 in guard cells, we placed AtPLDα1 under the control of a guard cell–specific promoter AtKatIpro (Figure 1a). AtKatIpro::PLDα1 was introduced into two canola cultivars, spring cultivar Westar and semi-winter cultivar Zhongshuang 9, via agrobacterium-mediated transformation. The expression of AtKatIpro::PLDα1 was verified by real-time PCR using RNA from canola leaves and AtPLDα1-specific primers, and AtPLDα1 was detected in OE3-3 and OE7-1 plants (Figure 1b). To detect changes in PLDα1 protein, we used abaxial epidermal peels of leaves to isolate protein because AtKatIpro::PLDα1 was expressed in guard cells. PLD antibodies raised against the terminal regions of Arabidopsis PLDα1 were used, and they react also with native PLDα1 in rapeseeds (Figure 1c). However, when the same amounts of proteins from the epidermal peels of WT and AtKatIpro::PLDα1 leaves were immunoblotted, these transgenic lines showed a higher level of PLDα1 protein than those of WT (Figure 1c and Figure S1).

Figure 1.

Expression of AtPLDα1 under a guard cell promoter in Brassica napus (cv Westar). (a) Transgene casette that contains AtKatIpro::PLDα1 in the binary vector pCAMBIA1300. (b) AtPLDα1 expression level by real-time PCR. (c) The protein expression in PLDα1-OE plants by immunoblotting. The protein was extracted from abaxial epidermal peels. The same amounts of proteins (2.5 μg/lane) were separated by 8% SDS-PAGE and transferred to a polyvinylidene difluoride membrane. PLDα1 was detected by PLDα1 antibodies. (d) Immuno-histochemical detection of PLDα1 in guard cell of abaxial epidermal peels. Red precipitates indicate the presence of PLDα1 as detected by PLDα1 antibodies that react with Arabidopsis and canola PLDα1. The experiment repeated twice. Bars = 5 μm. PLD, phospholipase D.

To confirm the AtPLDα1 expression in guard cells, leaf epidermal peels were immunoblotted in situ with anti-PLDα1 antibodies. The guard cells of AtPLDα1-OE stomata displayed more intensive staining than those of WT (Figure 1d). Guard cells of WT peels were also stained because of the cross-reactivity of the antibodies. The surrounding mesophyll cells were stained as BnPLDα1 in these cells could cross-reacted with the PLDα1 antibodies. Taken together, the results indicate that AtPLDα1 gene is expressed in Brassica napus and there is more PLDα1 protein present in guard cells of AtPLDα1-OE plant leaves than WT. Homozygous lines of both Westar and Zhongshuang 9 were used for further studies.

Decreased water loss in PLDα1-OE plants

To investigate the effect of AtPLDα1-OE on plant water loss, AtKatIpro::PLDα1 plants and corresponding Westar and Zhongshuang 9 WT were grown in soil in a growth room. The size of stomatal aperture of WT and PLDα1-OE leaves was comparable when the stomata were induced open (Figure 2a). However, after treatment with 50 μm ABA for 30 min, the stomatal aperture in PLDα1-OE leaves was smaller than that of WT (Figure 2a). When fully expanded leaves with similar size, age and position on plants at the six leaf stages were removed and weighed every 10 min to determine the rate of water loss, PLDα1-OE leaves displayed a lower rate of water loss than those of WT (Figure 2b). After 160 min, leaves from WT plants lost 30% more water than those from PLDα1-OE plants (Figure 2b). When water was withheld from 2-month-old plants, leaves of WT plants lost turgor faster than PLDα1-OE plants (Figure 2c,d). Relative water content of leaves from the PLDα1-OE plants was approximately 3% higher than that of WT under the drought condition (Figure 2e). Under well-watered conditions, no difference in relative water content was observed between PLDα1-OE and WT leaves (Figure S2a). To assess leaf damage under drought, we measured ionic leakage after drought treatments. Without drought, WT and PLDα1-OE leaves had comparable relative ionic conductivity (Figure S2b), but PLDα1-OE leaves had 30%–50% lower ionic leakage than WT leaves after drought treatments (Figure 2f).

Figure 2.

