Seeds of pepper (C. annuum), Arabidopsis thaliana (ecotype Col-0), and tobacco (Nicotiana benthamiana) were sown in a soil mix, including sand, loam soil, vermiculite, perlite, peat moss (10:10:5:3:2, v:v:v:v:v). These plants were grown in a fixed condition at 24 ± 1°C (16 h/8 h, light/dark).

The coding region of the CaAIEF1 gene was cloned into the pGBKT7 vector, which includes a nuclear localization and GAL4 DNA-binding domains. We used Saccharomyces cerevisiae strain AH109 for CaAIEF1 transactivation analysis, which have two reporter genes (Ade1 and His3) with GAL4 promoters. The AH109 cells were transformed with pGBKT7 vector that carried GAL4-CaAIEF1 fusion genes. The transformed yeast cells were selected into SD-adenine-histidine-leucine-tryptophan medium to confirm transcriptional activation.

For the loss-of function analysis of CaAIEF1, virus-induced gene silencing (VIGS) was conducted in pepper plants, as described by Lim et al. (2015b). Briefly, pTRV1 and pTRV2:CaAIEF1 or pTRV2:00 transformed to Agrobacterium tumefaciens strain GV3101. The Agrobacterium cells were infiltrated by syringe to the pepper cotyledons (OD₆₀₀ = 0.2 for each construct). Plants were grown in pepper growth condition described previously for spreading the virus.

For the elucidation of biological function of CaAIEF1, CaAIEF1 overexpressing (OX) Arabidopsis was generated. The full-length coding region of CaAIEF1 (accession no. KY652734) was inserted into pENTR/D-TOPO vectors (Invitrogen, Carlsbad, CA, United States). Through the LR reaction, the inserted genes were introduced into pK2GW7 to constitutively express each gene with the cauliflower mosaic virus (CaMV) 35S promoter (Karimi et al., 2002); the generated construct was transformed into Agrobacterium tumefaciens strain GV3101. The floral dip method was applied for transformation of Arabidopsis with the CaAIEF1 gene (Clough and Bent, 1998). Overexpressing plants were selected by germinating putative transformed seeds on Murashige and Skoog (MS) plates supplemented with 50 μg⋅mL^-1 of kanamycin as selection marker. Seeds of T3 plants were harvested from second-generation transgenic plants showing a 3:1 segregation ratio on MS plates supplemented with the same antibiotic.

To analyze the expression patterns of CaAIEF1 in pepper plants, which had been treated with ABA, NaCl, and dehydration, leaf samples were prepared as described by Lim et al. (2015b). For the ABA and NaCl treatment, the six-leaf-stage pepper plants were treated with 100 μM ABA, irrigated with a salt solution (200 mM). Whole plants were dried on 3-mm paper (Whatman, Clifton, United Kingdom), or the roots were removed and the aerial parts of plants were dried. After treatment, the third and fourth leaves were harvested at the several time points.

For the dehydration phenotype analysis, 5-week-old pepper plants and 3-week-old Arabidopsis plants were randomly placed and subjected to dehydration treatment by withholding watering for 12 and 11 days, respectively. After re-watering, the survival rate was calculated by counting the plant number with resumed growth. To determine the dehydration tolerance in a quantitative manner, rates of water loss were measured by drying leaves detached from gene-silenced pepper plants and Arabidopsis transgenic plants.

For quantitative real-time transcription-polymerase chain reaction (qPCR) analysis, 4-week-old wild-type and CaAIEF1-OX Arabidopsis were treated with dehydration stress, and harvested at the several time points.

The measurements of stomatal pore size and leaf temperature were performed as described previously (Lim and Lee, 2016). Briefly, pepper and Arabidopsis leaf peels were floated in stomatal opening buffer (SOS: 50 mM KCl and 10 mM MES-KOH, pH 6.15, 10 mM CaCl₂) in the light condition for 2.5 h. Stomatal closure was induced by replacing the buffer with fresh SOS containing various concentrations of ABA. After an additional 2.5 h of incubation in each buffer, 100 stomata per each sample were observed under a Nikon Eclipse 80i microscope. An infrared camera (FLIR systems; T420) were used for obtaining the thermal images, and FLIR Tools+ ver 5.2 software measured the leaf temperatures.

RNA samples were digested with RNA-free DNase to inhibit contamination of genomic DNA. The cDNA were synthesized using a Transcript First Strand cDNA Synthesis kit (Roche, Indianapolis, IN, United States). For qPCR analysis, the specific primers and CFX96 Touch^TM Real-Time PCR detection system (Bio-Rad, Hercules, CA, United States) were used (Supplementary Table S1). The PCR was programmed as follows: 95°C for 5 min; 45 cycles each at 95°C for 20 s and 60°C for 20 s; and 72°C for 20 s. To determine relative expression level, we used the ΔΔCt method (Livak and Schmittgen, 2001). For the normalization, the Arabidopsis AtACT8 and pepper CaACT1 genes were used.

