AtSK11 and AtSK12 Mediate the Mild Osmotic Stress-Induced Root Growth Response in Arabidopsis

Although most osmotic stresses are harmful to plant growth and development, certain drought- or polyethylene glycol (PEG)-induced mild osmotic stresses promote plant root growth. The underlying regulatory mechanisms of this response remain elusive. Here, we report that the GLYCOGEN SYNTHASE KINASE 3 (GSK3) genes ARABIDOPSIS THALIANA SHAGGY-RELATED KINASE 11 (AtSK11) (AT5G26751) and AtSK12 (AT3G05840) are involved in the mild osmotic stress (−0.4 MPa) response in Arabidopsis thaliana. When grown on plant medium infused with different concentrations of PEG to mimic osmotic stress, both wild-type (WT) and atsk11atsk12 plants showed stimulated root growth under mild osmotic stress (−0.4 MPa) but repressed root growth under relatively strong osmotic stress (−0.5, −0.6, −0.7 MPa) as compared to the mock condition (−0.25 MPa). The root growth stimulation of atsk11atsk12 was more sensitive to −0.4 MPa treatment than was that of WT, indicating that AtSK11 and AtSK12 inhibit the mild stress-induced root growth response. RNA-seq analysis of WT and atsk11atsk12 plants under three water potentials (−0.25 MPa, −0.4 MPa, −0.6 MPa) revealed 10 differentially expressed candidate genes mainly involved in cell wall homeostasis, which were regulated by AtSK11 and AtSK12 to regulate root growth in response to the mild stress condition (−0.4 MPa). Promoter motif and transcription factor binding analyses suggested that the basic helix-loop-helix (bHLH) transcription factor bHLH69/LJRHL1-LIKE 2 (LRL2) may directly regulate the expression of most −0.4 MPa-responsive genes. These findings indicate that mild osmotic stress (−0.4 MPa) promotes plant growth and that the GSK3 family kinase genes AtSK11 and AtSK12 play a negative role in the induction of root growth in response to mild osmotic stress.


Introduction
Plants live in constantly changing environments, which sometimes can be unfavorable for plant growth and development. In addition, climate changes and extreme water conditions are becoming more intense [1]. The sessile nature of plants requires that they evolve complex mechanisms to cope with various stress conditions such as drought-induced osmotic stress. Under severe osmotic stress, plants usually accelerate their life cycle to ensure reproduction before death, whereas under mild osmotic stress, such as moderate drought, plants usually adjust their growth to balance water uptake and water loss, and, more importantly, enhance root growth to promote water uptake [2][3][4].
Drought-or polyethylene glycol (PEG)-induced mild osmotic stress facilitates plant root growth. It is reported that drought stress during later stages of vegetative growth or early reproduction development increases root growth in soybean [5]. Similarly, in cotton, mild drought stress during the initial stage enhances root elongation [6]. A method was developed for producing plant medium
To investigate whether AtSK11, AtSK12, and AtSK13 are involved in osmotic response, we tested the responses of single mutants atsk11, atsk12, and atsk13 and double mutants atsk11atsk12 and atsk11atsk13 to PEG and mannitol treatment. When we used PEG to mimic osmotic stress conditions, WT and atsk11atsk12 showed stimulated root growth under the mild stress condition (−0.4 MPa) and reduced root growth under relatively strong osmotic stress conditions (−0.5, −0.6, and −0.7 MPa) compared to the mock condition without PEG (−0.25 MPa) ( Figure 2A). The atsk11atsk12 mutant was more sensitive in promoting root growth under the −0.4 MPa treatment than the WT ( Figure 2B), which indicates a negative role for AtSK11 and AtSK12 in this mild stress response. In addition, the root length of the other mutants (atsk11, atsk12, atsk13, atsk11atsk13) and gAtSK11-Flag and gAtSK12-Flag transgenic lines did not significantly differ from the WT under mock and treatment conditions ( Figure  S2, Supplementary Materials). However, under mannitol treatment, the WT and these mutants showed reduced root growth compared to the mock condition, which is different from PEG-mimicked osmotic stresses ( Figure 2C). These results indicate that AtSK11 and AtSK12 play a negative role in the mild stress (−0.4 MPa)-induced root growth regulation.
To investigate whether the protein levels of AtSK11 and AtSK12 are regulated by osmotic stress, we first performed immunoblot analysis using protein samples from the gAtSK11-Flag and gAtSK12-Flag roots. The results indicated that AtSK11 and AtSK12 protein levels were similar under different conditions in the same time point ( Figure S3A-D, Supplementary Materials). Therefore, the tested osmotic stress conditions may not influence the protein levels of AtSK11 and AtSK12 in the indicated time point.  replicates. Student's t-test was used to determine the significance of difference between WT and mutant under each treatment condition. Significant levels: * p < 0.05, ** p < 0.01, *** p < 0.001.

