Grapevine (V. vinifera L. cv. ‘Summer Black’ and ‘Venus’) was used in this study. Plantlets were grown on solid Murashige and Skoog (MS) medium under a 16-h light/8-h dark cycle and 70% relative humidity at 25°C in the greenhouse.

All transgenic lines of VvANN1 were developed in the A. thaliana Columbia (Col-0) background. Col-0 plants were used as the wild type in the present study. Arabidopsis thaliana seedings were germinated on MS for 6 days and transferred to soil pots (7cm×7 cm). The seedlings were grown in a light incubator (22°C, 16-h day/8-h night cycle and 70% relative humidity).

Five-week-old plantlets were transferred to liquid MS medium for 2 days and then planted in a fresh liquid MS medium with 10% (w/v) PEG6000 (NO. A610432, Sangon Biotech, Shanghai, China) to evaluate osmotic stress tolerance. Stem apex samples were collected for quantitative real-time polymerase chain reaction (RT-qPCR) analysis.

Six-day-old VvANN1 transgenic and Col-0 seedlings were germinated on MS medium, transferred to soil pots (7 cm×7 cm) for 9 days with regular water. These were dried for 7 days and then allowed a 3-day recovery. The survival rates were recorded and the seedings were photographed.

The total cDNA of VvANN1 and a 1,538-bp fragment of the VvANN1 promoter were cloned from five-week-old grapevine ‘Summer Black’ plantlets. The VvANN1 coding region sequence (cDNA) was digested and ligated with pCAMBIA1301-HA, modified pMDC83, and pET32a vectors to obtain Ubi::VvANN1-HA, 35S::VvANN1-GFP, and VvANN1-His constructs, respectively.

The promoter region (1,538 bp) upstream of the VvANN1 start codon was amplified and cloned into the vector to produce transgenic VvANN1Pro::GUS lines. Genomic fragments (285 bp) upstream of VvANN1 were amplified and cloned into pAbAi and pGreenII 0800-LUC vectors to generate pAbAi-VvANN1Pro and pGreenII 0800-LUC-VvANN1Pro, respectively.

VvbZIP45 cDNA was amplified and cloned into pCAMBIA1301-HA, pGreenII 62-SK, pGADT7 and pMDC83 vectors to generate Ubi::VvbZIP45-HA, pGreenII 62-SK-VvbZIP45, AD-VvbZIP45 and 35S::VvbZIP45-GFP, respectively.

Related constructs were introduced into Agrobacterium tumefaciens cells (GV3101) and grown at 28°C for 3 days before being transformed into Col-0. The primers used to produce the constructs are listed in Supplemental Table S1 .

Total RNA was extracted from grapevine stem apices and Arabidopsis leaves using TRIzol reagent (NO. B511311, Sangon Biotech, Shanghai, China). RNase-free DNase I (EN0521, Thermo Fisher Scientific, Waltham, MA, USA) was used to remove genomic DNA. RT-qPCR was performed using the ChamQ Universal SYBR^® qPCR Master Mix Kit (Q711-02, Vazyme, Nanjing, China) using the QuantStudio Q5 thermal cycler (Thermo Fisher Scientific, Waltham, MA, USA). The experiment was performed using three biological replicates. Relative quantitative results were calculated by normalization to AtACTIN2 and VvACTIN7, the internal controls in Arabidopsis and grapevine, respectively. All primer sequences are listed in Supplemental Table S1 .

GUS activity was detected via histochemical staining of tissues as previously described (Zhang et al., 2018a) but with slight modifications. All transgenic Arabidopsis tissues were incubated in GUS staining solution at 37°C with the corresponding time under dark conditions. All stained samples were washed with 70% ethanol to remove the residual dye and chlorophyll. Images were captured using a DVM6a 3D microscope (Leica, Wetzlar, Germany).

Transient expression assays were performed in grape protoplasts to determine the subcellular localization of VvANN1. Grape protoplasts were prepared, and transient expression assays were performed as described previously (Lee and Wetzstein, 1988; Wang et al., 2015). The recombinant plasmid 35S::VvANN1-GFP was introduced into grape protoplasts, followed by incubation at 28°C for 12 h. Fluorescence signals were observed using a Zeiss LSM710 laser scanning confocal microscope (Zeiss, Jena, Germany).