CaAIEF1-GFP proteins were expressed in the leaves of N. benthamiana epidermis cells by using infiltration of Agrobacterium tumefaciens strain GV3101 with strain carrying p19 suppressor (1:1 ratio; OD₆₀₀ = 0.5). The GFP signal was observed 2 days after infiltration, using a confocal microscope (510 UV/Vis Meta; Zeiss, Oberkochen, Germany).

To identify the novel ABA-induced transcription factor, we conducted an RNA-seq assay using control and ABA treated pepper leaves; we identified the putative pepper ABA-induced ERF. We designated gene name CaAIEF1 (C. annuum ABA Induced ERF 1) by domain analysis (Figure 1A and Supplementary Figure S1). CaAIEF1 encoded 180 amino acids. The predicted protein consisted of an AP2/ERF domain with 65 amino acids (59–123) in the central region and an acidic domain (AD) consisting of 21 amino acids (160–180); these domains specifically bind to promoter regions and activate target genes (Figure 1A) (Licausi et al., 2013). Moreover, a putative nuclear localization signal (NLS) was detected in the AP2/ERF domain. Amino acid sequence alignments of CaAIEF1 with ERF proteins revealed that these ERF protein members have a highly conserved AP2/ERF domain. CaAIEF1 shares identity with other ERF proteins of Solanum tuberosum (accession no. XP_006367196.1, 87.9% identity), N. tomentosiformis (accession no. XP_009619682.1, 81.3% identity), N. tabacum (accession no. XP_016487471.1, 80.7% identity), Solanum lycopersicum (accession no. XP_001233987.1, 79.1% identity) and A. thaliana (accession no. NP_188965.1, 32.9% identity). However, CaAIEF1 does not share any identity with pepper ERF proteins (Supplementary Figure S2).

Transcription factors have an activation domain, which functions in activation of the target gene. The activation domain is enriched in glutamine, proline, or acidic amino acids (Schwechheimer et al., 1998a,b; Jakoby et al., 2002; Cantin et al., 2003). Transactivation of acidic domains has universal functions in eukaryotic transcription factors (Hahn, 1993; Schwechheimer et al., 1998a). CaAIEF1 contains an acidic domain in the C-terminal region; hence, we used deletion constructs of the CaAIEF1 gene to determine whether the CaAIEF1 protein functions as a transactivation factor in yeast. The deletion constructs were inserted into the pGBKT7 vector carrying the GAL4 DNA-binding domain, expressed in S. cerevisiae strain AH109 (Figure 1B). Yeast cells carrying the C-terminal region (containing the acidic domain) grew well in the selection medium, suggesting activation of the reporter genes. Our results indicate that the C-terminal region of CaAIEF1 act as a potential transcriptional activator.

To examine whether CaAIEF1 is expressed in response to several stresses, including drought, ABA, and NaCl, we performed quantitative RT-PCR analysis (Figure 2A). In drought-treated pepper leaves, the expression level of CaAIEF1 was up-regulated at 2 h and thereafter gradually decreased. ABA is a key abiotic stress-related hormone involved in stress signal transduction—especially under drought stress conditions (Lee and Luan, 2012). In ABA-treated pepper leaves, CaAIEF1 was induced within 6–24 h. CaAIEF1 transcripts started to induce at 2 h and reached peak levels after 12 h by NaCl treatment. The CaAIEF1 is ethylene responsive factor; hence, we checked whether CaAIEF1 is induced by ethylene. As shown in Supplementary Figure S3A, the transcription level of CaAIEF1 was up-regulated by ethylene. Moreover, the CaAIEF1 expression was also regulated by developmental stages (Supplementary Figure S3B).

CaAIEF1 has an NLS in the AP2/ERF domain of the central region (amino acids 202–219); hence, we predicted that CaAIEF1 localizes in the nucleus. To confirm the subcellular localization of the CaAIEF1 protein (Figure 2B), we used the CaAIEF1 coding region with green fluorescent protein (GFP) cDNA. GFP fluorescence signals were detected in the nucleus of N. benthamiana cells, indicating that CaAIEF1 functions in the nucleus.

