The Arabidopsis small G‐protein AtRAN1 is a positive regulator in chitin‐induced stomatal closure and disease resistance

Abstract Chitin, a fungal microbial‐associated molecular pattern, triggers various defence responses in several plant systems. Although it induces stomatal closure, the molecular mechanisms of its interactions with guard cell signalling pathways are unclear. Based on screening of public microarray data obtained from the ATH1 Affymetrix and Arabidopsis eFP browser, we isolated a cDNA encoding a Ras‐related nuclear protein 1 AtRAN1. AtRAN1 expression was enriched in guard cells in a manner consistent with involvement in the control of the stomatal movement. AtRAN1 mutation impaired chitin‐induced stomatal closure and accumulation of reactive oxygen species and nitric oxide in guard cells. In addition, Atran1 mutant plants exhibited compromised chitin‐enhanced plant resistance to both bacterial and fungal pathogens due to changes in defence‐related genes. Furthermore, Atran1 mutant plants were hypersensitive to drought stress compared to Col‐0 plants, and had lower levels of stress‐responsive genes. These data demonstrate a previously uncharacterized signalling role for AtRAN1, mediating chitin‐induced signalling.

2019; Miya et al., 2007;Wan et al., 2008). Recognition of chitin by rice requires chitin elicitor binding protein (OsCEBiP), which contains an extracellular lysine motif but lacks an intracellular kinase domain (Akamatsu et al., 2013). An additional receptor, OsCERK1, interacts with OsCEBiP to form a heterodimer that transduces extracellular signals within cells (Shimizu et al., 2010). OsRacGEF1 and OsRLCK185 are direct substrates of OsCERK1 for transducing chitin signals (Akamatsu et al., 2013;Yamaguchi et al., 2013). In Arabidopsis, the cell surface receptor AtCERK1 binds to chitin, leading to homodimerzation of the receptor and subsequent activation of innate immunity (Liu et al., 2012;Wan et al., 2008). In addition to AtCERK1, the receptors AtLYK4 and AtLYK5 can bind to chitin and are essential to the chitin response (Wan et al., 2012). AtLYK5 has a higher affinity for chitin than AtCERK1, and is probably the primary receptor for chitin recognition . Although studies have been conducted to elucidate chitin signal transduction in induced disease resistance, many components of the transduction pathway, particularly those related to chitin-activated stomatal closure, remain unknown.
Stomata, which are natural openings in the plant surface bordered by guard cells, provide portals for pathogen penetration. In particular, foliar bacterial pathogens reach the interior of the plant mainly through stomata to cause disease. Active control of stomatal closure hampers pathogen invasion (Sawinski et al., 2013). When an Arabidopsis leaf is exposed to fungi or chitin, its guard cells respond by closing the stomatal aperture, hampering pathogen penetration (Lee et al., 1999). Chitin-induced stomatal closure is dependent on intracellular signalling via reactive oxygen species (ROS) and nitric oxide (NO), as demonstrated by pharmacological assays (Khokon et al., 2010). AtCERK1 is expressed in guard cells and is required for Fusarium oxysporum-triggered nonhost resistance and stomatal closure in Arabidopsis (Huang et al., 2017;Khokon et al., 2010;Liu et al., 2009). Furthermore, AtCERK1-PBL27 (receptor-like cytoplasmic kinase, RLCK)-SLAH3 (anion channel protein) constitutes a short signal transduction module that regulates chitin-induced stomatal closure and antifungal immunity (Liu et al., 2019). Characterization of more proteins involved in the guard cell and plant resistance signal transduction pathways will help to clarify the relationship between stomatal movement and disease resistance.

