Disruption of AtWNK8 Enhances Tolerance of Arabidopsis to Salt and Osmotic Stresses via Modulating Proline Content and Activities of Catalase and Peroxidase

With no lysine kinases (WNKs) play important roles in plant growth and development. However, its role in salt and osmotic stress tolerance is unclear. Here, we report that AtWNK8 is mainly expressed in primary root, hypocotyl, stamen and pistil and is induced by NaCl and sorbitol treatment. Compared to the wild-type, the T-DNA knock-out wnk8 mutant was more tolerant to severe salinity and osmotic stresses, as indicated by 27% and 198% more fresh weight in the NaCl and sorbitol treatment, respectively. The wnk8 mutant also accumulated 1.43-fold more proline than the wild-type in the sorbitol treatment. Under NaCl and sorbitol stresses, catalase (CAT) activity in wnk8 mutant was 1.92- and 3.7-times of that in Col-0, respectively. Similarly, under salt and osmotic stress conditions, peroxidase (POD) activities in wnk8 mutant were 1.81- and 1.58-times of that in Col-0, respectively. Taken together, we revealed that maintaining higher CAT and POD activities might be one of the reasons that the disruption of AtWNK8 enhances the tolerance to salt stress, and accumulating more proline and higher activities of CAT and POD might result in the higher tolerance of WNK8 to osmotic stress.


Introduction
Salinity generates both ionic and osmotic stresses in plants [1]. Ionic stress is caused by toxic levels of sodium (Na + ) in the cytoplasm and by deficiencies of other ions, such as K + [2]. Under both salt and osmotic stress conditions, increased generation of H 2 O 2 (hydrogen peroxide), O 2 − (superoxide), O 2 (singlet oxygen) and OH (hydroxyl) radicals enhance leakage of electrons to molecular oxygen [1,3]. These cytotoxic reactive oxygen species (ROS) can destroy normal metabolism through oxidative damage to lipids, proteins and nucleic acids, leading to cell membrane damage and malondialdehyde (MDA) production [4]. Accordingly, plants have evolved an oxygen-scavenging system, which consists of superoxide dismutase (SOD, EC 1.15.1.1), catalase (CAT, EC 1.11.1.6), peroxidase (POD, EC 1.11.1.7), and so on [5]. Under drought-induced osmotic conditions, plant roots sense water deficit signals and produce more abscisic acid (ABA), which is transported to leaves through vascular tissue, leading to stomata closure to reduce water loss [3]. One of the best characterized biochemical responses of plant cells to osmotic stress is the accumulation of organic osmolytes, such as proline and betaines [6]. These substances, in turn, stabilize membranes and maintain protein conformation at low leaf water potential [7].