Identification of the Mild Osmotic Stress-Responsive Genes Regulated by AtSK11 and AtSK12
To identify downstream genes that are regulated by the mild (−0.4 MPa) osmotic stress through AtSK11 and AtSK12, we performed RNA-seq of the atsk11atsk12 and WT plants exposed to the mock condition (−0.  Error bars indicate SD. Student's t-test was used to determine the significance of difference between WT and mutant under each treatment condition. Significant levels: *** p < 0.001. (C) Root length of WT and various AtSK11, AtSK12, and AtSK13 mutants under mannitol treatment. At least 50 seedlings were measured per genotype under each condition. Data are means of three independent biological replicates. Student's t-test was used to determine the significance of difference between WT and mutant under each treatment condition. Significant levels: * p < 0.05, ** p < 0.01, *** p < 0.001.

Identification of the Mild Osmotic Stress-Responsive Genes Regulated by AtSK11 and AtSK12
To identify downstream genes that are regulated by the mild (−0.4 MPa) osmotic stress through AtSK11 and AtSK12, we performed RNA-seq of the atsk11atsk12 and WT plants exposed    Figure 3C). Among these, 143 DEGs were identified between WT and atsk11atsk12 at −0.4 MPa, but not at −0.25 MPa or −0.6 MPa. Finally, we identified 12 genes overlapping between these 143 DEGs and the 36 DEGs (specifically induced by mild osmotic stress (−0.4 MPa) treatment in atsk11atsk12). These 12 genes were considered to be mild stress-responsive genes regulated by AtSK11 and AtSK12 ( Figure 3D).
The expression levels of these 12 mild stress-responsive genes were significantly downregulated in atsk11atsk12 vs. WT under mild stress treatment ( Figure 4A), indicating that they may work downstream of AtSK11 and AtSK12 to inhibit root growth under mild stress. Gene Ontology (GO) enrichment analysis suggested that these genes are involved in regulating root growth, including root hair differentiation, root epidermal cell differentiation, and plant cell wall organization ( Figure 4B). Besides two genes in the chloroplast genome, the 10 remaining genes include five extensin genes (EXT6, EXT10, EXT11, EXT12, and EXT13), one extensin-like family member gene (AT4G08410), one proline-rich protein gene (PRP3), a xyloglucan endotransglucosylase/hydrolase protein gene (XTH14), a gene expressed in root hair cells (AT3G49960), and TRANSPARENT TESTA 6 (TT6), encoding a flavanone 3-hydroxylase involved in flavonoid biosynthesis [28] ( Figure 4A). We measured the expression levels of EXT6, EXT10, EXT12, and EXT13 by qRT-PCR ( Figure S4, Supplementary Materials); the results were in accordance with the RNA-seq data ( Figure 4A). In summary, except for two chloroplast genome genes, we identified 10 mild stress-responsive genes regulated by AtSK11 and AtSK12 that might mediate changes in root growth under −0.4 MPa treatment. and atsk11atsk12 at -0.4 MPa, but not at -0.25 MPa or -0.6 MPa. Finally, we identified 12 genes overlapping between these 143 DEGs and the 36 DEGs (specifically induced by mild osmotic stress (-0.4 MPa) treatment in atsk11atsk12). These 12 genes were considered to be mild stress-responsive genes regulated by AtSK11 and AtSK12 ( Figure 3D). The expression levels of these 12 mild stress-responsive genes were significantly downregulated in atsk11atsk12 vs. WT under mild stress treatment ( Figure 4A), indicating that they may work downstream of AtSK11 and AtSK12 to inhibit root growth under mild stress. Gene Ontology (GO) enrichment analysis suggested that these genes are involved in regulating root growth, including root hair differentiation, root epidermal cell differentiation, and plant cell wall organization ( Figure  4B). Besides two genes in the chloroplast genome, the 10 remaining genes include five extensin genes (EXT6, EXT10, EXT11, EXT12, and EXT13), one extensin-like family member gene (AT4G08410), one proline-rich protein gene (PRP3), a xyloglucan endotransglucosylase/hydrolase protein gene (XTH14), a gene expressed in root hair cells (AT3G49960), and TRANSPARENT TESTA 6 (TT6), encoding a flavanone 3-hydroxylase involved in flavonoid biosynthesis [28] ( Figure 4A). We measured the expression levels of EXT6, EXT10, EXT12, and EXT13 by qRT-PCR ( Figure S4, Supplementary Materials); the results were in accordance with the RNA-seq data ( Figure 4A). In summary, except for two chloroplast genome genes, we identified 10 mild stress-responsive genes regulated by AtSK11 and AtSK12 that might mediate changes in root growth under -0.4 MPa treatment.