Stomatal movements are regulated by abiotic and biotic stresses,
including light, drought, CO 2 , humidity, microbes, plant hormones, and elicitors (Murata et al., 2015). The mechanisms underlying regulation of stomatal movement involve receptors, protein kinases, ion channels, and second messengers such as ROS and NO, indicating that stomatal guard cells of the plants harbour dynamic regulatory networks in plants (Shi et al., 2015). Molecular factors related to guard cell signalling induced by chitin could be exploited in the future as switches to activate resistance in the design of efficient strategies to protect crops.
By high-throughput down-regulation screening of a plant cDNA library, we identified several genes driving stomatal closure or plant cell death in response to elicitors, including NbVPE, NbGa, NbMAPKKKa, NbMEK2, NbWIPK, and NbALY916 (Teng et al., 2014;Zhang et al., 2010Zhang et al., , 2012aZhang et al., , 2012b. The evolutionarily conserved small-G protein RANs have emerged as key signalling proteins regulating multiple cellular processes in animals and yeast, including nuclear translocation of proteins and RNA, cell cycle regulation, and nuclear envelope maintenance (Reiner & Lundquist, 2018). RANs associated with guanine nucleotide exchange factors (GEFs) and GTPase-activation proteins (GAPs) act as molecular switches that are activated by GTP and inactivated by the hydrolysis of GTP to GDP (Vernoud et al., 2003). In animals, exportin-1 binds RAN and nuclear export signal (NES)-containing proteins to form a complex used for nuclear export and antiviral immunity (Heaton et al., 2019;Li et al., 2019). In plants, RAN can interact with AtXPOI, an Arabidopsis exportin-1 homologue, and AtRanBP1, an NES-containing protein, in biochemical assays (Haasen et al., 1999). RANs have been found to play roles in seed development, vegetative growth, and abiotic stress responses Xu & Cai, 2014;Xu et al., 2016), but their biological function has not been fully elucidated. Here, we found that Ras-related nuclear protein 1 (AtRAN1), which is a small G-protein, is highly expressed in Arabidopsis guard cells. Atran1 mutant plants exhibited impaired in chitin-induced stomatal closure, resistance to bacterial and fungal pathogens, and drought tolerance.
Our results suggest a role in chitin signalling for a small G-protein in Arabidopsis.

| Structure of AtRAN1
AtRAN1 contained five exons and four introns, a 5′ untranslated region of 80 nucleotides (nt), and a 3′ untranslated region of 347 nt ( Figure 1a). AtRAN1 is predicted to encode a protein of 221 amino acids containing a RAN domain protein ( Figure 1b). AtRAN1 contained four ATP/GTP-binding motifs, two effector molecule-binding motifs (GAP and GEP interaction sites), and two switch regions (Switch I and II). GTPase-activating protein and guanine-nucleotideexchange factor interaction sites (GEFs) were located at positions 44-49 and 97-104, respectively, which are conserved among various plant RANs. The highly acidic C-terminal motif DDDDDIFE was present at positions 214-219, although it differed slightly from other plant and insect RANs (Figure 1c).
Previous studies have shown that chitin triggers stomatal closure. Arabidopsis AtRAN1 is ubiquitously expressed during plant development, regulating seed development as well as abiotic stress tolerance (Haizel et al., 1997;Liu et al., 2014Liu et al., , 2019, and raising the possibility that AtRAN1 may be involved in the response to chitin. To study the potential role of AtRAN1 in chitin-induced stomatal closure, in silico analysis of AtRAN1 expression in Arabidopsis guard cells was performed with publicly available expression data using the ATH1 Affymetrix protocal (MIAMExpress, accession number: E-MEXP-1443, Yang et al., 2008) and Arabidopsis eFP browser (http://www.bar.utoro nto.ca/efp/cgi-bin/efpWeb.cgi; Winter et al., 2007), which contains microarray databases for guard cells and other tissues. AtRAN1 had a high expression level in guard cells ( Figure 2a). As expected from transcriptomic data, AtRAN1 mRNA was detected through quantitative reverse transcription PCR (RT-qPCR) in all tissues tested, and was up-regulated upon chitin treatment (Figure 2b,c,d).  which makes it a powerful system for testing subcellular distribution (Rolland, 2018). Confocal imaging indicated that the fluorescence signal was primarily confined to the nucleus, with a weak signal in the cytoplasm, whereas the control GFP was distributed throughout the cytoplasm and nucleus at 36 hr postinfiltration. Thus, AtRAN1 was mainly localized to the nucleus (Figure 2f). This finding is consistent with those from the rice and mammalian counterparts of AtRAN1 (Quimby & Dasso, 2003;Xu et al., 2016).