Protein kinases have been reported to regulate responses of plants to salt and osmotic stresses. For instance, Ca 2+ signals, triggered by salt stress, are perceived by SOS3-the binding of calcium and myristoylation of SOS3 are necessary for SOS3 function-and then, SOS3 activates the SOS2 kinase; the subsequently activated SOS2 kinase phosphorylates the SOS1 Na + /H + antiporter, which then pumps Na + out of the cytosol [8].
With no lysine kinase (WNK) is a subfamily of serine/threonine protein kinases existing in both animals and plants. A lysine residue in subdomain II of kinase is very important for Adenosine triphosphate (ATP) binding and highly conserved among all the other kinase subfamilies, which is missed in this WNK subfamily, but replaced by another lysine in subdomain I [9,10]. Human WNKs are genetically linked to the regulation of blood pressure [11]. However, WNKs are not found in yeast, indicating WNKs are restricted to multicellular organisms [12]. However, the roles of WNKs in plants are poorly understood.
In Arabidopsis, 10 WNK genes (AtWNK1-10) are supposed to encode WNKs, whereas the rice genome contains seven WNKs (OsWNK1-7) [13]. Compared with mammalian WNKs, plant WNKs are much smaller proteins with a predicted molecular weight of about 60-70 KDa. With the exception of AtWNK1 and OsWNK5, which consist only of a highly conserved N-terminal kinase domain, all the other plant WNKs have a highly divergent C-terminal domain of about 300 amino acids [13]. AtWNK1 phosphorylates the APRR3 member of the APRR1/TOC quintet, which is involved in the circadian rhythm [14]. On the other hand, the transcription of AtWNK2, AtWNK4 and AtWNK6, together with AtWNK1, are controlled by circadian rhythm [14]. AtWNK8 (At5g41990), a member of the WNK family, interacts with subunit C of the vacuolar H + -ATPase in vitro via a special short C-terminal domain and phosphorylates Arabidopsis vacuolar H + -ATPase subunit C at multiple sites [15], implicating that AtWNK8 is a participant in stress responses. AtWNK8 is localized in the nucleus and interacts physically with Arabidopsis enhanced downy mildew 2 (EDM2), which appears to act downstream of AtWNK8 in controlling floral transition by modulating expression of the floral repressor gene, flowering locus C (FLC, At5g10140) [16]. Recently, a soybean root-specific WNK homolog, GmWNK1, had been identified to regulate root development, possibly through mediating ABA homeostasis in vivo [17]. Hence, we hypothesized that AtWNK8 might mediate Arabidopsis responses to osmotic and salt stresses.
To date, the roles of AtWNK8 in abiotic stress responses remain unclear. In this study, we employed a T-DNA insertion mutant of WNK8 and an overexpression line of WNK8 (WNK8-OE) to study the roles of WNK8 in salt and osmotic stresses at physiological and molecular levels. We revealed that disruption of AtWNK8 enhances tolerance of Arabidopsis to salt and osmotic stresses mightbe via modulating proline content and activity of CAT and POD.