The Ca²⁺-binding assay was conducted by detecting fluorescence measurements of VvANN1 according to a method described previously (Qiao et al., 2015). The recombinant plasmid VvANN1-His was introduced into Escherichia coli (E. coli) strain BL21. Total VvANN1-His protein was induced by isopropyl β-D-thiogalactoside (IPTG; NO. A100487, Sangon Biotech, Shanghai, China) at 28°C for 12 h. E. coli cells were lysed via ultra-sonication, and the obtained samples were ultra-centrifuged at 12,000 ×g at 4°C for 15 min. The supernatant was collected and purified via affinity chromatography on Ni-agarose columns (Cat. No. 30210, Qiagen, Duesseldorf Germany). The assay media contained 2 μM VvANN1-His protein and 0 mM or 2 mM Ca²⁺. Fluorescence spectroscopy was carried out using a fluorescence spectrophotometer (F-4600; Hitachi, Tokyo, Japan).

The pAbAi-VvANN1Pro (containing 285 bp partial promoter sequences with the ABRE element) vector was transformed into the yeast strain Y1HGold as bait. Positive yeast strains were diluted and spread onto selection medium (SD; Code No:630411, Clontech, Mountain View, CA, USA) lacking Ura containing various concentrations of Aureobasidin A (AbA; Code No.630466, Clontech, Mountain View, CA, USA) to screen for an appropriate concentration to eliminate self‐activation. AD-VvbZIP45 was transformed into Y1HGold with pAbAi-VvANN1Pro, and the pGADT7 plasmid was used as a negative control. Different experimental groups were cultured on SD/-Leu medium with or without 80 ng/mL AbA at 30°C for 3 days.

A dual-luciferase reporter assay was conducted to test the transcriptional repression activity of VvbZIP45 in tobacco (Nicotiana tabacum) leaves. The pGreenII 62-SK and pGreenII 62-SK-VvbZIP45 were used as effector plasmids, and pGreenII 0800-LUC-VvANN1Pro was used as the reporter plasmid. The plasmids were mixed and expressed in tobacco leaves via A. tumefaciens GV3101 strain injection. After 2 days, total protein was extracted from infiltrated tobacco leaves, and the LUC/REN activity ratio was measured using the dual luciferase reporter assay system (E1960, Promega, Madison, WI, USA).

Approximately 1 g of plant tissue was harvested from two-week-old transgenic hybrid progeny seedlings that were germinated on MS plates in a light incubator. Samples were prepared according to previous reports (Yamaguchi et al., 2014; Zhao et al., 2022). ChIP experiments were performed using Abcam ChIP Kit - Plants (ab117137, Abcam, Cambridge, MA, USA) according to the manufacturer’s instructions. Chromatin was immunoprecipitated using anti-HA (ab9110, Abcam, Cambridge, MA, USA). Following immunoprecipitation, samples were analyzed by RT-qPCR. The specific primers used are listed in Supplemental Table S1 .

Leaves were collected from VvANN1 transgenic Arabidopsis plants and Col-0 plants grown under normal conditions for 15 days, then grown with or without water for 5 days. H₂O₂ and O₂ ^·- levels were examined via histochemical staining with 3, 3′‐diaminobenzidine (DAB; CAS No: 91-95-2, Sigma-Aldrich, St. Louis, MO, USA) or nitro blue tetrazolium (NBT; CAS No: 298-83-9, Sigma-Aldrich, St. Louis, MO, USA), as previously described (Zhang et al., 2021). All measurements were determined using three independent biological replicates.

Leaves were collected from VvANN1 transgenic Arabidopsis plants and Col-0 plants cultivated under normal conditions for 15 days (control), then without water for 5 days (drought-exposed).