| Chitin-induced stomatal closure is compromised in Atran1 mutant plants
The increase in AtRAN1 expression in response to chitin prompted us to assess the role of AtRAN1 in chitin signalling. We employed a homozygous line containing T-DNA insertions of AtRAN1 from the SALK T-DNA insertion collection. The Atran1 mutants were verified through diagnostic PCR screening and DNA sequencing, and contained a T-DNA insertion 492 bp upstream of the ATG start codon.
Using RT-qPCR, we verified a large reduction in AtRAN1 transcript levels ( Figure 1d). Two rounds of backcrossing were performed to reduce possible genetic background effects. Atran1 mutants had no apparent phenotypic differences compared to Col-0 plants under physiological growth conditions. Various biotic and abiotic stresses can induce stomatal closure (Vahisalu et al., 2008). Given that AtRAN1 was up-regulated by chitin

| AtRAN1 acts upstream of ROS and NO production to generate chitin-induced stomatal closure
Both ROS and NO are essential second messengers that function in guard cell signalling. We examined the effects of a ROS scavenger (catalase) and a NO scavenger (cPTIO) on the chitin-induced stomatal closure ( Figure 4a). Chitin-induced stomatal closure was significantly inhibited by exogenous application of catalase and cPTIO, whereas catalase or cPTIO alone did not induce stomatal movement (data not shown). We further monitored ROS and NO synthesis in guard cells using the ROS-and NO-sensitive fluorescent dyes H 2 DCF-DA and DAF-2DA, respectively, by confocal microscopy. As shown in Figure 4, chitin affected both ROS and NO production, with obvious fluorescence observed in guard cells. Catalase sharply reduced chitin-induced ROS production, whereas cPTIO had no effect on chitin-induced ROS production ( Figure 4b). Similarly, chitin-triggered NO accumulation was markedly inhibited by catalase and cPTIO ( Figure 4c). As expected, exogenously applied hydrogen peroxide (H 2 O 2 , a major type of ROS) enhanced NO fluorescence in guard cells ( Figure S2), which is consistent with previous reports (Xie et al., 2014). Together, the correlation between stomatal movement and changes in ROS and NO levels indicate that ROS mediate chitin-induced stomatal closure via synthesis of NO. F I G U R E 2 AtRAN1expression in plants and AtRAN1 subcellular localization. (a) The expression ofAtRAN1in guard cells was monitored by microarray data from ATH1 Affymetrix (MIAMExpress, accession number: E-MEXP-1443; Yang et al.,2008) andArabidopsiseFP browser (http://www.bar.utoro nto.ca/efp/cgi-bin/efpWeb.cgi; Winter et al.,2007). (b) Gene Chip Operating Software-normalized expression levels ofAtRAN1in leaf cells and guard cells were retrieved usingArabidopsiseFP Browser. (c) Quantitative reverse transcription PCR (RT-qPCR) analysis of theAtRAN1expression in dry seeds, rosette leaf, cauline leaf, flower, root, and stem. (d) RT-qPCR analysis ofAtRAN1expression upon chitin treatment. In all cases,AtRAN1transcript levels were normalized toS16andEF1αlevels. Error bars indicateSE. Each bar represents an average of three independent reactions, including both biological and technical replicates. (e) Visualization ofAtRAN1promoter::green fluorescent protein (GFP) expression in the epidermis of transgenicArabidopsislines (AtRAN1promoter::GFP-1, 2, 3) by confocal microscopy. (f) Transient expression of AtRAN1-GFP fusion protein inNicotiana benthamianaleaves. AtRAN1-GFP expression was driven by cauliflower mosaic virus 35S promoter and transiently expressed inN. benthamianaleaves by agroinfiltration Next, we investigated the relationships among AtRAN1, ROS, and NO in the guard-cell chitin-signalling network. We observed obvious ROS fluorescence in Col-0 guard cells after chitin treatment.
Only low-intensity ROS fluorescence was observed after mock treatment. Fluorescence intensity in Atran1 mutant guard cells was