Extensin Genes and TT6 Inhibit Root Growth in Response to Mild Osmotic Stress Treatment
To investigate the roles of the mild stress-responsive genes in root growth, we created a set of knockdown lines of extensin genes (EXT6, EXT10, EXT12, and EXT13) and TT6, PRP3, XTH14, and AT3G49960 by microRNA interference. Each artificial microRNA vector was driven by its own promoter. The gene expression level of each microRNA target in selected knockdown lines was reduced at least 50% ( Figure S5A,B, Supplementary Materials). As these genes were downregulated by mild stress treatment in atsk11atsk12 ( Figure 4A), we predicted that the knockdown lines would have longer roots under the mild stress condition than the WT. We did observe that the transgenic lines EXTp::miR-EXT6 (targeting EXT6, EXT10, EXT12, and EXT13) and TT6p::miR-TT6 had longer roots under -0.25 and -0.4 MPa than the WT. By contrast, the root lengths in the knockdown lines of PRP3, XTH14, and AT3G49960 were not significantly different from that of the WT at -0.4 MPa

Extensin Genes and TT6 Inhibit Root Growth in Response to Mild Osmotic Stress Treatment
To investigate the roles of the mild stress-responsive genes in root growth, we created a set of knockdown lines of extensin genes (EXT6, EXT10, EXT12, and EXT13) and TT6, PRP3, XTH14, and AT3G49960 by microRNA interference. Each artificial microRNA vector was driven by its own promoter. The gene expression level of each microRNA target in selected knockdown lines was reduced at least 50% ( Figure S5A,B, Supplementary Materials). As these genes were downregulated by mild stress treatment in atsk11atsk12 ( Figure 4A), we predicted that the knockdown lines would have longer roots under the mild stress condition than the WT. We did observe that the transgenic lines EXTp::miR-EXT6 (targeting EXT6, EXT10, EXT12, and EXT13) and