Activities of the AtWNK8 Promoter at Tissue Level
To investigate temporal and spatial activities of the AtWNK8 promoter, we analyzed beta-glucuronidase (GUS) staining at the tissue level using transgenic Arabidopsis harboring the promoter of AtWNK8 and a fused GUS reporter gene. The histochemical GUS staining results revealed that the promoter of AtWNK8 was active in stem and cotyledon of two-day-old seedlings, especially in the junction of hypocotyl and primary root (PR) (Figure 1a). In seven-day-old seedlings, a GUS signal was detected in roots, hypocotyls and leaves (Figure 1b). Results in Figure 1c, d indicated that the promoter of AtWNK8 was mainly active in hypocotyls near shoot apex and veins of cotyledon. GUS signal was also high in stamens and pistils of four-week-old plants ( Figure 1e). Figure 1f, g showed the activity of WNK8 promoter in roots, except root cap. Furthermore, root cross-section showed that the activity of WNK8 promoter was mainly confined in vascular tissues ( Figure 1h). In short, the WNK8 promoter is universally active in the hypocotyls, roots, stems and flowers, indicating that WNK8 appears to play important roles in plant growth and development.

Induction of AtWNK8 by NaCl and Sorbitol Stresses in Arabidopsis
In mammals, WNKs are ion or osmotic stress-responsive genes [22]. To clarify whether AtWNK8 transcription is induced by salt or osmotic stresses, Arabidopsis seedlings were subjected to a high concentration of NaCl or sorbitol-generated stress. Quantitative real-time polymerase chain reaction (qRT-PCR) was further employed to analyze the expression of AtWNK8. Figure 1. Expression patterns of the AtWNK8 gene. Arabidopsis Col-0 was transformed with a binary vector containing a proAtWNK8::GUS construct. Homogenous T3 transgenic plants were stained with X-gluc for 12 h, following the standard methods. The activities of AtWNK8 promoter were detected in two-day-old (a) or seven-day-old (b-d, f-h) or four-week-old (e) Arabidopsis, as illustrated: (a) two-day old seedling; (b) seven-day-old seedling; (c) hypocotyl; (d) leaves; (e) flowers; (f) root elongation zone; (g) root tip; and (h) cross section of root tip. Table 1. Bioinformatics analysis of cis-elements in the WNK8 (At5g14990) promoter region. Considering RD29A (At5g52310) is a stress responsive marker gene, it is induced by drought, salinity, cold and ABA [23]. We firstly detected the expression of AtRD29A with RNA samples from whole seedlings. Results in Figure 2a indicated that RDA29A was induced very quickly by NaCl and sorbitol treatment and then decreased. Whereas the WNK8 expression was induced rapidly and peaked at 1 h after high NaCl and sorbitol treatments, the duration of higher WNK8 expression level was short. The AtWNK8 expression gradually decreased to a normal level at 12 h after treatment ( Figure 2b). Moreover, salt stress more strongly induced WNK8 expression than sorbitol at 1 h after treatment. Generally, the responsive patterns of WNK8 were similar to RD29A (Figure 2a,b). To further determine the responses of AtWNK8 to the above-mentioned stresses at tissue level, the transgenic seedlings were treated with two concentrations of NaCl and sorbitol for 12 h. As shown in Figure 2c, under control conditions, the GUS staining can be detected in roots, leaves, stems and shoot apexes. The activities of the WNK8 promoter were induced in flowers and leaves, as well as the upper part of primary roots (PRs) under salt stress. In addition, 50 mM sorbitol treatment only stimulated the activities of WNK8 promoter in the root tips, but 400 mM sorbitol upregulated the activities of WNK8 promoter in the aerial part, root tips and elongation zone of PRs. Whereas a previous study showed that the transcripts of WNK8 were higher in roots at the seedling stage (two-week-old) and in stems at the flowering stage (six-week-old), they were almost undetectable in other organs based on RT-PCR results [13]. Altogether, data in Figure 2 indicated that AtWNK8 responded to the changes of environmental NaCl and sorbitol concentration.

Figure 2.
Responses of the AtWNK8 expression to salt and osmotic stresses. The expression levels of AtRD29A (a) and AtWNK8 (b) were determined by quantitative real-time polymerase chain reaction (qRT-PCR), and the beta-glucuronidase (GUS) signal in transgenic Arabidopsis harboring proWNK8::GUS was observed through standard methods (c). Total RNA was extracted from whole seedlings. Seven-day-old seedlings were treated with NaCl or sorbitol, as the indicated time (a and b), or treated with NaCl or sorbitol for 12 h (c); Scale bar, 100 μm. Data were the means ± SE of three independent experiments. Dehydration, cold, salt and osmotic stresses, as well as ABA induced the expression of RD29A [23]. The 120 bp promoter region of RD29A contains the DRE, DRE/CRT-core motif (A/GCCGAC) and ABRE element [23]. Consistently, the WNK8 promoter contains a DRE-and cold-responsive cis-element (Table 1). It is well known that salt and osmotic stresses result in dehydration, so we reasoned that salt and osmotic stress-induced WNK8 expression seems to be related to DRE elements. Furthermore, identification of WNK8 promoter-binding TFs would facilitate deciphering functions of WNK8 in salt and osmotic stress responses.