Samples (each weighing 0.2 g) were homogenized in 1 mL of sodium phosphate buffer (50 mM phosphate, 1 mM EDTA-Na₂, 1% (w/v) polyvinyl pyrrolidone; pH 7.4). Centrifugation was performed at 10,000 ×g at 4°C for 20 min, and the supernatant was used to detect antioxidant enzyme activity. The activities of SOD, POD, and CAT were determined according to methods described previously (Zhang et al., 2017).

Samples (each weighing 0.2 g) were homogenized in 2 mL of 10% thiobarbituric acid (TBA). Following centrifugation at 12,000 ×g at 4°C for 15 min, the MDA contents were detected according to the method described by Zhou et al. (2018).

There are 14 predicted annexin genes in grapevine (Jami et al., 2012), however, the function of ANNEXIN family in grapevine remains unidentified. We isolated a putative grapevine annexin gene from Vitis vinifera via RT-PCR and named it VvANN1. Based on the results of bioinformatics analyses, the genomic sequence of VvANN1 is composed of five exons and four introns and produces a 930-bp coding sequence transcript encoding a protein of 309 amino acids with a molecular weight of 35 kDa ( Figure 1A ). This protein contains four annexin domain architectures ( Figure 1B ) with a type-II Ca²⁺ binding site in the first annexin conserved repeat domain.

A phylogenetic tree generated using the ANNEXINs of A. thaliana and V. vinifera suggests that VvANN1 has the maximum homology with AtANN1 of A. thaliana, so the gene was designated as VvANN1. This finding implies that VvANN1 proteins may regulate stress response processes similar to its homologous AtANN1 ( Figure 1C ).

Transcript levels in different organs of the grapevine ‘Venus’ were performed via RT-qPCR to analyze the tissue-specific expression of VvANN1. The expression of VvANN1 was higher in the flower, stem, and tendril than in the root, fruit, and leaf ( Figure 2A ). Furthermore, we also generated Arabidopsis transgenic lines (#1, #4, #7 and #9) containing the VvANN1Pro::GUS, and detected GUS activities in seedlings, flowers, which was consistent with the findings shown by RT-qPCR ( Figure 2B ). Moreover, important discrepancies were noted when comparing the GUS staining observed in different VvANN1Pro::GUS transgenic lines, such as #1 and #9. This result suggested the cassette VvANN1::GUS might be inserted various sites in different Arabidopsis transgenic lines and resulted in different GUS expression level based on the position effect, for example, there might be diverse enhancers near to the insertion sites.

We transformed the 35S::VvANN1-GFP vector transiently into grapevine protoplasts to investigate the subcellular localization of VvANN1. VvANN1-GFP signal could be detected in cytoplasm, whereas GFP signals were ubiquitously distributed in the grapevine protoplast ( Supplemental Figure S1 ).

Furthermore, we carried out a VvANN1-His recombinant protein fluorescence experiment to verify the Ca²⁺ binding ability of VvANN1. First, we transformed the VvANN1-His plasmid into E. coli strain BL21. The VvANN1-His recombinant protein was successfully expressed in BL21 by adding IPTG, purified via Ni-NTA affinity chromatography, and further detected via SDS-PAGE ( Supplemental Figure S2 ). Next, a UV spectrophotometer was used to measure the fluorescence intensity of the VvANN1-His recombinant protein when incubated with or without Ca²⁺. The highest fluorescence intensity was approximately 300 A.U. at 390 nm. By contrast, fluorescence intensity was reduced to 150 A.U. at 390 nm in the presence of Ca²⁺ ( Supplemental Figure S3 ), indicating that VvANN1 might have Ca²⁺-binding capacity.

AtANN1 is up-regulated in a Ca²⁺-dependent manner to regulate drought stress responses synergistically. Here, we investigated whether VvANN1 was responsive to drought stress. We performed RT-qPCR to detect the VvANN1 expression pattern in plantlets (Vitis spp. cv ‘Summer Black’ and ‘Venus’) treated with or without 10% PEG6000. The results showed that the expression of VvANN1 in ‘Summer Black’ and ‘Venus’ was significantly induced by PEG treatment up to 11-fold and 4-fold, respectively, at 24 h ( Supplemental Figure S4A ).

