The B. distachyon Bd21 was cultured and used for gene cloning and expression analysis. The Arabidopsis thaliana ecotype Colombia (Col-0) was used as WT control. The pSN1301-BdPP2CA6 plasmid was transformed into Arabidopsis using Agrobacterium tumefaciens strain EHA105 as described previously (Clough and Bent, 1998). Arabidopsis transformed with pSN1301 empty vector was used as vacant vector control (VC). Seeds of transformed Arabidopsis were selected using MS medium supplemented with 20 mg/L hygromycin. Homozygous lines of T₃ and T₄ generations were used for further analysis.

All plant materials were grown under a 14 h light/10 h dark cycle and 70% relative humidity at 22°C. For organ-specific expression assay, different organs including root, stem, leaf, and caryopsis were taken, respectively, from 3-week-old B. distachyon plants. For abiotic stress treatments, 3-week-old B. distachyon were treated with 100 μM ABA, 200 mM NaCl, or 20% (w/v) PEG-6000, respectively, and then leaves were sampled at 0, 1, 3, 6, 12, and 24 h after treatments.

To identify PYLs and clade A PP2Cs in B. distachyon, PYL and clade A PP2C gene sequences of Arabidopsis were downloaded from TAIR database¹, which were then used to do BLAST searches against Ensembl Plants database². All resulting proteins were confirmed to contain certain domains using Pfam database³. Identified sequences of B. distachyon were aligned using ClustalX2.0 with default parameters (Larkin et al., 2007). Phylogenetic trees were constructed using bootstrap neighbor-joining (NJ) method and bootstrap analysis (1,000 replicates) by MEGA6.0 (Tamura et al., 2013). Promoter sequences and exon-intron data were obtained from Plaza database⁴. Abiotic stress-responsive cis-elements in promoters were analyzed using PlantCare database⁵. Expression heatmaps of BdPP2CAs and BdPYLs were constructed as described previously (Chen L. et al., 2016).

A total of 7 BdPP2CAs and 12 BdPYLs were amplified by RT-PCR. For yeast two-hybrid assay, ORFs of BdPP2CAs and BdPYLs were amplified from B. distachyon cDNA and cloned into vectors pGBKT7 and pGADT7, respectively. The pGADT7-BdSnRK2.2, pGADT7-BdSnRK2.3, and pGADT7-BdSnRK2.6 vectors were obtained from our laboratory (Wang et al., 2015). The ORFs of the AtSnRK2.2, AtSnRK2.3, and AtSnRK2.6 were amplified from Arabidopsis Col-0 WT cDNA and were cloned into vector pGADT7. For BiFC assay, ORFs of BdPP2CA6 were cloned into vector pUC-SPYNE, while BdPYL11, and AtSnRK2.3 were cloned into vector pUC-SPYCE. For GFP fusion protein expression, 35S::BdPYL11::GFP, 35S::BdPP2CA6::GFP, and 35S::AtSnRK2.3::GFP were constructed. For plant transformation, ORF of BdPP2CA6 was cloned into vector pSN1301 driven by CaMV 35S promoter.

All primers used are listed in Supplementary Table S2 and all amplified fragments were confirmed by sequencing. All accession numbers of the genes mentioned in manuscript are listed in Supplementary Table S3.

Each pair of constructed plasmids were co-transformed into yeast strain AH109 following the manufacturer’s protocol (Clontech Inc., USA) and subsequently plated onto SD minimal medium without leucine and tryptophan (SD/-Leu-Trp). Then yeast cells were transferred to specific selection media for yeast growth assessments.

The constructs of BdPYL11-GFP and BdPP2CA6-GFP were transformed, and constructs of BdPP2CA6-YNE and BdPYL11-YCE or AtSnRK2.3-YCE were co-transformed into onion epidermal cells by using biolistic method (Biolistic PDS-1000/He Particle Delivery System, Bio-Rad, USA), respectively (Hanson and Kohler, 2001). Fluorescence was observed by fluorescence microscope (IX71, Olympus, Japan) 1 day after transformation. DAPI (4′,6-diamidino-2-phenylindole) staining was performed as described previously (Sun et al., 2015).

Total RNAs were isolated using Plant RNA Extraction Kit (ZOMANBIO, China). The cDNAs were obtained by using FastQuant RT Kit (TIANGEN, China). The qRT-PCR analysis was performed using SYBR Green Kit (TIANGEN, China) on a real-time PCR machine (CFX96^TM Real-time Detection System, Bio-Rad, USA). Gene-specific primer sets were listed in Supplementary Table S2 and some of them were designed according to previous studies (Fujii et al., 2007; Feng et al., 2014; Chen Z.H. et al., 2016). All amplified fragments were confirmed by sequencing. The qRT-PCR analysis was performed as described previously (Hu et al., 2013).