| AtRAN1 loss-of-function affects chitin-induced disease resistance to bacterial and fungal pathogens
To further investigate the roles of chitin signalling and AtRAN1 in plant defence, both Col-0 and Atran1 mutant plants were pretreated with chitin and their responses to different pathogens were examined. We tested the effects of chitin pretreatment on infection by Pseudomonas syringae pv. tomato (Pst) DC3000 of Col-0 and Atran1 mutant plants.
One day after chitin treatment, plants were dip-inoculated with 10 8 cfu/ml of Pst DC3000. Mock-treated Atran1 mutant plants were F I G U R E 3 Sensitivity of theAtran1mutant to chitin treatment. Leaves of Col-0,Atran1, and gAtRAN1plants with open stomata were exposed to either mock or 50 µg/ml chitin. Representative images from one of these experiments are shown. The stomatal apertures were measured after a 2-hr incubation. At least 60 stomata were measured for each genotype per replication. Data were means ± SEof three independent experiments more susceptible than Col-0 to Pst DC3000, as indicated by a 7.62fold greater number of bacteria present 4 days after inoculation as well as increased chlorotic spots on day 5 (Figure 6a). In addition, bacterial abundances in chitin-treated Atran1 mutant plants were reduced by 2.59-fold compared to those in mock-treated Atran1 plants, reflecting the general induction of plant innate immunity in both Col-0 and  did not differ before drought stress, but were 1.2-fold to 1.5-fold lower in response to drought in Atran1 mutant plants than in Col-0 plants (Figure 8).

| D ISCUSS I ON
The observations described here convincingly show that Arabidopsis AtRAN1 is highly expressed in guard cells and plays essential roles in chitin-responsive control of stomatal closure, activation of ROS or NO production, and disease resistance, as well as drought stress tolerance. Therefore, we suggest that AtRAN1 could be explored for improving plant tolerance to abiotic and biotic stresses.
RAN is an evolutionarily conserved eukaryotic GTPase. RAN is considered an essential factor for nuclear transportation of protein F I G U R E 6 The resistance ofAtran1mutant plants toPseudomonas syringaepv.tomato(Pst) DC3000 andSclerotinia sclerotiorumin response to chitin. (a) Disease symptoms caused by Pst DC3000 on individual leaves of indicatedArabidopsislines. Plants were photographed 5 days postinoculation and 1 day after chitin treatment. Representative two leaves from one of three experiments are shown. Quantitative analysis of bacterial growth in planta after inoculation with Pst DC3000. (b) Five-week-old mature plants were pretreated (24 hr before pathogen inoculation) with chitin, and then, leaves were inoculated withS. sclerotiorumby attaching a 4 mm diameter hyphal plug onto surface of detached leaves. Representative two leaves from one of three experiments are shown. Lesion size was measured 24 hr after inoculation. Means ± SEof three independent experiments are shown. Statistical analysis was carried out using a two-way analysis of variance (ANOVA) with GraphPad Prism software. Different letters denote statistically significant differences among different treatments as determined by ANOVA (LSD test,p < .05) in animal cells and plays a role in animal immunity against virus infection (Han & Zhang, 2007). Recent studies have demonstrated that plant RAN might participate in cell cycle control, postmitotic nuclear assembly, auxin sensitivity, and response to environmental stress (Chen et al., 2011;Wang et al., 2006;Zang et al., 2010). The results presented here support the pivotal roles of Arabidopsis RAN in stomatal movement, disease resistance, and the drought stress response. These results are consistent with previous findings that Arabidopsis and rice plants overexpressing RAN show improved cold tolerance (Xu & Cai, 2014;Xu et al., 2016), and provide compelling evidence that these RANs are essential to the abiotic stress response, revealing both conserved and plant-specific roles in accordance with their counterparts in animals. and AtRANGEFs have the capacity for GEF activity, further identification of AtRANGEFs that activate AtRAN1 and kinases that function downstream of AtRAN1 will help to clarify chitin signalling in Arabidopsis. We found that Arabidopsis AtRAN1, similar to AtCERK1, can regulate chitin-induced stomatal closure, as well as plant resistance. Atran1 mutation did not affect chitin-induced AtCERK1 expression, and that interaction analyses for AtRAN1 with AtCERK1 and other chitin signalling components should be conducted. As similar chitin-recognition patterns mediated by CERK1 are found in Arabidopsis and rice, whether Arabidopsis AtRAN1 employs a similar mechanism to that of rice OsRac1 in response to chitin remains to be investigated.