GUS staining was performed to determine further the expression of VvANN1 in six-day-old transgenic VvANN1Pro::GUS plants under PEG treatment. Histochemical staining revealed that, VvANN1Pro::GUS signals were mainly expressed in the vascular tissues of roots under normal conditions. This expression pattern of GUS activity was increased following treatment with 10% PEG6000 for 12 h ( Supplemental Figure S4B ), suggesting that VvANN1 expression may be induced by PEG treatment.

To further determine the role of VvANN1 in modulating plant osmotic stress, we further generated 3 independent transgenic Arabidopsis lines driven by the CaMV35S promoter (L2, L3 and L4) and 3 independent transgenic Arabidopsis lines driven by the Ubi promoter (L5, L6 and L7). The transcript level of VvANN1 was analyzed with RT-qPCR and three homozygous transgenic Arabidopsis lines (L2, L3 and L6) were used in subsequent experiments ( Supplemental Figure S5 ). VvANN1 transgenic and Col-0 seedlings were cultivated on MS medium with 0 mM, 250 mM or 300 mM mannitol. The germination rates showed no apparent difference between VvANN1 transgenic and Col-0 seedlings in MS medium after 84 h. Conversely, all VvANN1 transgenic seedlings showed higher germinating rates than Col-0 in the presence of mannitol ( Figures 3A, B ), suggesting that, during the germination stage, VvANN1 transgenic seedlings were less sensitive to osmotic stress than Col-0 seedlings. Based on these results, we concluded that VvANN1 might positively regulate osmotic stress in Arabidopsis.

Fifteen-day-old transgenic Arabidopsis and Col-0 plants were subjected to a drought treatment to evaluate the function of VvANN1 in drought tolerance. Once watering was stopped for 7 days, VvANN1 transgenic plants exhibited less wilting than Col-0 plants ( Figure 4A ). After rewatering, VvANN1 transgenic plants had higher survival rates (58%, 54% and 60% for L2, L3 and L6, respectively) than Col-0 plants (36%) ( Figure 4B ), indicating that VvANN1 transgenic plants significantly improved drought tolerance in Arabidopsis.

We analyzed the VvANN1 promoter using the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare) to gain further insights into the regulatory mechanism of VvANN1 and found several stress-responsive cis-elements (e.g., MYC, MYB and ABRE). Among these are two ABRE cis-elements in the 0- to 300-bp region of the VvANN1 promoter. Studies have shown that VvbZIP45 can bind to the ABRE element in the promoter region of the target gene and enhance drought stress tolerance in Arabidopsis (Liu et al., 2014; Nicolas et al., 2014; Liu et al., 2019). Thus, we hypothesized that VvbZIP45 could bind to the promoter region of VvANN1 and regulate its expression.

We performed a Y1H assay, where the VvANN1 promoter sequence (from -1 to -285 bp) containing two ABRE motifs was first constructed into the pAbAi vector (pAbAi-VvANN1Pro). Next, this bait plasmid was transformed into yeast strain Y1HGold. Finally, the pGADT7-VvbZIP45 and pGADT7 vectors were transformed into the yeast Y1HGold carrying the bait plasmid. The results showed that Y1HGold carrying pAbAi-VvANN1Pro and transformed with the pGADT7-VvbZIP45 plasmid grew on SD/-Leu medium containing 80 ng/mL AbA, whereas yeast cells carrying the pGADT7 vector did not ( Figure 5A ). This result showed that VvbZIP45 could bind to the promoter region of VvANN1.

We performed a ChIP-qPCR assay to investigate whether VvbZIP45 could bind directly to the ABRE cis-element in the promoter region of VvANN1. Ubi::VvbZIP45-HA transgenic Arabidopsis plants were crossed with VvANN1Pro::GUS transgenic plant, and hybrid transgenic A. thaliana seedlings were verified via PCR. Four fragments spanning different regions of the VvANN1 promoter with or without the ABRE motif were selected for qPCR analysis ( Figure 5B ). Figure 5C shows that 1, 3 and 4 fragments of the VvANN1 promoter were markedly enriched in ChIP-qPCR assay with anti-HA compared with anti-IgG.