Decreased water loss in AtKatIpro::PLDα1 plants. (a) Stomatal aperture of AtKatIpro::PLDα1 and WT Westar plants and its response to abscisic acid (ABA). Leaf discs (0.5 cm2) from 3-week-old plants were incubated in a buffer under light for 1 h to induce stomatal opening, and then 50 μm ABA was added and incubated for 30 min. One hundred to 150 stomata were measured. Values are means ± SD. (b) Water loss of detached leaves of Westar cultivar. Water loss is expressed as the percentage of initial fresh leaves weight. Values are means ± SD (= 3 in four independent experiments). (c) Plant phenotype of AtKatIpro::PLDα1 and WT Westar plants after drought treatments. Water was withheld from 60-day-old plants for 5 days. The experiments were repeated three times, and representative phenotypes were shown. (d) Plant phenotype of AtKatIpro::PLDα1 and WT Zhongshuang 9 plants after drought treatments as in (c). (e) Relative water content in Westar WT and transgenic leaves after drought treatments as in (c). (f) Relative ionic conductivity of Westar WT and transgenic leaves after drought treatments as in (c). Values are means ± SD (= 3 in three independent experiments). * and ** denote significance at P < 0.05 and P < 0.01, respectively, compared with WT based on Student's t test. PLD, phospholipase D.

We further examined the expression of PLDα1 and several genes involved in drought responses. Homozygous transgenic and WT Westar plants were grown in soil for approximately 60 days under normal watering condition in a growth room, and then irrigation was halted for 5 days. Under the well-watered condition, the expression levels of these genes were comparable between WT and PLDα1-OE plants (Figure 3, upper panels). However, under drought, the expression levels for the ABA- and drought stress-inducible genes, including canola BnPLDα1, ABI1, ABI2, RAB18, RD29B and ERA1, were expressed 20%–90% less in PLDα1-OE plants than that of WT (Figure 3, lower panels).

Figure 3.

Expression of PLDα1 and drought-inducible genes in Westar WT and AtKatIpro::PLDα1 plants with and without drought stress. Water was withheld from 60-day-old plants for 5 days, RNA was extracted from leaves, and DNA-free RNA was used for real-time PCR. The expression levels were normalized to BnActin. Values are means ± SD (= 3). PLD, phospholipase D.

Impacts of PLDα1-OE on plant flowering and growth in fields

To test the effect of PLDα1-OE in the field, seeds of homozygous transgenic OE lines and WT plants were sowed in the field with a randomized field plot design using three plots for WT and each PLDa1-OE lines. The rainfall in that grown season was limited, and the soil water content was dropped to 12.8% (clay soil with the full water-holding capacity at 58% soil water content). At this water deficit condition, AtKatIpro::PLDα1 plants displayed higher relative water content and a tendency of lower ionic leakage than WT (Figure 4a,b). AtKatIpro::PLDα1 plants showed lower stomatal conductance and less water transpiration than WT plants (Figure 4c,d).

Figure 4.

Water retention and transpiration loss for Westar WT and AtKatIpro::PLDα1 plants in the field. (a) and (b) Relative water content and relative ionic conductivity, respectively. The leaf discs were sampled from Westar WT and AtKatIpro::PLDα1 plants in fields when the soil water content was at 12.8% (clay soil with the full water-holding capacity at 58% soil water content). Values are means ± SD (= 3). Stomatal conductance (c) and water transpiration (d) were measured from 58-day-old plants in the field. The leaf areas (6.5 cm2) of similar positions were measured at 10:00–11:00 a.m. under 101.3 kPa atmospheric pressure, 430–460 μmol/m2s light intensity, at 25.6–26.3 °C, and 40%–41% humidity. Values are mean ± SD (= 6–10). * and ** denote significance at P < 0.05 and P < 0.01, respectively, compared with WT based on Student's t test. PLD, phospholipase D.

When Westar plants grew up to about 145 days and entered the flowering stage, we recorded the percentage of flowering plants in the population of each line as flowering rate. Compared with WT plant population, the number of flowering PLDα1-OE plants was 20%–70% fewer in Westar (Figure 5a). The flowering time of PLDα1-OE plants was delayed 3–6 days from that of WT in Westar. It took more time for Zhongshuang 9 to enter the flowing stage. The percentage of flowering plants was delayed 2–4 days in one OE line from that of WT Zhongshuang 9 plants (Figure 5b). The plant height was measured every 2 days from the time plants started bolting, and PLDα1-OE plants in both cultivars grew faster than that of WT (Figure 5c,d).

Figure 5.