The expression level of CaAIEF1 was induced by ABA and abiotic stresses; hence, we investigated the in vivo function of CaAIEF1 using virus-induced gene silencing (VIGS) in pepper plants and Arabidopsis transgenic plants. Semi-quantitative RT-PCR analysis revealed that CaAIEF1 transcripts were not expressed in CaAIEF1-silenced pepper plants in normal and drought stress conditions (Figure 3A). We investigated the biological function of CaAIEF1 in drought stress condition (Figure 3B). Under well-grown conditions, we did not observe any phenotypic differences in both plants (Figure 3B, upper panel). However, when we subjected pepper plants to dehydration for 12 days and rehydration for 3 days, CaAIEF1-silenced pepper plants showed a more wilted phenotype than control plants (Figure 3B, middle and lower panels). In addition, after rehydration, the 80% of control and only 10.0% of CaAIEF1-silenced pepper plants resumed their growth. To examine whether the drought-sensitive phenotype was cause by rapid transpirational water loss, the fresh weight of rosette leaves were measured during 8 h after detachment (Figure 3C). The higher transpirational water loss was detected in CaAIEF1-silenced plants than in control plants. Under dehydrated conditions, ABA is synthesized in several tissues and accumulates in the leaves; this leads to stomatal closure via regulation of guard cell turgor pressure (Lee and Luan, 2012). To examine the biological role of CaAIEF1 in drought stress-induced ABA signaling, we measured the stomatal apertures and leaf temperatures of control and CaAIEF1-silenced pepper plants with or without ABA (Figures 3D–G). In the absence of ABA, we did not observe differences between control and CaAIEF1-silenced pepper plants. However, after exposure to 10 and 20 μM ABA, the stomatal pore sizes were large in CaAIEF1-silenced pepper plants compared with control plants (Figures 3D,E). Moreover, the leaf temperatures were significantly lower in CaAIEF1-silenced plants than in control pepper plants, implying that increased evaporative cooling was caused by decreased stomatal closure (Figures 3F,G). Our data indicated that the decreased ABA sensitivity of CaAIEF1-silenced pepper plants leads to increased water loss and a drought-sensitive phenotype.

To further examine the biological function of CaAIEF1, we generated CaAIEF1-OX Arabidopsis plants. Semi-quantitative RT-PCR analysis revealed that CaAIEF1 was only expressed in CaAIEF1-OX plants (Supplementary Figure S4). CaAIEF1 is induced by ABA; hence, we predicted that enhanced expression of CaAIEF1 modulates ABA signaling. To examine whether CaAIEF1 influences the ABA-mediated signaling, we analyzed the ABA sensitivity during the germinative- and post-germinative stages, including seed germination, seedling establishment, and root growth (Figure 4). We did not observe marked differences in germination rates between both plants on growth media without ABA. In the presence of 1.5 μM ABA, CaAIEF1-OX seeds germinated slower than wild-type seeds (Figure 4A). Moreover, CaAIEF1-OX plants were hypersensitive to ABA at seedling stages in seedling establishment (Figures 4B–D) root growth (Figures 4E,F). To determine whether ABA hypersensitivity in root growth and seedling establishment in CaAIEF1-OX plants is caused by late germination, the seedlings at 3 days after germination in the absence of ABA were transferred to medium including ABA. As shown in Supplementary Figure S5, the CaAIEF1-OX plants also exhibited ABA sensitive phenotype. The ABA-hypersensitive phenotype displayed by CaAIEF1-OX plants indicates that CaAIEF1 modulates the ABA-mediated response.

CaAIEF1-silenced pepper plants showed a drought-sensitive phenotype; hence, we examined whether CaAIEF1-OX plants exhibit drought tolerant phenotype (Figure 5). Under well-grown conditions, we did not observe phenotypic differences in both plants (Figure 5A, left panel). However, when we treated drought stress by dehydration for 11 days and rehydration for 2 days, CaAIEF1-OX plants showed a less wilted phenotype than wild-type plants (Figure 5A, middle and right panels). In addition, the survival rate of CaAIEF1-OX plants (76.6–93.8%) was higher than that of wild-type plants (10.7%). To examine whether different transpirational water loss leads to altered drought tolerance, the leaf fresh weight was measured (Figure 5B). In accordance with the drought tolerance, the transpirational water loss was lower in CaAIEF1-OX plants than in wild-type plants. To examine the biological role of CaAIEF1 in drought stress-induced ABA signaling, we measured the stomatal apertures and leaf temperatures of wild-type and CaAIEF1-OX plants with or without ABA (Figures 5C–E). We did not observe any differences between both plants. ABA treatment led to an increase in stomatal closure in wild-type and transgenic plants. However, after exposure to 10 and 20 μM ABA, the stomatal pore sizes were small in CaAIEF1-OX plants compared with wild-type plants (Figures 5C,D). We measured leaf temperatures, as indicator of stomatal opening/closure via evaporative cooling. The leaf temperatures were significantly higher in CaAIEF1-OX plants than in wild-type plants (Figure 5E). Our results imply that the ABA hypersensitivity displayed by CaAIEF1-OX plants leads to decreased water loss and contributes to a drought-tolerant phenotype.

Several previous studies suggested that the expression of stress-responsive genes influence stress tolerance or stress sensitivity; hence, we examined whether stress-responsive genes in pepper and Arabidopsis are directly or indirectly regulated by the expression level of CaAIEF1 (Figure 6). We subjected plants to dehydration stress by removing the roots from the soil. Contrary to our expectations, q-PCR assay revealed that the induction of stress-responsive genes was higher in CaAIEF1-silenced pepper leaves than in control pepper leaves (Figure 6A). However, the induction of several stress-responsive genes, including NCED3, RD29A, RD29B, COR15A, RAB18, and KIN1, was lower in CaAIEF1-OX leaves than in wild-type leaves (Figure 6B). Our data imply that CaAIEF1 expression negatively affected the expression of stress-responsive genes.