LRL2 Functions Downstream of AtSK11 and AtSK12 to Regulate -0.4 MPa-Responsive Gene Expression
As kinases, AtSK11 and AtSK12 are unlikely to directly regulate gene expression; therefore, we reasoned that transcription factors must be involved to mediate their effects on the mild stressresponsive gene expression. To identify these transcription factors, we performed promoter motif analysis of the mild stress-responsive genes ( Figure 6) and identified two conserved motifs. Motif 1 was enriched in the promoters of all 10 genes except PRP3; and motif 2 was enriched in the promoters of EXT6, EXT10, EXT11, EXT12, EXT13, AT4G08410, and XTH14 (Tables S1 and S2, Supplementary Materials). We compared the two motifs to the motifs in the DAP-seq databases. The basic helix-loophelix (bHLH) transcription factor bHLH69 was predicted to bind to motif 1 (with p-value of 2.13403 × 10 −5 ( Figure 6). bHLH69, also known as LJRHL1-LIKE 2/ DEFECTIVE REGION OF POLLEN 2 (LRL2/DROP2), regulates root hair and sperm cell development [29,30]. To examine whether LRL2 regulates the expression of the mild stress (-0.4 MPa)-responsive genes, we conducted dual-luciferase assays ( Figure 7A) by transiently expressing various effector and reporter combinations in Nicotiana benthamiana mesophyll cells. We observed that LRL2 could inhibit the promoter activity of EXT6, XTH14, AT3G49960, and TT6 ( Figure 7B). Furthermore, when co-transfected with AtSK11-GFP or AtSK12-GFP, the inhibition of EXT6 promoter activity by LRL2 was repressed ( Figure 7C,D). These results suggested that LRL2 functions downstream of AtSK11 and AtSK12 to regulate the expression of mild stress-responsive genes in Arabidopsis roots. Error bars indicate SD. Student's t-test was used to determine the significance of difference between WT and mutant under each treatment condition. Significant levels: * p < 0.05, ** p < 0.01, *** p < 0.001.

LRL2 Functions Downstream of AtSK11 and AtSK12 to Regulate −0.4 MPa-Responsive Gene Expression
As kinases, AtSK11 and AtSK12 are unlikely to directly regulate gene expression; therefore, we reasoned that transcription factors must be involved to mediate their effects on the mild stress-responsive gene expression. To identify these transcription factors, we performed promoter motif analysis of the mild stress-responsive genes ( Figure 6) and identified two conserved motifs. Motif 1 was enriched in the promoters of all 10 genes except PRP3; and motif 2 was enriched in the promoters of EXT6, EXT10, EXT11, EXT12, EXT13, AT4G08410, and XTH14 (Tables S1 and S2, Supplementary Materials). We compared the two motifs to the motifs in the DAP-seq databases. The basic helix-loop-helix (bHLH) transcription factor bHLH69 was predicted to bind to motif 1 (with p-value of 2.13403 × 10 −5 ( Figure 6). bHLH69, also known as LJRHL1-LIKE 2/ DEFECTIVE REGION OF POLLEN 2 (LRL2/DROP2), regulates root hair and sperm cell development [29,30]. To examine whether LRL2 regulates the expression of the mild stress (−0.4 MPa)-responsive genes, we conducted dual-luciferase assays ( Figure 7A) by transiently expressing various effector and reporter combinations in Nicotiana benthamiana mesophyll cells. We observed that LRL2 could inhibit the promoter activity of EXT6, XTH14, AT3G49960, and TT6 ( Figure 7B). Furthermore, when co-transfected with AtSK11-GFP or AtSK12-GFP, the inhibition of EXT6 promoter activity by LRL2 was repressed ( Figure 7C,D). These results suggested that LRL2 functions downstream of AtSK11 and AtSK12 to regulate the expression of mild stress-responsive genes in Arabidopsis roots. ( Figure 5A,B, Figure S6A-C, Supplementary Materials). These findings suggest that the extensin genes and TT6 may function downstream of AtSK11 and AtSK12 to inhibit the root elongation under mild osmotic stress (-0.4 MPa).