Plant growth and development for Westar WT and AtKatIpro::PLDα1 plants in the field. (a) Percentage of flowering plants of AtKatIpro::PLDα1 and WT Westar plants after 145-day growth. (b) Percentage of flowering plants of AtKatIpro::PLDα1 and WT Zhongshuang plants after 160-day growth. Values are means ± SD (= 3). (c) Plant growth indicated by plant height of Westar plants from days 145 to 159. Values are means ± SD (= 11–14, = 3; (d) Plant growth indicated by plant height of Zhongshuang 9 plants from days 157 to 175. Values are means ± SD (= 8–10, = 3). PLD, phospholipase D.

WT and PLDα1-OE plants were harvested from the field after seed maturation for further measurements. The height of PLDα1-OE plants in Westar was 5, 21 and 39% greater in the lines of OE3-3, OE5-4 and OE7-1, respectively, as compared to WT plants (Figure 6a), but only one line OE5-4 exhibited longer inflorescence significantly (i.e. stem bearing flowering and siliques; Figure 6a). In addition, OE7-1 and OE3-3 have more productive branches (branches bearing seeds per plant; Figure 6c). In Zhongshuang 9, PLDα1-OE plants were approximately 5% higher (Figure 6b) and 7% longer in primary inflorescence (Figure 6b), and had 10% more productive branches (Figure 6c) than WT. The silique number of primary inflorescence (Figure S3a,c) and silique length (Figure S3b,d) were similar between PLDα1-OE and WT plants in both cultivars.

Figure 6.

Enhanced agronomic traits by PLDα1 overexpression in guard cells. (a) and (b) Plant height and primary inflorescence length of WT and AtKatIpro::PLDα1 plants in Westar and Zhongshuang 9, respectively. (c) and (d) Number of effective branches (seed-bearing) OE and WT plants of Westar and Zhongshuang 9, respectively. Values are means ± SD (= 6–10, = 3 for Westar and = 6–9, = 3 for Zhongshuang 9). Bars with different letters denote significant difference at < 0.05, based on ANOVA analysis and Duncan's multiple range test. PLD, phospholipase D.

Enhanced seed production in PLDα1-OE plants in fields

The seed yield per plant was measured in harvested plants grown in the field. There was no significant difference in the seed number per silique (Figure S4a,c) or the weight per 1000 seeds (Figure S4b,d) between PLDα1-OE and WT plants in both cultivars. But the silique number per plant was increased by approximately 15% in PLDα1-OE lines of Westar and by 20%–50% in PLDα1-OE lines of Zhongshuang 9 (Figure 7a,c). The total seed yield per plant was 20%–50% greater in Westar PLDα1-OE lines and approximately 50% more in Zhongshuang 9 PLDα1-OE lines than corresponding WT (Figure 7b,d). PLDa1-OE plants produce more seeds than WT by increasing flowering stems and silique number (Figure 6a,b). In some lines, PLDα1-OE plants also have more productive branches than that of WT (Figure 6c,d).

Figure 7.

Seed production of AtPLDα1-OE and WT plants in the field. (a) and (b) Silique number and seed yield per plants, respectively, for Westar transgenic and WT plants. (c) and (d) Silique number and seed yield per plants, respectively, for Zhongshuang 9 transgenic and WT plants. Values are means ± SD (= 6–10, = 3 for Westar and = 6–9, = 3 for Zhongshuang 9). Bars with different letters denote significant difference at < 0.05, based on ANOVA analysis and Duncan's multiple range test. PLD, phospholipase D.

Seed oil is the most important economic trait of canola production. The oil content of two PLDα1-OE seeds in Zhongshuang 9 was similar to that of WT, whereas the OE13-24 plants had lower oil content than that of WT (Figure 8a, inset). However, the total oil content of two PLDα1-OE Westar lines displayed an increase of 9%, but another line had similar oil content as WT (Figure 8b, inset). In both cultivars, PLDα1-OE seeds accumulated more monounsaturated oleic acid (C18:1), which accounted for the increase in oil content in Westar (Figure 8b). In Zhongshuang 9, the increase in oleic acid was accompanied by a decrease in polyunsaturated α-linolenic acid (C18:3; Figure 8a).

Figure 8.

Seed oil content and fatty acid composition of PLDα1-OE and WT plants. (a) and (b) Seed oil content and fatty acid composition of Zhongshuang 9 and Westar cultivars, respectively. Total oil was extracted from the same number of seeds and detected by GC. Values are means ± SD (= 3). * and ** denote significance at P < 0.05 and P < 0.01, respectively, compared with WT based on Student's t test. PLD, phospholipase D.