Responses of Knock-Out Mutant wnk8 and Overexpression Line of WNK8 (WNK8-OE) to Salt and Osmotic Stresses
To obtain further evidence of the in vivo functions of WNK8 responding to salt and osmotic stresses, initially we employed two T-DNA insertion mutants (SALK_024887 and SALK_058925; see [13]) and two independent WNK8 overexpression lines (WNK8-OE) to explore the performance of those lines under 150 mM NaCl and 300 mM sorbitol. Moreover, we found that the two independent T-DNA insertion lines or two overexpression lines showed similar responses under the two adverse conditions (data not shown), so we just employed one T-DNA insertion line (SALK_024887) as wnk8 mutant and one overexpression line as WNK8-OE in the follow-up experiments, such as measuring the contents of chlorophyll (Chl), relative water content (RWC), Fresh weight (FW), proline content and activities of POD and CAT.
The loss of the function mutants in WNK8 (SALK_024887 and SALK_058925) were characterized by previous studies [13], and the overexpression lines (WNK8-OE) were generated (see Experimental Section and Figure S1). PCR analysis verified that the homozygous AtWNK8 T-DNA insertion mutant wnk8 was a 100% knock-out mutant that did not transcribe the WNK8 gene (data not shown, see [13]). qRT-PCR analysis showed that the expression level of WNK8 was over three-fold higher in the homozygous T3 WNK8-OE line than that in the wild-type Col-0 ( Figure S1).
Grown in 300 mM sorbitol (Figure 3a) for 30 days or 150 mM NaCl (Figure 3b) for 10 days, the wild-type plants progressively turned yellow, exhibited chlorosis and ceased growth, but the wnk8 mutant plants remained relatively green and grew slowly on petri dishes, suggesting that WNK8 is a negative player for salt and osmotic stress adaptations.
Next, to investigate whether the altered stress responses occurred in natural conditions, the wild-type Col-0, wnk8 and WNK8-OE seedlings were planted in soil. Figure 3c shows that under the two stresses, progressive chlorosis, reduced leaf size and general growth inhibition of wild-type and WNK8-OE were observed after all treatments through top irrigation and eventually died. In contrast, the wnk8 mutant survived longer. Compared to sorbitol treatment, NaCl treatment had a relatively smaller adverse effect on wnk8 growth. In addition, the wnk8 mutant also showed more tolerance to mannitol treatment (data not shown). On the other hand, the difference of the tolerance to NaCl was not as obvious as the tolerance to sorbitol among Col-0, wnk8 and WNK8-OE. This indicated that WNK8 plays more negative roles in the responses to osmotic stress than to salt stress.
Numerous studies demonstrated that root system architecture is dramatically influenced by environmental factors, including salt and osmotic stresses [24,25]. To investigate the involvement of AtWNK8 in root growth and development as affected by the two adverse conditions, uniform seedlings were grown in 1/2 MS media containing 150 mM NaCl or 300 mM sorbitol for seven days in vertically placed petri dishes. The length of PR and the number of lateral roots (LRs) were measured. As shown in Figure 3d, both NaCl and sorbitol applications significantly decreased PR length and the number of LRs in Col-0, wnk8 and WNK8-OE, whereas the PR length of wnk8 was longer than Col-0 and WNK8-OE under salt stress, and the PR length of WNK8-OE was shorter than Col-0 and wnk8 under osmotic stress. This indicated that WNK8 negatively regulated PR growth under the two stressed conditions. For the number of LRs, no significant differences among Col-0, wnk8 and WNK8-OE were found under normal and adverse conditions (Figure 3d). These data implied that WNK8 appeared to play a small role in LR development. Figure 3. The wnk8 mutant was more tolerant to salt and osmotic stresses. Seeds from wild-type and WNK8 plants were germinated on standard 1/2 MS medium for three days; then, uniform seedlings were transferred to vertically placed plates containing 300 mM sorbitol for 30 days (a) or 150 mM NaCl for 10 days (b). Three-day-old seedlings of Col-0, wnk8 and WNK8-OE grown in standard 1/2 MS medium were transferred to soils and grown for 20 days and then treated with 150 mM NaCl or 300 mM sorbitol via top irrigation once every three days. Photos were taken after a 15-day treatment (c). Stratified seeds were directly germinated in 1/2 MS media containing 150 mM NaCl or 300 mM sorbitol for seven days in vertically placed petri dishes, and then, the length of PR and the number of lateral roots were counted (d). Data are the means ± SE from three independent experiments, and each experiment included 20 seedlings tested. Asterisks on bars indicate significant differences in contrast to Col-0 under the same treatment (Student's t-test, p < 0.05). Number of lateral root (no/seedling)