For stomatal aperture measurement, leaves of 4-week-old WT and transgenic plants were collected and incubated in solution containing 30 mM KCl, 10 mM MES-Tris, and 50 μM CaCl₂ (pH 6.15) with light for 2 h. Then 50 μM ABA or 150 mM NaCl was added to the solution. After 2 h treatments, leaf samples were examined using microscope (IX71, Olympus, Japan). Widths and lengths of stomata were measured to figure out stomatal apertures.

For ABA treatment, 7-day-old seedlings germinated on 1/2 MS medium were transferred to MS medium or MS media supplemented with 1, 5, or 10 μM ABA for 10 days. For salinity stress tolerance assay at early development stage, 7-day-old seedlings were transferred onto MS medium or MS medium supplemented with 150 mM NaCl for 10 days. For physiological index measurement and qRT-PCR analysis, 4-week-old seedlings cultured in soil were watered with or without 150 mM NaCl for 10 days and then leaf samples were collected. Survival rates of the treated Arabidopsis seedlings were measured 35 days post NaCl treatment and pictures of the seedlings were taken 30 days post NaCl treatment.

For ion content determination, treatment of plants and sampling of leaves were performed as described previously with slight modification (Deng et al., 2013). One-week-old seedlings culturing on MS medium were transferred onto MS medium supplemented with 150 mM NaCl for 1 week and then leaves were sampled. Measurement was carried out by using Atomic Absorption Spectrometry (ContrAA 700, Analytik Jena, Germany).

Measurements of ion leakage (IL), H₂O₂, proline, chlorophyll, and malondialdehyde (MDA) content were performed as described previously (Huang et al., 2015).

Statistical analyses were performed using Excel (Microsoft, USA) and SPSS (IBM, USA). Figures were plotted by using OriginPro (OriginLab, USA).

To identify BdPYL and clade A BdPP2C proteins from B. distachyon, BLAST searches were performed. Twelve putative BdPYL proteins and eight putative clade A BdPP2C proteins were identified from the B. distachyon genome. Phylogenetic analyses were performed to construct phylogenetic trees by using full length amino acid sequences of BdPYL family or clade A BdPP2C subfamily, and the results showed that key domains of BdPYL proteins or clade A BdPP2C proteins are evolutionary conserved (Figure 1).

To analyze the genomic distribution of BdPYLs and clade A BdPP2Cs, their approximate positions on each chromosome were marked. The results showed that six BdPYLs distributed on chromosome 2 and the rest members were localized on chromosomes 1 and 3, while six clade A BdPP2Cs distributed on chromosome 2 and the rest members were mapped on chromosomes 1 and 4 (Supplementary Figure S1). All BdPYLs and clade A BdPP2Cs were designated according to their distribution on chromosomes as listed in Supplementary Table S1.

The exon–intron structures of BdPYLs and BdPP2CAs were also analyzed (Figure 1). For BdPYL family, only three BdPYLs contains two or more introns and the rests contain no intron at all. For BdPP2CA subfamily, all members except BdPP2CA8 have two or three introns. These conserved intron and extron numbers in both BdPYL family and BdPP2CA subfamily support their close evolutionary relationships.

To identify abiotic stress-responsive cis-elements in promoters, 1,500 bp sequences upstream of BdPYL and BdPP2CA CDSs were obtained, and several abiotic stress-responsive cis-elements (ABRE, CE3, CGTCA-motif, GARE-motif, HSE, TGACG-motif, MBS, TCA-element) were identified. The results demonstrate that 50% of the BdPYLs contain ABRE elements in their promoter regions and only BdPP2CA5 has no ABRE element in its promoter region, implying that many BdPYLs and BdPP2CAs maybe ABA responsive. Results also show that these genes have one or more additional stress-responsive cis-elements in their promoter regions, suggesting their potential involvements in various stress responses (Supplementary Table S1).

To analyze the potential functions of BdPYLs and BdPP2CAs under environmental stresses, public microarray data were obtained to analyze their expressions under cold, heat, drought, and high-salinity conditions (Priest et al., 2014). The results showed that expressions of most BdPYLs were not obviously changed under stress treatments, while BdPP2CA6 was significantly up-regulated under high-salinity stress and BdPP2CA8 was significantly up-regulated under drought stress (Supplementary Figure S2).