Effect of PLDα1-OE on seedling growth in response to ABA, NaCl and drought

To examine how the AtKatIpro::PLDα1 transgenic canola plants perform under hyperosmotic conditions, 13-day-old seedlings of homozygous AtKatIpro::PLDα1 and WT were transferred to 1/2 MS liquid media containing 100 mm NaCl, 10% PEG6000 or 50 μm ABA. The medium was replaced with fresh solutions containing the same stressor every week, and the biomass of plants was measured 4 weeks after the initial treatment. Among the stressors, salt stress (NaCl) inhibited plant growth the most (Figure 9), decreasing biomass accumulation of WT plants by 85%, whereas ABA and PEG6000 decreased biomass by approximately 45%, as compared to the normal growth condition (Figure 9a). All three PLDα1-OE lines had higher dry and fresh weights than WT under the control, 50 μm ABA, 100 mm NaCl and 10% PEG grown media (Figure 9a,b). In particular, PLDα1-OE plants accumulated more than twice biomass than WT at 100 mm NaCl (Figure 9a,b).

Figure 9.

Biomass accumulation in roots, leaves and whole plant of AtPLDα1-OE and WT plants of Zhongshuang 9 cultivar. Seeds were germinated on water for 13 days, and then the seedlings were transferred into liquid media with 0, 50 μm abscisic acid (ABA), 100 mM NaCl or 10% PEG6000 for 30 days. Liquid medium was replaced every week. Values are means ± SD (n = 10–20). * and ** denote significance at P < 0.05 and P < 0.01, respectively, compared with WT based on Student's t test.

Discussion

To explore the use of PLDα1-mediated signalling for crop improvement, we placed PLDα1 under the control of a guard cell–specific promoter to avoid stress-induced PLDα1 degradation of membrane lipids in other tissues (Hong et al., 2008a), such as roots and mesophyll cells. Testing the construct in spring and semi-winter cultivars of canola shows that AtKatIpro::PLDα1 plants display reduced water loss and improved plant growth under drought. AtKatIpro::PLDα1 plants have higher relative water content and less membrane leakage than WT plants during drought stress. In addition, the expression level of drought-inducible genes, BnABI1, ABI2, RAB18, RD29B, ERA1, including BnPLDα1, was lower in AtKatIpro::PLDα1 than WT plants under drought. These patterns of gene expression are consistent with the physiological observation that AtKatIpro::PLDα1 plants experienced less drought stress than WT. Moreover, we have tested these plants in native field conditions where drought is often a common stress for canola growth in the Yangtze basin of China. In the field, AtKatIpro::PLDα1 plants of both cultivars showed significantly higher seed yield per plant than WT. The increase in seed production in AtKatIpro::PLDα1 plants results from increases in flowering stems and silique number, whereas the seed number per silique or seed weight is not altered.

The increased flowering stems and silique number are likely associated with better growth due to the maintenance of water status in AtKatIpro::PLDα1 plants. When the soil water content was only 12.8% during the flowering period, the relative water content of leaves in AtKatIpro::PLDα1 plants was higher than that of WT. The flowering time of AtKatIpro::PLDα1 plants was delayed for a few days in both Westar and Zhongshuang 9. The improved water retention and thus better vegetative growth may explain, at least in part, the alterations in flowering time, flowering stems and silique numbers in AtKatIpro::PLDα1 plants. To determine whether the increased PLDα1 expression in guard cells may have direct regulatory roles in plant development would require further investigation.

A potential added value from the AtKatIpro::PLDα1 manipulation is the increase in oleic acid content because a higher level of oleic acid with a lower level of C18:3 is a desired trait of cooking oil as it decreases the rancidity and increases oxidative stability. The improved vegetative growth may also contribute to the increased seed oil content in AtKatIpro::PLDα1 plants. OE plants kept a longer vegetative growth than WT, and decreased water loss from leaves compared with WT. These changes would allow a longer time and more photosynthate for seed filling. An alternative, but not mutually exclusive, hypothesis could be that increased PLDα1 in seeds may promote phospholipid, particularly phosphatidylcholine (PC) turnover. As PC is the major substrate for TAG synthesis, the increased turnover may promote TAG production and change in fatty acid composition. However, the latter hypothesis remains to be tested because PLDα1 expression is directed towards guard cells and we did not observe increased oil content in well-watered plants.