Effects of Salt and Osmotic Stresses on Fresh Weight (FW), Relative Water Content (RWC) and Chlorophyll Content of Col-0, wnk8 Mutant and WNK8-OE Plants
As shown in Figure 4a, under normal growth conditions, there were no significant differences among Col-0, wnk8 and WNK8-OE, as indicated by FW. By contrast, salt and osmotic stresses significantly inhibited growth of Col-0, wnk8 and WNK8-OE, resulting in less FW. Under salt stress conditions, the FW of WNK8 was the highest, followed by Col-0 and WNK8-OE, and the FW of WNK8 was 1.27-times that of Col-0 (p < 0.05). Under osmotic stress conditions, the FW of wnk8 was 2.98-times that of Col-0 (p < 0.05). (a) Three-day-old seedlings germinated in standard 1/2MS media were transferred to 150 mM NaCl or 300 mM sorbitol for treatment. FW was measured after seven days of treatment. Data are the means ± SE from three independent experiments, and each experiment included 20 seedlings tested. Asterisks on bars indicate significant differences in contrast to the wild-type under the same treatment (Student's t-test, p < 0.05). (b,c) Three-day-old seedlings grown in 1/2 MS media were transferred to soil and grown for 20 days and then treated with 150 mM NaCl or 300 mM sorbitol via irrigation once every three days. After a 15-day treatment, RWC (b) and chlorophyll content (c) were measured. Data are the means ± SE from three independent experiments, and each experiment included 20 seedlings tested. Asterisks on bars indicate significant differences in contrast to the wild-type under the same treatment (Student's t-test, p < 0.05).  Figure 4b showed that as to RWC, no significant differences among Col-0, wnk8 and WNK8-OE were found under control conditions. However, under salt stress, RWC of WNK8-OE was significantly lower than wnk8 and Col-0. More importantly, subjected to osmotic stress, the RWC of wnk8 was higher than that of Col-0; consistently, the RWC of WNK8-OE was lower than that of Col-0 (p < 0.05).

Results in
Compared with the control, NaCl and sorbitol treatments significantly decreased the chlorophyll content of Col-0 and WNK8-OE, but not the wnk8 mutant. Under salt and osmotic stress conditions, the chlorophyll content of WNK8 was significantly higher than Col-0 and WNK8-OE (p < 0.05, Figure 4c).
These results indicated that WNK8 was a negative regulator for Arabidopsis in response to salt and osmotic stresses. The relatively high level of RWC and chlorophyll content of WNK8 under the two adverse conditions will maintain cell activity and stabilize photosynthesis, thus leading to more FW. Taken together, the WNK8 mutant appeared to be less sensitive, but more tolerant, to salt and osmotic stresses in terms of physiological and growth parameters.

Higher Tolerance of wnk8 to Salt and Osmotic Stresses Was Partially Achieved by Accumulating More Proline
Apart from acting as a compatible solute to protect cells from dehydration damage, proline can act as a free radical scavenger, thus increasing plant tolerance to abiotic stresses [26]. Figure 5 showed that Col-0, wnk8 and WNK8-OE plants all accumulated proline under salt and osmotic stresses, implying that accumulation of proline is a universal mechanism of Arabidopsis to cope with the two abiotic stresses.

Figure 5.
Proline content in Col-0, wnk8 and WNK8-OE under salt and osmotic stress conditions. Three-day-old seedlings grown in standard 1/2 MS media were transferred to soil and further cultivated for 20 days and then treated with 150 mM NaCl or 300 mM sorbitol via irrigation once every three days. After 15 days, the proline content was measured. Data are the means ± SE from three independent experiments, and each experiment included 20 uniform seedlings tested. Asterisks on bars indicate significant differences in contrast to the wild-type under the same treatment (Student's t-test, p < 0.05).