We used a luciferase reporting system to determine whether VvbZIP45 could activate VvANN1 expression in vivo ( Figures 5D, E ). A pGreenII 0800 vector harboring a dual-luciferase reporter gene driven by the VvANN1 promoter was co-transformed into tobacco leaves with pGreenII 62-SK or pGreenII 62-SK-VvbZIP45. The results showed that tobacco plants expressing pGreenII 62-SK-VvbZIP45 exhibited significantly higher LUC/REN activity than control plants.

To further assure the regulation of VvbZIP45 in VvANN1 expression in vivo, we also generated transgenic Arabidopsis that constitutively expressed the VvbZIP45 gene (35S::VvbZIP45) and further produced VvANN1Pro::GUS/35S::VvbZIP45 Arabidopsis plants via crossing. The lower GUS expression level VvANN1Pro::GUS lines #1 and #4 (as shown in Figure 2C ) were used as the male parents, 35S::VvbZIP45 Arabidopsis as female parent. The results of histochemical GUS assays showed that staining intensity was significantly higher in hybrid plants (VvANN1Pro::GUS/VvbZIP45) than in VvANN1Pro::GUS transgenic plants ( Figure 5F ). Interestingly, 10% PEG6000 treatment could be resulted in much more strong GUS staining in VvANN1Pro::GUS/35S::VvbZIP45 hybrid Arabidopsis plants comparing to VvANN1Pro::GUS transgenic plants itself ( Figure 6 ). These results indicated that VvbZIP45 could act as a transcription factor to promote VvANN1 expression under normal or drought stress conditions.

Further examination of the expression level of VvbZIP45 in five-week-old grapevine plantlets under osmotic stress showed that VvbZIP45 expression was rapidly activated by PEG treatment and reached the highest level at 12 h ( Supplemental Figure S6 ). This result suggests that VvbZIP45 could be up-regulated under drought stress.

In plants, ROS serve as signal molecules at low levels but can cause cell damage at extreme doses. Drought stress could lead to ROS accumulation. Here, H₂O₂ accumulation was detected via DAB staining. The results showed no difference in DAB staining of VvANN1 transgenic plants and Col-0 plants under normal conditions. However, after 5 days of drought treatment, weaker staining was observed in VvANN1 transgenic leaves than in Col-0 leaves ( Figure 7A ). The accumulation of O₂ ^.- detected via NBT staining showed similar results under drought stress treatment ( Figure 7B ). These results indicated that less H₂O₂ and O₂ ^.- were produced in VvANN1 transgenic plants than in Col-0 plants under drought stress.

Previous studies have shown that ROS accumulation affects intracellular environment stability and severely damages the plant cell membrane (Zhang et al., 2018b). The level of MDA is an important indicators of cell membrane damage (Gao et al., 2020). In the present study, the MDA levels in VvANN1 transgenic plants were similar to Col-0 plants under normal conditions. Following drought treatment for 5 days, three VvANN1 transgenic plants had significantly lower MDA content (3.42, 3.80 and 3.51 μmol/g) than Col-0 (4.54 μmol/g) ( Figure 7C ). These results suggest that VvANN1 plays a vital role in the reduced accumulation and damage of ROS in cells under drought stress.

The activities of ROS scavenging enzymes (SOD, POD and CAT) were examined to analyze the mechanism of VvANN1 function in regulating the level of ROS ( Figures 7D–F ). The activities of SOD, POD, and CAT were not significantly different in both VvANN1 transgenic plants and Col-0 plants under normal conditions. However, after drought stress treatment, the enzymatic activities of SOD, POD, and CAT were significantly higher in VvANN1 transgenic plants than in Col-0 plants. These results suggested that the expression of VvANN1 driven by CaMV35S or Ubi promoter in Arabidopsis could improve the activity of SOD, POD and CAT to eliminate excess ROS, subsequently assisting in the maintenance of plasma membrane integrity and improving drought stress tolerance.