In addition to drought, AtKatIpro::PLDα1 plants displayed increased tolerance to NaCl in vegetative growth. The transgenic plants grew better and accumulated more biomass than WT under salt stress, and the difference grew greater with increased stress. PLDα1 and PA have been implicated in mediating salt stress signalling through the regulation of mitogen-activated protein kinase cascades in Arabidopsis (Yu et al., 2010). It will be of interest to examine whether the increased PLDα1 expression in guard cells alters salt signalling pathways.

These data indicate that the guard cell manipulation of PLDα1 has the potential to improve crop plant growth and production under drought. However, under well-watered conditions, there is no overt difference between AtKatIpro::PLDα1 and WT plants in water retention and growth. In addition, traits associated with drought tolerance may have a dual effect, depending on the degree of severity of drought, such as positive effect under severe drought and negative effect under mild drought stress (Tardieu, 2012). Therefore, further studies are needed to investigate the transgenic plant performance under different levels of water deficits and at multiple field locations to determine the broad applicability of the AtKatIpro::PLDα1 strategy to improving crop water-use efficiency and drought tolerance.

Experimental procedures

Transgene constructs and canola transformation

The guard cell–specific expression vector was constructed by cloning the AtKatI promoter (At5g46240, 3770bp) using a forward primer (5′-GGATCCTCCTTACGATTTTGACCCTACTC-3′) and a reverse primer (5′-TCTTTAGAGATCATCAAAAAGATGTC-3′), and the full length AtPLDα1 (At3g15730, 2640 bp) using a forward primer (5′-ATGGCGCAGCATCTGTTGC-3′) and a reverse primer (5′-CTGCCTCCAATCCTTACAACCTAA-3′) to pCAMBIA1300 binary vector (Figure 1a). The construct was introduced into Agrobacterium tumefaciens GV3101, which was used to infect canola hypocotyls. Seeds of canola cultivars Westar or Zhongshuang 9 were immersed with 75% ethanol for 1 min, ethanol was then removed, and seeds were incubated with 1.5% (w/v) HgCl2 containing 0.1% (v/v) Tween-20 for 15 min. After three washes with sterile water, seeds were germinated in sterile petri dishes containing 1/2 MS solid medium (pH 6.0, 1% agar). The seeds were germinated in dark for 7 days at 23 °C, and then hypocotyls were cut to a length of 6–8 mm for infection with Agrobacterium GV3101. Agrobacteria were grown to 0.6 OD, diluted with MS medium (pH 5.8) containing 3% sucrose, and 100 μm acetosyringone (AS), and incubated with hypocotyls for 30 min. The infected hypocotyls were dried with sterile filter paper and placed on a solid MS medium (pH 6.0) containing 3% sucrose, 1.8% mannitol, 1 mg/L 2,4-dichlorophenoxyacetic acid (2,4-D), 0.3 mg/L kinetin, 100 μm AS and 3 g/L Phytagel, for 2 days in the dark at room temperature. The hypocotyls were used to produce callus in the same medium containing 30 μm silver thiosulfate, 300 mg/L timentin and 25 mg/L kanamycin in 8 h night/16 h day at room temperature for 2–3 weeks. The shoots were generated from kanamycin resistant callus grown in MS (1×, pH6.0) containing 1% glucose, 0.25 g/L xylose, 0.6 g/L MES, 2 mg/L trans-zeatin, 0.1 mg/L IAA, 300 mg/L timentin and 25 mg/L kanamycin for 4–6 weeks. The kanamycin-resistant shoots were transferred to MS solid medium (1×, pH 6.0) with 1% sucrose, and 300 mg/L timentin for rooting. The transgenic plants were identified by PCR using AtPLDα1-specific primers 5′-CCATAAAGGTCCCAACAGA-3′ (forward) and 5′-GTCATCCCAAACAAGCAAA-3′ (reverse). The transgenic plants were grown in soil for seed collection.