Control
NaCl Sorbitol Under control conditions, the content of proline in wnk8 and WNK8-OE was higher than that in Col-0. Although salt stress stimulated proline accumulation in Col-0, the proline content was similar among Col-0, wnk8 and WNK8-OE under salt stress. This suggested that alteration of WNK8 transcription had no effect on proline accumulation under salt stress, and the increased tolerance of the wnk8 mutant to salinity was independent of proline accumulation. More interestingly, under osmotic stress, the content of proline in wnk8 was the highest and that in WNK8-OE was the lowest, indicating that the greater tolerance of wnk8 to osmotic stress seems to be dependent on proline accumulation. It was documented that P5C synthase (P5CS) and P5C reductase catalyzed proline biosynthesis [27], and proline dehydrogenase (ProDH) and P5C dehydrogenase mediated proline degradation in plants [28]. Further studies are needed to clarify whether loss of function of WNK8 altering the proline level in Arabidopsis is dependent on P5CS and P5C reductase under salt and osmotic stresses.

The wnk8 Mutant Maintained Higher Activities of CAT and POD in Salt and Osmotic Stresses
The ability to scavenge ROS in a short time is crucial for plants to tolerate stress and survive. Antioxidative CAT and POD are important enzymes for quenching ROS in plants [29]. As shown in Figure 6a, under control conditions, no difference of CAT activity among Col-0, wnk8 and WNK8-OE can be found. However, under salt stress, the WNK8 mutant had 1.92-fold higher CAT activity than Col-0, but no difference between Col-0 and WNK8-OE. When subjected to sorbitol stress, the activity of CAT in wnk8 mutant was 3.66-fold of that in the wild-type (p < 0.05), but the difference between Col-0 and WNK8-OE was insignificant (p > 0.05, Figure 6a). We noticed that both salt and sorbitol treatments significantly increased CAT activity in WNK8. Figure 6b showed that under control conditions, the POD activity in Col-0 was significantly higher than that in WNK8 and WNK8-OE, but opposite in salt stress; the POD activity in wnk8 was 1.81-times higher than Col-0 and 5.33-times higher than WNK8-OE. On the other hand, under sorbitol stress, relative to Col-0, the POD activity in wnk8 was increased by 58%, and that in WNK8-OE was significantly lower (p < 0.05). This indicated that WNK8 is a negative regulator for POD activity under salt and osmotic stress conditions.
CAT and POD are involved in quenching ROS in plant tissues. Results in Figure 6 implied that maintaining higher CAT and POD activity might be one of the reasons why wnk8 mutant was more tolerant to salt and osmotic stresses. Since AtWNK8 is a kinase, we cannot rule out that CAT and POD may be directly phosphorylated by it or indirectly mediated by WNK8 through other unknown proteins.
More interestingly, Urano et al. (2012) demonstrated that WNK8 phosphorylates AtRGS1 (regulator of G-protein signaling 1) and then stimulates the endocytosis of AtRGS1, and the induced AtRGS1 endocytosis by D-glucose, but not L-glucose, is dependent on WNK8 [30]. Accordingly, the wnk8 mutant has decreased sensitivity to 6% D-glucose, and overexpression of WNK8 results in more sensitivity to D-glucose [30]. At this point, WNK8 seems to be not involved in glucose-generated osmotic stress responses, but is involved in glucose signaling. On the other hand, overexpression of GmWNK1 in Arabidopsis increases the tolerance to salt and osmotic stresses [31], and GmWNK1 is more close to AtWNK1 [17]. We noticed that AtWNK8 is most divergent form AtWNK1 [30]. Given the opposite effects of AtWNK1 and AtWNK8 in osmotic and salt stresses, it is tempting to reveal underlying mechanisms. Figure 6. Activities of catalase (CAT) and peroxidase (POD) in Col-0, wnk8 and WNK8-OE under salt and osmotic stress conditions. Three-day-old seedlings grown in standard 1/2 MS media were transferred to soil and grown for 20 days and then treated with 150 mM NaCl or 300 mM sorbitol via irrigation once every three days. After 15 days, the activities of CAT (a) and POD (b) were determined. Data are the means ± SE from three independent experiments, and each experiment included 20 uniform seedlings tested. Asterisks on bars indicate significant differences in contrast to the wild-type under the same treatment (Student's t-test, p < 0.05).