LRL2 Functions Downstream of AtSK11 and AtSK12 to Regulate -0.4 MPa-Responsive Gene Expression
As kinases, AtSK11 and AtSK12 are unlikely to directly regulate gene expression; therefore, we reasoned that transcription factors must be involved to mediate their effects on the mild stressresponsive gene expression. To identify these transcription factors, we performed promoter motif analysis of the mild stress-responsive genes ( Figure 6) and identified two conserved motifs. Motif 1 was enriched in the promoters of all 10 genes except PRP3; and motif 2 was enriched in the promoters of EXT6, EXT10, EXT11, EXT12, EXT13, AT4G08410, and XTH14 (Tables S1 and S2, Supplementary Materials). We compared the two motifs to the motifs in the DAP-seq databases. The basic helix-loophelix (bHLH) transcription factor bHLH69 was predicted to bind to motif 1 (with p-value of 2.13403 × 10 −5 ( Figure 6). bHLH69, also known as LJRHL1-LIKE 2/ DEFECTIVE REGION OF POLLEN 2 (LRL2/DROP2), regulates root hair and sperm cell development [29,30]. To examine whether LRL2 regulates the expression of the mild stress (-0.4 MPa)-responsive genes, we conducted dual-luciferase assays ( Figure 7A) by transiently expressing various effector and reporter combinations in Nicotiana benthamiana mesophyll cells. We observed that LRL2 could inhibit the promoter activity of EXT6, XTH14, AT3G49960, and TT6 ( Figure 7B). Furthermore, when co-transfected with AtSK11-GFP or AtSK12-GFP, the inhibition of EXT6 promoter activity by LRL2 was repressed ( Figure 7C,D). These results suggested that LRL2 functions downstream of AtSK11 and AtSK12 to regulate the expression of mild stress-responsive genes in Arabidopsis roots.

Discussion
This study provides several lines of evidence to demonstrate that GSK3 family members AtSK11 and AtSK12 have a negative role in mediating the increased root growth in Arabidopsis induced by mild osmotic stress (−0.4 MPa) treatment. First, the promoter-GUS analysis of 10 GSK3 family members revealed that only AtSK11, AtSK12, and AtSK13 are highly expressed in root tips, which are

Discussion
This study provides several lines of evidence to demonstrate that GSK3 family members AtSK11 and AtSK12 have a negative role in mediating the increased root growth in Arabidopsis induced by mild osmotic stress (−0.4 MPa) treatment. First, the promoter-GUS analysis of 10 GSK3 family members revealed that only AtSK11, AtSK12, and AtSK13 are highly expressed in root tips, which are important for sensing water [2,[23][24][25]. Second, using PEG to mimic osmotic stress conditions, the atsk11atsk12 double mutant was more sensitive in promoting root growth to −0.4 MPa treatment than the WT, whereas root growth in the other mutants and overexpression lines was similar to that of the WT. Third, RNA-seq analysis identified 12 downstream genes that might be responsible for the mild stress-induced root growth in atsk11atsk12. Fourth, phenotyping of the knockdown mutants of the mild stress-responsive genes suggested that extensin genes and TT6 might function downstream of AtSK11 and AtSK12 to inhibit root elongation under −0.4 MPa conditions. Fifth, promoter motif and transcription factor binding analyses of the 10 genes predicted that the bHLH transcription factor LRL2 may regulate the expression of stress-responsive genes. Indeed, LRL2 inhibits the promoter activity of EXT6, TT6, XTH14, and AT3G49960, as confirmed in a dual-luciferase assay, and AtSK11 and AtSK12 can largely suppress the inhibitory effect of LRL2 on the promoter activity of EXT6 in a dual-luciferase assay.
It is likely that AtSK11 and AtSK12 function in the root growth response under mild stress conditions by regulating cell wall homeostasis. Among the 10 genes regulated by AtSK11/AtSK12 under mild stress conditions, five are extensin genes (EXT6, EXT10, EXT11, EXT12, and EXT13), and one is an extensin-like family gene (AT4G08410). Among the 20 classical extensins in Arabidopsis, nine are specifically expressed in roots based on the Arabidopsis eFP Browser (expression data for four extensins are not available in the database). Notably, four of the nine root-specific extensins are present in our list ( Figure S8, Supplementary Materials) [31,32]. Extensin proteins are members of the cell wall hydroxyproline-rich glycoprotein (HRGP) superfamily that play key roles in cell wall homeostasis by functioning as scaffolds to facilitate self-assembly of the cell wall [33,34]. In addition to extensins, PRP3, a proline-rich structural cell wall protein [35], and XTH14, whose family members cleave xyloglucan chains and re-form bonds to the non-reducing ends of available xyloglucan molecules in primary cell walls, are also important for cell wall homeostasis [36]. It has been reported that the cell wall plays important roles in osmotic stress responses [3,34,37]. Osmotic stress leads to the accumulation of reactive oxygen species, resulting in cell wall stiffening through the crosslinking of cell wall glycoproteins such as extensins [3,38]. In addition, osmotic stress upregulates the expression of genes encoding xyloglucan-modifying enzymes to remodel the cell wall [3,38]. Therefore, it is likely that AtSK11 and AtSK12 take part in the root growth responses at −0.4 MPa by affecting cell wall homeostasis.
Several lines of evidence support the conclusion that LRL2 mediates the effect of AtSK11/AtSK12 on the expression of most mild stress-responsive genes. First, promoter motif and transcription factor binding analyses suggested that the transcription factor LRL2 binds to the promoters of all 10 mild stress-responsive genes except PRP3. Second, in dual-luciferase assays, LRL2 inhibited the promoter activity of EXT6, TT6, XTH14, and AT3G49960; this inhibitory function was repressed by co-transfection with AtSK11 or AtSK12 ( Figure 7B-D). Third, LRL2 is expressed in roots and root hairs, and functions in root and root hair development [29]. GSK3 family members are also involved in root hair development [30], and AtSK11 and AtSK12 are highly expressed in roots, including root tips (Figures 1B,C and S1B). Because AtSK11 and AtSK12 can inhibit LRL2 transcriptional activity in dual-luciferase assays, further studies could be conducted to explore how AtSK11 and AtSK12 regulate LRL2 activity.
In this study, we revealed the regulation mechanism of mild stress-induced root growth by GSK3-LRL2-cell wall-related genes, which could be useful in improving crop adaptation to drought stress. Osmotic stresses are caused by extreme water potentials. For sustainable agricultural production, it is critical to maintain soil water potential lower than −0.03 MPa but significantly greater than −1.5 MPa [39]. However, which water condition is best for plant growth remains elusive. Here, we demonstrated that the mild stress condition, −0.4 MPa, could facilitate root growth, and may thus be an ideal water potential for Arabidopsis, which could be useful information for future work to optimize agriculture irrigation.