Plant growth, agronomic traits and soil water content

The transgenic and WT seeds were germinated on watered gauze fixed on a tray (13.5 × 30.5 × 40.5 cm). Thirteen-day-old seedlings were transferred to soil or 1× MS liquid (1/2 MS solid) media with/without stressors in growth room at 26/22 °C under 14 h light/10 h dark or native field conditions. For the field study, plots were arranged by a randomized field plot design with three replications for WT and each PLDαa1-OE lines. Each plot area was 1.4 × 1.5 m, and 15 plants were planted in each area with the same spacing from one plant to the other. Flowering time and plant height were observed during growth and inflorescence stages (grown in field for 3–5 months). Other agronomic traits, such as silique number, seed yield and seed weight, were tallied from harvested plants grown in field. Six to ten plants from each plot were harvested, and data were analysed using the analysis of variance (ANOVA) and Duncan's multiple range test.

Soil water content was measured by random sampling from three locations where plants were growing. Soil was sampled from 5 to 18 cm in depth for the clay soil using a puncher. The similar amount of soil was collected from three different locations, immediately transferred to a sealed plastic bag, and weighed as Wf and the dry soil as Wd. The soil water content was calculated as (Wf − Wd)/Wd × 100%. When it was saturated with water, the clay soil water content is 58%, regarded as its water-holding capacity.

RNA extract and real-time PCR

Full expanded leaves were sampled, and total RNA was extracted using a TransZol reagent according to the manufacturer's instruction (Transgen, http://www.transgen.com.cn). Total RNA was digested with DNaseI to remove DNA and was used for the first-strand cDNA synthesis using TIANscript RT Kit (Tiangen, http://www.tiangen.com). The resulting cDNA was diluted five times and used as template (2 μL) for PCR. The primers for Brassica napus Actin (BnActin) are 5′-AGCTGGAGACGGCTAAGAG-3′ (forward) and 5′-GTTGGAAAGTGCTGAGGGA-3′ (reverse), and those for AtPLDα1 are 5′-TCTCTGCTTTGCTGCTGTTGTAGC-3′ (forward) and 5′-CACAAAGCTACATTCTCTCACCACGTC-3′ (reverse). Real-time PCR was tested using the MyiQ single-colour real-time PCR detection system (Bio-Rad; http://www.bio-rad.com) using SYBR green to monitor dsDNA synthesis. The primers for real-time PCR were listed (Table S1). The PCR conditions were as follow: 50 °C for 10 min for cDNA synthesis, 95 °C for 5 min for iScript reverse transcriptase inactivation and DNA denaturation, 45 cycles at 95 °C for 10 s, 60 °C for 30 s to PCR cycling and detection, and then 95 °C for 1 min, 55 °C for 1 min, 80 cycles at 55 °C for 10 s and each cycle increasing 0.5 °C to analysis melt curve. The gene expression level was normalized to that of BnActin.

Protein extraction and immunoblotting

The protein was extracted from abaxial epidermal peels collected from fully expanded leaves and ground by motor pestle at 4 °C in precooled buffer A (50 mm HCl–Tris pH 7.5, 10 mm KCl, 1 mm EDTA, 0.5 m PMSF and 2 mm DTT), containing 1 tablet/50 mL protease inhibitors mixture (Roche, Roche Applied Science, Mannheim, Germany; 11 836 170 001). The homogenates were centrifuged at 10 000 g for 10 min at under 4 °C, and the supernatant was used for immunoblotting and activity assays. Protein concentration was measured spectrophotometrically using the Bradford protein assay reagent (Bio-Rad; 500-0006) according to the manufacturer's instruction. Total protein was separated with 8% SDS-PAGE gel (3.5% stacking) and transferred to a PVDF membrane. The membrane was preblotted with 5% nonfat milk and then incubated with AtPLDα1 antibodies raised against the terminal regions of Arabidopsis PLDα1 (1 : 1000 v/v) at 4 °C overnight. After three washes with phosphate-buffered saline (PBS), the membrane was incubated with the second antibody IgG-alkaline phosphatase (Sigma, St. Louis, MO; A7539, 1 : 3000 v/v) at room temperature for 2 h. After washing with PBS three times, the membrane was incubated with a alkaline phosphatase substrate containing nitroblue tetrazolium 5-bromo-4-chloro-3-indolyl phosphate (1 : 1 v/v) at room temperature for colour development.