Studies of several crop species, including rice, tomato, and
Arabidopsis, showed that proteins can regulate elicitor-triggered F I G U R E 7 Expression analyses of the indicated defence-responsive genes inAtran1mutant plants in response to chitin. Samples (n = 5) with or without chitin treatment were collected. Values represent means ± SEfrom three independent experiments; different letters denote statistically significant differences among different treatments as determined by ANOVA (LSD test,p < .05) stomatal movement and plant resistance (Du et al., 2014;Qiu & Yu, 2009). The tomato NAC transcription factor JA2 is highly expressed in guard cells of tomato leaves. JA2-silenced tomato plants show impaired Pst DC3000-and flg22-induced stomatal closure (Du et al., 2014). Furthermore, JA2-silenced plants contain greater pathogen biomass than wild-type plants during pathogen assays (Du et al., 2014). In our study, AtRAN1 is a guard-cell expressed protein and stomatal closure was less sensitive to chitin treatment in the Atran1 mutant than in Col-0 plants. Consistent with a positive role in chitin-triggered stomatal closure, Atran1 mutant plants contained more Pst DC3000 biomass than wild-type plants upon chitin treatment. This result may be interpreted as due to compromised stomatal closure in Atran1 mutant plants, which provides a portal through which more bacteria move into the stoma and get into leaf apoplast to cause disease. As chitin is an important component of the fungal cell wall, enhanced susceptibility to S. sclerotiorum upon mutation of AtRAN1 is unsurprising. These data raise the possibility that AtRAN1 may also be involved in the response to bacterial pathogen-associated molecular patterns (PAMPs), although this hypothesis remains untested. In addition, Atran1 mutant plants showed reduced tolerance to drought stress, suggesting AtRAN1 plays dual roles in biotic and abiotic signalling, and is involved in the crosstalk between signalling pathways activated by abiotic and biotic stresses. In a recent research study, the plant drought stress is influenced by the hormone abscisic acid (ABA) (Yan et al., 2020), and thus further investigation of ABA signalling in relation to AtRAN1-mediated chitin signalling in the context of stomatal closure or drought stress will help to identify the common and unique processes in the response to biotic and abiotic stresses. Plants are simultaneously subjected to a variety of stresses. Understanding the crosstalk in converging stress response pathways may be helpful for increasing crop stress tolerance and crop production.  Trujillo et al., 2006). Our results showed that the expression pattern of RbohD was similar in Atran1 mutant and Col-0 plants with and without chitin treatment, while RbohF expression was down-regulated in Atran1 mutant upon chitin treatment, suggesting involvement of Rboh in an AtRAN1-dependent pathway related to chitin-induced ROS accumulation. NIA is important in the biosynthesis of NO, and Arabidopsis NIA-dependent NO production is key to the establishment of resistance to disease caused by Pst DC3000 (Lozano-Juste & Leon, 2010;Vitor et al., 2013). The expression level of NIA was reduced in the Atran1 mutant after chitin treatment. This result suggests that chitin-induced AtRAN1mediated stomatal closure and resistance to pathogens may be related to Rboh-dependent ROS accumulation and NIA-dependent NO accumulation. Our previous work and this study indicate that elicitors with different origins can induce ROS and NO production (Zhang et al., 2010(Zhang et al., , 2016a(Zhang et al., , 2016b. Therefore, these results indicate that intracellular ROS and NO accumulation is shared among all signalling pathways in guard cells, and signalling components such as AtRAN1 mediate the intracellular accumulation of ROS and NO in guard cells. Such pathways may interconnect, with specificity of plant signal transduction achieved through transcriptional changes in response to different pathogen elicitors. Previous studies have reported the collinearity of chitin, ROS, and NO in guard cell signalling (Khokon et al., 2010). Pharmacological evidence indicates that ROS-induced NO generation in guard cells is related to both nitric oxide synthase and nitrate reductase activity (Khokon et al., 2010). This finding suggests further complexity of chitin-guard cell signalling, which has not yet been fully described. Collectively, these data suggest a role for AtRAN1 upstream of ROS and NO production in chitin-mediated stomatal closure signalling.  (Clarke et al., 2000;Robert-Seilaniantz et al., 2011). In addition, AtRAN1 mutation did not affect AtCERK1 expression upon chitin treatment, suggesting that AtRAN1 may act downstream of AtCERK1 or in an AtCERK1independent manner in chitin signalling. Thus, AtRAN1 may function as a positive regulator of SA signalling pathways related to chitin-triggered plant defences. Further investigations of the roles of AtRAN1 in disease resistance to various biotic stresses, such as bacterial flg22 and fungal chitin, will help to reveal the relationships among AtRAN1-mediated stomatal movement, stomatal immunity, and apoplast immunity in plants.