Although we tried hard to clone BdPP2CA3 from the B. distachyon cDNA, we failed to clone the full length of it. The rest 7 BdPP2CAs and all identified BdPYLs were successfully cloned. In Arabidopsis, ABA-bound PYLs interact with clade A PP2Cs and inhibit their activities (Melcher et al., 2009; Miyazono et al., 2009; Yin et al., 2009). To investigate their interaction networks in B. distachyon, we examined the interactions between BdPP2CAs and BdPYLs in yeast two-hybrid assay. Data of those BdPP2CAs/BdPYLs which could not interact with any BdPYLs/BdPP2CAs were not shown. The result showed that BdPP2CA4, BdPP2CA5, BdPP2CA6, BdPP2CA8 interact with 7 or 8 or 9 BdPYLs when supplied with exogenous ABA (Figure 2A). In the absence of exogenous ABA, fewer BdPYLs interact with these four BdPP2CAs (BdPP2CA4, BdPP2CA5, BdPP2CA6, and BdPP2CA8) and only BdPYL11 can interact with all of them (Figure 2B).

As interactions between BdPYL11 and BdPP2CA4, BdPP2CA5, BdPP2CA6, or BdPP2CA8 are strong and ABA independent, qRT-PCR analyses were performed to study their potential functions in response to abiotic stresses. Organ-specific expression analysis was also carried out, and the results demonstrated that BdPP2CA8 expressed only in roots while BdPYL11, BdPP2CA4, BdPP2CA5, and BdPP2CA6 were detected in roots, as well as in leaves, stems, and caryopses (Figure 3A).

In Arabidopsis, PYLs were strongly down-regulated while clade A PP2Cs were strongly up-regulated by ABA treatment (Merlot et al., 2001; Saez et al., 2004; Santiago et al., 2009). The qRT-PCR results reveal that expression patterns of the investigated genes in B. distachyon are consistent with their homologs in Arabidopsis. BdPYL11 was down-regulated under ABA, NaCl and PEG6000 treatments while BdPP2CA4, BdPP2CA5, BdPP2CA6, and BdPP2CA8 were up-regulated (Figures 3B–E). These results suggest that these genes respond to various abiotic stresses.

Without exogenous ABA, BdPP2CA6 only interacts with BdPYL11 in yeast two-hybrid assay. So we were more interested in interaction between these two proteins. GFP fusion vectors were constructed and transferred into onion epidermal cells to investigate their subcellular localizations. BdPYL11-GFP fusion protein was observed in the nucleus, plasma membrane, and cytoplasm, while BdPP2CA6-GFP fusion protein was observed only in nucleus (Figure 4A). To examine whether BdPP2CA6 interacts with BdPYL11 in vivo, we further carried out BiFC assay. YFP fluorescence was observed in nucleus, suggesting that BdPP2CA6 interact with BdPYL11 in vivo (Figure 4B).

BdPP2CA6, a clade A PP2C gene, was greatly induced by ABA, PEG6000 and NaCl treatments (Figure 3). BdPP2CA6 exhibited a strong interaction with BdPYL11 in the absence of ABA (Figures 2A and 4A). Transient expression of BdPP2CA6-GFP fusion protein in onion epidermal cells demonstrates that BdPP2CA6 localizes only at the nucleus (Figure 4B). Therefore we focused on investigating BdPP2CA6 and obtained BdPP2CA6 overexpression Arabidopsis plants for further analyses.

To investigate whether overexpression of BdPP2CA6 could affect ABA response, 7-day-old WT, VC and BdPP2CA6-overexpression Arabidopsis seedlings were placed on MS medium and MS media supplemented with different concentrations of ABA, respectively, for 10 days. ABA-mediated inhibition of root growth was affected by overexpression of BdPP2CA6 (Figures 5A–D). Root elongations of BdPP2CA6-overexpression plants were inhibited with higher degrees than that of WT and VC by all concentrations of ABA. Fresh weights and root lengths were measured and were significantly higher for WT and VC plants than for transgenic plants at all ABA concentrations (Figures 5G,H). The results demonstrated that overexpression of BdPP2CA6 enhanced ABA sensitivity at early growing stage of plants.

Whether BdPP2CA6 could regulate ABA-mediated stomatal closure was investigated. Under normal condition, stomatal apertures of transgenic seedlings were a little bit smaller than that of WT. After ABA treatment, stomatal apertures of all seedlings decreased, with a much more significant decrease in transgenic plants compared with WT (Figure 6). These results suggest that overexpression of BdPP2CA6 promotes ABA-mediated stomatal closure.