Plant Materials, Plant Growth, and Generation of Multiple Mutants
Arabidopsis thaliana ecotype Col-0 was used as the wild type (WT) in this study. T-DNA insertion mutants atsk11 (SALK_014382), atsk12 (CS332559), and atsk13 (CS340333) were obtained from the Arabidopsis Biological Resource Center (ABRC), Columbus, OH, USA. Homozygous mutants were identified by PCR, and the corresponding primers are listed in Table S5 (Supplementary Materials).
Seeds were surface sterilized in 75% ethanol for 10 min, followed by absolute ethanol for 10 min. For general plant growth, the seeds were plated on 1/2 Murashige and Skoog (MS) medium (Phyto Technology Laboratories, Shawnee Mission, KS, USA) containing 0.5% sucrose and 0.8% agar. The pH of the medium was adjusted to 5.7 with KOH. After 2 days of stratification at 4 • C, the plates were transferred to a growth chamber (HiPoint, 740FLED, Gaoxiong, Taiwan, China) at 23 • C under a 16-h light/8-h dark cycle. After about 10 days of growth, the seedlings were moved to soil and grown in a greenhouse at 23 • C under a 16-h light/8-h dark cycle.
The atsk11atsk12 and atsk11atsk13 double mutants were generated by hybridization experiments with atsk11 as female parent and atsk12 and atsk13 as male parents. Then, we identified the wild-type, heterozygous, and homozygous mutant plants (ratio about 1:2:1) by genotyping PCR. For the atsk12atsk13 mutant, hybridization experiments were conducted with atsk12 as female parent and atsk13 as male parent. For the atsk11atsk12atsk13 mutant, hybridization experiments were conducted with atsk11atsk12 as female parent and atsk13 as male parent. However, atsk12atsk13 and atsk11atsk12atsk13 were not identified.