Control
NaCl Sorbitol

Plant Materials and Growth Conditions
All the Arabidopsis thaliana plants used in this study were Columbia-0 (Col-0) ecotype. Genomic DNA was extracted via the cetyl trimethylammonium bromide (CTAB) method [32]. We amplified the promoter fragment of AtWNK8 1804 bp upstream of the ATG initiation code through PCR with the primer pair composed of the following forward primer (5'-AAAggatccCTCTGCTGCGTTCTTTGGG C-3') and the reverse primer (5'-AAAccatggCAAACAAAGCAATCGAGAAC-3'). The pCAMBIA 1305.2 plasmid [33] DNA digested by BamHI and NcoI was cut with the original 35S promoter and ligated with the AtWNK8 promoter digested by BamHI and NcoI enzyme and then transformed with the constructed pCAMBIA 1305.2 vector containing the proWNK8::GUS construct via the floral dip method [34]. Transformed lines were selected in hygromycin-containing medium. Homozygous T3 seedlings were used in the subsequent experiments.
Total RNA was extracted through a TRIZol method and subjected to reverse transcriptase polymerization to get cDNA. The forward primer (5'-AAAggatccgATGGCTTCTGGTTCTGGATT-3') and the reverse primer (5'-AATtctagaAGAGATGTTAACTGCTTTTTGCT-3') were used to clone the full length WNK8 cDNA open reading frame (ORF) into pCAMBIA1380, which harbors a 35S promoter. Col-0 was transformed via the floral dip method [34] to get homozygous WNK8-OE lines.

NaCl and Sorbitol Treatments
For sterilized media growth, seeds of wild-type, wnk8, and WNK8-OE were surface sterilized with 70% ethanol for 1 min and then with 10% bleach for 5 min before being washed five times with sterilized water. The seeds were germinated and grown on vertically placed petri dishes containing sterilized half strength MS media (pH 5.7), as described in Wang et al. (2008) [13]. Uniform 3-day-old plants were transplanted into 1/2 MS media supplemented with NaCl or sorbitol at indicated concentrations and grown for the indicated time. For taking pictures to indicate the differences between lines, seedlings were treated with NaCl for 10 days or with sorbitol for 30 days.
For measuring root parameters, stratified seeds were grown in 1/2 MS media containing 150 mM NaCl or 300 mM sorbitol for 7 days in vertically placed petri dish before being scanned. The lengths of PR were measured using ImageJ 1.43 software [36], and the numbers of LR were counted under microscopy. For soil culture, 3-day-old plants, transferred from 1/2 MS media, were grown in a growth chamber with a 16/8 h photoperiod at 22/18 °C, 75% RH and 100 μmol m −2 s −1 light for 20 days, and then they were irrigated with 150 mM NaCl or 300 mM sorbitol solution once every 3 days. Photos were taken after 15 days of treatment just before harvesting.

Determination of WNK8 Promoter Activity
For detecting AtWNK8 promoter activity using transgenic plants, 2-day-old or 7-day-old or four-week-old homozygous T3 transgenic plant tissues were directly submerged into GUS staining solution, as described by Jefferson et al. (1987) [33], and incubated at 37 °C . The chlorophyll of plant samples was cleared in ethanol, and photographs of GUS activity were taken using a digital camera under a stereomicroscopy (Leica DMB5000). Treated seedlings grown in half strength 1/2 MS media containing NaCl or sorbitol, as indicated, were taken at indicated time points before being photographed.