To monitor PLDα1 in guard cells, abaxial epidermal peels were collected and incubated for 1 h in a solution containing 5 mm MES-KOH pH 6.1, 22 mm KCl and 1 mm CaCl2. The peels were fixed in a buffer containing 1.5% formaldehyde, 0.5% glutaraldehyde, 0.1 m PIPES pH 6.9, 5 mm EGTA, 2 mm MgCl2 and 0.05% Triton X-100 for 35 min with gentle shaking. The fixed peels were washed three times with fresh PBS for 30 min (Sang et al., 2001), and then were transferred onto microscope slides and freeze-shattered followed with a procedure (Wasteneys et al.,1997). Briefly, the peels adhered on slides were incubated with enzymes containing 1% cellulase, 1% pectinase and 2% driselase in PBS for 30 min at 37 °C, and then were incubated with 5 mg/mL proteinase for 10 min at 37 °C. Triton X-100 (1%) was added to incubated peels for 1.5 h, the peels were incubated with a PBS buffer containing 50 mm glycine for 30 min, and then they were blocked with 3% BSA for 30 min. The peels were blotted with PLDα1-specific antibody (1 : 100 dilution) at 4 °C overnight, followed by the second antibody conjugated with alkaline phosphatase (1 : 50 dilution) for 2 h. After rinsing with PBS, the peels were incubated for 10 min in the substrate Fast Red TR/Naphthol AS-MX (Sigma; F4523). The slides were rinsed three times with PBS and observed under microscope.

Water loss, stomatal aperture, relative water content and ionic leakage

The same size, position and age of fully expanded leaves were excised from the transgenic and WT plants when plants were at the 5–6 leaf stage. Removed leaves were weighed promptly (Wf), left in the same growing conditions and then weighed every 10 min (Wn). The leaves were dried in an oven and weighed (Wd). The relative water loss of leaves was calculated as (Wf − Wn)/(Wf − Wd) × 100%.

Leaf discs (0.5 cm2) were isolated and put into a induced buffer (10 mm KCl, 0.2 mm CaCl2 and pH 6.15 10 mm MES) from 23 days old plants. The leaves were exposed at nature light (800–820 μmol/m2s) for 1 h, and 50 μm ABA was added for 30 min. Stomatal aperture was observed under microscope and calculate using a software ImageJ (National Institutes of Health, Bethesda, MD). Water transpiration and stomatal conductance were measured using a portable photosynthesis system (Yaxin-1102).

Seeds were germinated on a wetting filter for 7 days, and then seedlings were transplanted into pots and grown for 50–60 days in greenhouse. Drought treatments were initiated by withholding irrigation until leaves in WT plants began wilting. To measure ionic leakage of the stressed plants, leave discs (1 cm2) from the same size and position of same age leaves were immersed in 30 mL distilled water and were incubated with gentle agitation for 20 min. The initial ionic conductivity of bathing solution with leaf discs was measured. Then, the bathing solution with leaf discs was boiled for 30 min and cooled to room temperature, and the total ionic conductivity was measured. The relative ionic conductivity was represented as the percentage of the initial ionic conductivity versus the total ionic conductivity. To measure relative water content, another set of leaves with similar size, position and age were collected and weighed (Wf). The leaves were then submerged in water for 5 h and weighed (Ws) as leaves were saturated with water. Finally, the leaves were dried in an oven and weighed (Wd). The relative water content was calculated as (Wf − Wd)/(Ws − Wd) × 100%.

Seed oil content

The seed oil content was quantified by measuring fatty acid content using GC-FID. Briefly, seed samples were immersed in 1.5 mL methanol containing 5% H2SO4, 40 μL internal standard glyceryl triheptadecanoate (Sigma; T2151) and 300 μL methylbenzene in tubes. The tubes were incubated in a water bath at 85 °C for 3 h and then were cooled to room temperature. NaCl (0.9%, 1.5 mL) was added and mixed, which was followed by the addition of 2 mL hexane containing 0.2% BHT. After setting at room temperature for 15 min, upper phase was transferred to a new glass tube. The extraction was repeated twice, combined and dried by a stream of nitrogen. Finally, 0.4 mL hexane with 0.2% BHT was added to dissolve the FAME again, and 1 μL of FAME solution was injected to GC through the injection port with detector temperature at 280 °C and oven temperature at 180 °C for 2 min, and then increased by 10 °C/min up to 220 °C for 5 min.

Acknowledgements

We thank Deng Xianjun and Shen Qingwen for their assistance in this study. This work was supported by grants from the National Science Foundation of China (30871303, 30971852), the US Department of Agriculture (2007-35318-18393) and the Chinese National Key Basic Research Project (2012CB114200).

Ancillary