Osmotic Stress Treatment
For mannitol (Sigma-Aldrich, Saint Louis, MO, USA) treatment, seeds were plated on 1/2 MS medium, pH 5.7, containing 1% agar, vertical cultured for 4 days, and transferred to 1/2 MS medium containing 0, 100, 200, and 300 mM mannitol; the position of root tips was marked. Root length was measured after 3 days of growth. For PEG treatment, PEG (PEG8000, Sigma-Aldrich, Saint Louis, MO, USA)-infiltrated plates were prepared according to Verslues et al. [8] using 10 × 10 square Petri dishes and the PEG-infused plates were allowed to rest for 12 hours before use. The water potential of the plates was measured with a Dewpoint Potential Meter (WP4C, Pullman, WA, USA). Seeds were vertically grown in 1/2 MS medium (pH 5.7) containing 6 mM MES (Sigma-Aldrich, Saint Louis, MO, USA) and 1.5% agar (Sigma-Aldrich, Saint Louis, MO, USA) for 4 days, transferred to PEG-infiltrated plates; the position of root tips was marked and grown for 3 days. The plates were photographed, and root length from the marked position to the new position of the root tip was measured with ImageJ 1.48v software (National Institutes of Health, USA).

RNA-Seq Sample Preparation and Data Analysis
Four-day-old WT and atsk11atsk12 seedlings were transferred to PEG-infiltrated plates (−0.25 MPa, −0.4 MPa, −0.6 MPa). Following 3 days of growth, root tips (~5 mm) were collected from the plants. Each sequencing sample had two replicates. RNA was extracted from the samples using the TRIzol (Invitrogen, Dún Laoghaire, Dublin, Ireland) method. RNA-seq was performed using the Illumina HiSeq4000 platform (San Diego, CA, USA). Each sample generated approximately 23,000,000 clean reads (150 bp, PE). An average of 94.6% clean reads were mapped to the Arabidopsis thaliana genome from TAIR10 (ftp://ftp.arabidopsis.org/home/tair/Sequences/whole_chromosomes/) using TopHat2 [40] with default parameters. Differential expression analysis was performed with Cuffdiff software (Cambridge, MA, USA) (from the Cufflinks package [41], FDR < 0.05). The Fragments Per Kilobase Million (FPKM) value of WT and atsk11atsk12 under three water potentials are listed in Tables S3 and S4 (Supplementary Materials). Heatmaps were generated with pheatmap in R, which normalizes FPKM values by Z-score. GO enrichment analysis was performed with agriGO (http://systemsbiology.cau.edu.cn/agriGOv2/index.php). The RNAseq Raw data have been deposited in the National Center for Biotechnology Information (NCBI) database with Bio-Project ID PRJNA636228.

Quantitative Real Time-PCR
For Figure 1E, RNA was extracted from the root of 10-day-old vertical growth seedlings of WT and atsk11, atsk12, atsk13 homozygous mutants. For Figure S4, 4-day-old WT and atsk11atsk12 vertical growth seedlings were transferred to PEG-infiltrated plates (−0.25 MPa, −0.4 MPa, −0.6 MPa), following 3 days of growth; root tips (~5 mm) were collected from each genotype under different treatment conditions. For Figure S5, RNA was extracted from the root of 10-day-old vertical growth seedlings of microRNA homozygous lines (T3). For Figure S7 (Supplementary Materials), RNA was extracted from the root of 10-day-old vertical growth seedlings of gAtSK11-Flag and gAtSK12-Flag transgenic homozygous plants (T3). RNA was extracted by TRIzol method, and the first strand cDNA was synthesized using the Takara first-strand cDNA synthesis kit (2641A, Beijing, China). cDNA was mixed with the SYBR Green Master Mix (Vazyme, Q511-02, Nanjing, Jiangsu, China) for qRT-PCR. The qRT-PCR primers were designed with NCBI Primer-Blast (http://www.ncbi.nlm.nih.gov/tools/primer-blast/), and are listed in Table S5. Three biological replicates were generated for each sample, and qRT-PCR results were analyzed using the 2 −∆∆Ct method.