To evaluate the effects of salinity stress on growth of BdPP2CA6-overexpression plants, 7-day-old WT, VC and BdPP2CA6-overexpression plants were placed on MS medium and MS medium supplemented with 150 mM NaCl, respectively. After 10 days, no obvious difference was discovered at post-germination developmental stage on MS media. The addition of 150 mM NaCl significantly inhibited root elongation of WT but not of BdPP2CA6-overexpression plants (Figures 5A,E). Fresh weights and root lengths were evaluated and the average fresh weight and root length of BdPP2CA6-overexpression plants were higher than that of WT and VC at 150 mM NaCl concentration (Figures 5G,H). In addition, growth of 2-week-old seedlings under salinity treatment was measured. Two-week-old WT and BdPP2CA6-overexpression plants growing on MS medium were transferred to Hoagland solution supplemented with 0 or 150 mM NaCl. Most WT plants wilted after 1 day, while most BdPP2CA6-overexpression transgenic Arabidopsis remained green, indicating that BdPP2CA6 is involved in responses to salinity stress at the early growing stage (Figure 7B).

For NaCl treatment, 3-week-old WT, VC and BdPP2CA6-overexpression plants growing in soil were watered with 150 mM NaCl for 30 days. Most WT plants turned purple or wilted whereas most transgenic plants remained green (Figure 7A). Survival rates of transgenic seedlings were measured 35 days post NaCl treatment and were significantly higher than that of WT and VC plants (Figure 7C). As we found that overexpression of BdPP2CA6 could promote ABA-mediated stomatal closure, we assumed that BdPP2CA6-overexpression plants might be able to withhold more water to cope with external salinity through stomatal closure. To confirm this possibility, we measured stomatal apertures of leaves from WT and BdPP2CA6-overexpression plants under NaCl treatment. Under normal condition, the differences between WT and BdPP2CA6-overexpression plants were not significant, the stomatal aperture indices of transgenic plants after salinity treatment decreased more rapidly than that of WT, indicating that transgenic plants could withhold more water to cope with salinity stress (Figure 6).

A higher K⁺/Na⁺ ratio is important for plants to cope with salinity stress (Zhu, 2003). To investigate whether overexpression of BdPP2CA6 could affect Na⁺ and K⁺ accumulations in plants, the Na⁺ and K⁺ contents were measured. No significant difference in Na⁺ contents was observed between WT and transgenic plants under normal condition or NaCl treatment. Whereas K⁺ contents of transgenic plants were significantly higher than that of WT under both normal condition and NaCl treatment (Figures 7I–K). The results demonstrated that overexpression of BdPP2CA6 increased K⁺ accumulation, thus increasing the ratio of K⁺/Na⁺ under salinity stress condition.

Several physiological indices including IL, H₂O₂, proline, chlorophyll, and MDA contents were measured to investigate the physiological mechanisms in BdPP2CA6-overexpression plants in response to salinity stress. BdPP2CA6-overexpression plants exhibited lower contents of IL, MDA, and H₂O₂, and higher levels of chlorophyll and proline compared with WT under NaCl treatment (Figures 7D–H). The results suggested that transgenic seedlings suffered less membrane damage and intracellular oxidative stress, therefore were more tolerant to external salinity stress.

To further understand the molecular mechanisms of BdPP2CA6 function in abiotic stress tolerance, interactions between BdPP2CA6 and AtSnRK2/BdSnRK2 were investigated. SnRK2.2/2.3/2.6 were selected as potential targets because SnRK2.2/2.3/2.6 are essential in ABA signaling and interact with several clade A PP2Cs in Arabidopsis (Fujita et al., 2009; Nakashima et al., 2009). We chose SD/-Trp-Leu-His/0.1 μM 3-AT to assess growth status of yeast cells co-transformed with BdPP2CA6 and BdSnRKs because they were not able to grow on SD/-Trp-Leu-His-Ade. The results showed that BdPP2CA6 had a mild interaction with BdSnRK2.2 but had a strong interaction with AtSnRK2.3 (Figures 8A,B). This strong interaction between BdPP2CA6 and AtSnRK2.3 was further confirmed by BiFC assay (Figure 8C).

To identify molecular events involved in BdPP2CA6-mediated signal pathway, expressions of several ABA- and stress-responsive genes were investigated by qRT-PCR technique. The results showed that certain genes (ABF2, ABF3, DREB2A, MYB15, RD26, GORK, and SLAH3) were significantly induced by salinity stress and were induced to higher expression levels in transgenic plants than WT while expression levels of MYB44, RD29B, and ABF4 were not significantly different between transgenic plants and WT under salinity stress (Figure 9). These results imply that BdPP2CA6 positively regulate the response to salinity stress by modulating expression of diverse stress-related genes in Arabidopsis plants.