Determination of WNK8 Transcript Level under Salt and Osmotic Stress
Seven-day-old plants cultivated in 1/2 MS media were transferred to 1/2 MS media supplemented with NaCl or sorbitol at tested concentrations for the indicated time. Total RNAs were extracted, as described by Wang et al. (2008) [13]. qRT-PCR was performed using the ToYoBo qRT-PCR SYBR Green Mix kit (ToYoBo, Japan) with a Corbett Research Rotor-Gene 2000 cycler. The reference housekeeping gene, AtEF1a (At1g07920), was used as the control and used to normalize qRT-PCR results. Accordingly, the expression level of AtEF1a was set as 1.0, the relative expression levels of other genes were determined via the comparison with that of AtEF1a. A standard curve was obtained for each gene using different dilutions of cDNA template mix. Reaction conditions for thermal cycling were: 95 °C for 2 min, 40 cycles of 95 °C for 15 s, 56 °C for 15 s and 72 °C for 45 s. Fluorescence data were collected during the cycle at 72 °C. Reactions were technically repeated three times independently. Quantification of each gene was performed using the Corbett Research Rotor-Gene software. Related primers were list in Table S1.

Determination of FW, RWC and Chlorophyll Content
Three-day-old seedlings were germinated in the standard 1/2 MS media containing 150 mM NaCl or 300 mM sorbitol. After 7 days, FWs were measured. Three-day-old seedlings germinated in standard 1/2 MS agar plates were transferred to pots and cultivated for 20 days, then irrigated with 150 mM NaCl or 300 mM sorbitol every three days. After 15-days' treatment, the chlorophyll was then extracted with 100% aqueous acetone, and total chlorophyll content (including chlorophyll a and b) was measured, as described previously [37]. Six leaves from each plant treated as mentioned above were used for RWC determination; after FW determination, the tissues were placed in distilled water for 24 h at room temperature. The hydrated shoot tissues were weighed to determine the turgid weight (TW). The tissues were subsequently dried in an oven at 60 °C for 48 h and weighed to determine the dry weight (DW). RWC was calculated as RWC = (FW − DW)/(TW − DW) × 100, as described by Smart (1974) [38].

Measurement of Proline Content and Activities of CAT and POD
Three-day-old seedlings germinated in standard 1/2 MS agar plates were transferred to pots and further cultivated for 20 days, then irrigated with 150 mM NaCl or 300 mM sorbitol every three days. After 15-days' treatment, plants were harvested for proline, CAT and POD analysis.
Proline content was measured, as described by Bates (1973) [39]. Briefly, about 0.3 g shoots from both control and treatment groups were homogenized with liquid nitrogen. Tissue powders were suspended in 1 mL of 3% sulfosalicylic acid and centrifuged at 1000 × g for 5 min at 4 °C. 0.1 mL. The supernatant was mixed with 0.2 mL acid ninhydrin, 0.2 mL 96% acetic acid and 0.1 mL 3% sulfosalicylic acid. The mixtures were incubated at 96 °C for 1 h, mixed with 1 mL toluene and further centrifuged at 1000 × g for 5 min at 4 °C. Upper phases were collected, and the absorbencies were read at 520 nm. The proline concentration was determined by using an extinction coefficient of 0.9999 that was derived from a standard curve.
For the activities of CAT and POD, leaf samples from control and stress treatments were homogenized with liquid nitrogen and suspended in phosphate buffer (pH 7.8) for enzyme activity analysis. The suspensions were centrifuged at 12,000 × g for 20 min at 4 °C, and the supernatants were used for activity analysis. The protein content in shoot extracts was determined with Bradford's method [40], using bovine serum albumin as a standard. Catalase (CAT) activity was determined accordingly [41]. Samples containing 100 μg protein were suspended in 1 mL of 50 mM Tris-HCl solution at pH 7.8. The assay medium consisted of 50 mM potassium phosphate buffer at pH 7 and 10 mM H 2 O 2 . The decrease of the H 2 O 2 absorbance was followed for 90 s at 240 nm at room temperature. Nanomoles of hydrogen peroxide consumed per minute were defined as one unit of CAT. POD activity was determined according to Shannon et al. (1966) [42]. The reaction mixture consisted of 3 mL of 0.