Plasmid Construction and Plant Transformation
The plasmids used to produce Flag-, GFP-, and GUS-tagged proteins were generated in the pCAMBIA 2306, pCAMBIA 2302, and pCAMBIA 1391 backbones, respectively. The promoter region of each gene included the region 1.5 kb upstream of the start codon, unless the sequence was <1.5 kb (in which case the longest sequence before the neighboring gene sequence was utilized). The genomic and promoter sequences of each gene were amplified from Col-0 genome DNA and cloned into the plasmid. Artificial microRNA was designed using the http://wmd3.weigelworld.org website and cloned into pCAMBIA1300 driven by its native promoter. The primer sequences used in this study are listed in Table S5. The constructs were transformed into Agrobacterium tumefaciens GV3101 cells, and transgenic plants were generated using the floral dip method [42]. The T0 seeds were screened by germinating on 1/2 MS solid medium with antibiotic selection. For each artificial microRNA transformation, at least six individual transgenic T3 lines were selected. Following qRT-PCR, the lines with the most downregulated genes were chosen for phenotypic analysis ( Figure S5).

GUS Staining and Immunoblot Analysis
For Figure 1B and Figure S1A, 5-day-old vertical growth of T3 GSK3pro::GUS transgenic homozygous lines were used for GUS staining, and at least three transgenic lines were analyzed. Histochemical staining of roots harboring the GUS reporter was performed as described [43].
For immunoblot analysis, the roots of gAtSK11-Flag and gAtSK12-Flag transgenic plants were ground to a fine powder in liquid nitrogen and solubilized with 5% SDS. After incubation at 95 • C for 5 min, the extracts were centrifuged at 12,000 rpm for 10 min. The resultant supernatant with 1× protein loading was separated on a 10% SDS-PAGE gel, then transferred to a nitrocellulose membrane (GE Healthcare Life Sciences, Marlborough, MA, USA), and the protein was detected with the corresponding antibodies. β-actin was used as an internal control. The gAtSK11-Flag and gAtSK12-Flag protein were detected by anti-Flag antibody (Abmart, Berkeley Heights, NJ, USA).

Confocal Microscopy
For Figures 1C and S1B,C, gAtSK11-GFP, gAtSK12-GFP, and gAtSK13-GFP seedlings were grown vertically in 1/2 MS for 5 days, and the roots were imaged under a Leica TCS SP8 Confocal Microscope (Buffalo Grove, IL, USA). GFP signals were observed in the epidermal, cortex, and stele cells of the root tip. For each transgenic material, three transgenic lines were observed, and for each line, at least five plants were observed.

Motif enrichment and Promoter Binding Analyses
The promoter sequences of the −0.4 MPa-responsive genes from 1500 bp upstream of their ATGs were extracted from the TAIR database and used to predict enriched motifs with an E-value < 0.05 in MEME software (V4.11.4, Reno, NV, USA) [44]. The enriched motifs were compared with known motifs from the DAP-seq database [45] using TOMTOM software (V4.11.4, Reno, NV, USA) [46] with a p-value < 10 −4 and a q-value < 0.05.

Dual-Luciferase Assay
The dual-luciferase assay was performed as described [43]. Agrobacterium strain GV3101 cells carrying the constructs were transformed into 4-week-old Nicotiana benthamiana leaves. Three days after infiltration, total proteins were extracted from lower leaf mesophyll cells and subjected to dual-luciferase analysis (E1910; Promega, Beijing, China), and three biological replicates were conducted.

Statistical Analysis
For Figures 1E and 7B-D, Figures S3B,D and S4, student's t-test was used to determine the significance level of the difference. For Figures 2 and 5, Figures S2 and S6, the F test was used to determine the variance and the two-tailed t-test with equal variance or unequal variance was used to determine the significance level of the difference. Asterisks indicate a significant difference (* p < 0.05, ** p < 0.01, *** p < 0.001).