Mutation of
Here, we identified and characterized AtRAN1 as a guard cell-localized small G-protein that responds to chitin and drought. AtRAN1 mutation impaired chitin-induced stomatal closure, leading to lower ROS and NO accumulation in guard cells. Moreover, AtRAN1 mutation compromised chitin-induced plant resistance to Pst DC3000 and S. sclerotiorum, while also decreasing expression levels of defence-related genes. In addition, AtRAN1 may be a positive regulator of the drought response via pathways mediated by stress responsive genes. Follow-up research will aim to identify the targets of AtRAN1 to further elucidate the signalling pathways used by plants exposed to abiotic and biotic stresses.

| Bioinformatics analysis of AtRAN1
The entire AtRAN1 open reading frame (ORF) was amplified from Arabidopsis cDNA by RT-PCR using primers (AtRAN1-F and AtRAN1-R), according to AGI ID At5g20010. The PCR product was gel purified, ligated into pMD19 (TaKaRa), and transformed into  (Winter et al., 2007). First, choose the data source of tissue specific from the Arabidopsis eFP Browser, then fill the AGI ID (Arabidopsis Genome Initiative identifier) of candidate genes. Upon submission in the "Absolute" mode, the plant tissues are coloured ranging from yellow to red, according to the expression level of the gene of interest. The genes with high expression level in guard cells were selected to screen the most reliable reference genes for chitin-triggered expressions via RT-qPCR.

| Plasmid construction and Arabidopsis transformation
The AtRAN1 promoter was cloned by PCR using gene-specific primers (AtRAN1promoter-F and AtRAN1promoter-R) that amplified nucleotides −612 to +6, and the resulting AtRAN1 promoter fragment containing all the required cis-regulatory elements was directly cloned into the pBIGFP (S65T) binary vector (Kang et al., 2008) using the SalI and BamHI unique restriction sites.
To make transient expression construct pBinGFP2::AtRAN1, the AtRAN1 gene was cloned using cDNA from Arabidopsis with primers AtRAN1-TL-F and AtRAN1-TL-R. The PCR products were inserted into KpnI-and SalI-digested pBinGFP2, creating a fusion with the GFP coding region.
For the molecular complementation experiment, a 2.202-kb fragment that included a 557-bp region upstream of the start codon and a 261-bp region downstream of the stop codon of AtRAN1 was amplified by PCR from genomic DNA using the primers AtRAN1-ResF and AtRAN1-ResR, which introduced KpnI and BamHI sites at each end, respectively. The resulting PCR product was cloned into pCAMBIA1300 (CAMBIA) using the KpnI and BamHI sites, producing plasmid pCAM-BIA1300::gAtRAN1. Constructs were verified by sequencing and used to transform Col-0 and Arabidopsis Atran1 mutant plants using a modified floral dip procedure (Clough & Bent, 1998).

| Analysis of GFP expression
Tissues from plants containing GFP constructs were analysed using an FV1000 confocal laser scanning microscope (Olympus) to determine GFP expression in guard cells. The abaxial epidermal strips from independent seedlings leaves were peeled gently, mounted on a microscope slide, and examined with the microscope, with an argon laser at a wavelength of 488 nm and emission between 500 and 530 nm, as described by Zhang et al. (2018) and Liang et al. (2005), and they displayed coherent results.

| Pathogen assays
Pst DC3000 was provided by Barbara Kunkel (Washington University, USA). Bacteria were cultivated at 28 °C and 220 rpm in King's B medium containing 50 mg/ml rifampicin. Pst DC3000 inoculation assays were performed as described previously (Yao et al., 2013). A bacterial suspension (10 8 cfu/ml) was dip-inoculated onto rosette leaves of 5-week-old A. thaliana. Leaf discs collected 4 days postinoculation (dpi) were washed twice with sterile water and homogenized in 10 mM MgSO 4 . Bacteria were enumerated at 4 dpi as described by Yang et al. (2015). Each biological repeat comprised nine leaf discs from three separate plants.
S. sclerotiorum NGA4 was grown on potato dextrose agar as described in the method of Zhang et al. (2014). Leaves were detached from 5-week-old A. thaliana and transferred onto filter paper saturated with sterile distilled water in a Petri dish. A 4-mm diameter hyphal plug was then placed upside down onto each leaf and kept under a 16 hr day/8 hr night regime at 25 °C.
Pictures of the lesions were taken at 1 dpi, and lesion diameters were recorded.

| Stomatal bioassays
Stomatal closure bioassays were performed as described previously (Zhang et al., 2016b). Two whole leaves were excised per genotype

| RT-qPCR analysis
Total RNAs were extracted from seedlings using TRIzol (Invitrogen) according to the manufacturer's instructions. To eliminate genomic DNA contamination, RNA was treated with DNase I (TaKaRa) for 20 min.
First-strand cDNA was synthesized from total RNA using the TaKaRa RNA PCR kit. Real-time PCRs were performed on an ABI 7500 realtime PCR system using IQ5 multicolor real-time PCR master mix (Toyobo). Each PCR (30 µl) contained 15 µl 2 × SuperReal Premix Plus, 0.25 µl each primer, and appropriate cDNA. PCR was performed under the following conditions: 95 °C for 15 minfollowed by 40 cycles of 95 °C for 10 s, 55-60 °C for 30 s, and 72 °C for 32 s. Melt curves were collected by the ABI 7500 system and followed by 95 °C for 15 s, 60 °C for 1 min, and then 95 °C for 30 s. Ribosomal S16 (S16, At2g09990) and elongation factor 1α (EF1α, At5g60390) were used as internal references for all PCR analyses. The gene-specific primers used for RT-qPCR are listed in Table S1.

| Drought stress and water loss
To apply drought stress, the protocol described by Seki et al. (2001) was used with minor modification. Seedlings were grown under well-watered conditions for 4 weeks and subsequently deprived of water for 2 weeks. Next, the plants were rewatered for at least 3 days and photographed. For water-loss assays, rosette leaves were collected from 5-week-old plants as test samples. The samples were weighed immediately on a piece of paper and placed on the laboratory bench (relative humidity 50%, 22-23 °C). The weight lost by each sample at preassigned time points (0, 1, 2, 3, 4, and 5 hr) was recorded.

| Statistical analysis
For statistical analysis, one-way analysis of variance (ANOVA) or two-way ANOVA tests were performed with GraphPad prism 7.00 software.

ACK N OWLED G M ENTS
This work was supported in part by the National Key R&D Program

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available in Arabidopsis eFP Browser at http://bar.utoro nto.ca/efp_arabi dopsi s/ cgi-bin/efpWeb.cgi, reference number At5g20010. The GenBank accession number and Arabidopsis Genome Initiative (AGI) locus identifier of the AtRAN1 gene are as follows: NM_122008.3 (GenBank), At5g20010 (AGI locus). Other data that support the findings of this study are available from the corresponding author upon reasonable request.