The table-grape cultivar Crimson seedless (C133–199 × Emperor) was used in this study. The sterilized seedlings were grown on 1/2 MS solid medium (Murashige and Skoog, 1962) in glass bottles (9 cm height × 6 cm diameter) at a 25/20°C growth room with white light illumination (120 μmol photons per m² s¹) under a 16/8 h light/dark photocycle. Shoot tips with a minimum of one bud were used to subculture every month. After 30 days, uniform plants were selected as the test material and were transferred into Hoagland solution (2 mM MgSO₄, 5 mM Ca(NO₃)₂, 5 mM KNO₃, 1 mM KH₂PO₄, and microelements contained 0.2 mM Fe-EDTA, 2 × 10⁻³ mM MnCl₂, 2.5 × 10⁻⁴ mM H₃BO₃, 5 × 10⁻⁴ mM CuSO₄, 2 × 10⁻³ mM ZnSO₄ and 5 × 10⁻⁴ mM H₂MoO₄) for 7 days. The growth conditions were kept the same as above. The test seedlings were treated with 50 mM NaCl in Hoagland solution for 3, 6, 12, 24, 48, and 72 h, respectively. And the seedlings grown in Hoagland alone were used as the control (0 h). Samples were then harvested at the same time in the same day to minimize diurnal effects wherever possible, then frozen in liquid nitrogen and stored at −80°C.

Plasma membrane-enriched vesicles were isolated by discontinuous sucrose gradient centrifugation, according to the method of Wang et al. (2001) and Mandala and Taiz (1985) with modifications. Roots were washed with cold deionized water and homogenized in extraction medium containing 50 mM Tricine-Tris (pH 7.8), 3 mM EGTA, 3 mM MgSO₄, 0.5% (V/V) PVP, 2 mM DTT, 0.2 mM PMSF, 5% (V/V) glycerol, and mannitol, which produced the same osmotic potential as within leaves. Two milliliters of the medium was used for each 2 g of fresh material. The homogenate was filtered through four layers of cheesecloth and centrifuged at 10,000 g for 20 min (Beckman JA-20). The supernatant was loaded on a 0%: 36%: 45% (W/W) sucrose gradient solution (5 mM HEPES adjusted to pH 7.5 with Tris, and 1 mM DTT) and centrifuged at 100,000 g for 2 h (Beckman SW40Ti). The plasma membrane-enriched vesicles were located at the 36%: 45% (W/W) sucrose interface. These vesicles were carefully collected and diluted 3–4-fold with dilution buffer (3 mM MgSO₄, 10 mM HEPES–Tris pH 7.5, 1 mM DTT) and centrifuged at 100,000 g for 30 min (Beckman type 65). The pellets were suspended in a storage buffer (40% (V/V) glycerol, 2 mM DTT, 10 mM HEPES, adjusted to pH 7.0 with KOH).

Protein content was determined using bovine serum album as a protein standard according to the Bradford method (Bradford, 1976).

The highly conserved domain “GDGVNDAPALK” was defined as the probe sequence according to the alignment results of published PM H⁺-ATPase amino acid sequences in plants (Palmgren, 2001). The non-redundant protein database of V. vinifera was retrieved through blastp on the NCBI website (National Center for Biotechnology Information, www.ncbi.nlm.nih.gov). Then the PM H⁺-ATPase functional domains of these candidate proteins were predicted using Pfam software (http://pfam.sanger.ac.uk; Finn et al., 2010). The corresponding gene sequences and cDNA sequences of these domain-verified PM H⁺-ATPases were acquired using accession numbers.

The phylogenetic tree was constructed using MEGA 5.0 software and making comparisons between candidate VvPMAs and previously identified AHA proteins from A. thaliana.

Total grape RNA was extracted using the CTAB method as previously described (Gambino et al., 2008), with a few modifications. Root samples (200 mg) were ground with mortar and pestle under liquid nitrogen. The powder was transferred to an RNase-free tube containing 2 mL CTAB extracting buffer (2% CTAB, 2.5% PVP-30, 2 M NaCl, 100 mM Tris-HCl (pH 8.0), 25 mM EDTA, and 2% 2-mercaptoethanol) and mixed. Then the samples were incubated at 65°C for 10 min. After centrifugation (4°C, 10, 000 g, 10 min), the supernatant was then transferred to an equal volume of extraction cocktail [chloroform/isoamyl alcohol (24/1)] in order to extract the RNA. The mixture was centrifuged at 10, 000 g for 10 min, after which, the supernatant was mixed with LiCl (3 M) and stored at −20°C for 12 h. After centrifugation (4°C, 10, 000 g, 20 min), the pellets were suspended in SSTE buffer [1 M NaCl, 1% SDS, 1 mM EDTA, and 10 mM Tris-HCl (pH 8.0)]. A chloroform/isoamyl alcohol (24/1) mixture was added, and then the samples were centrifuged at 10, 000 g for 10 min. The supernatant was transferred to a new RNase-free tube and mixed with 0.7 volume isopropanol. After centrifugation at 10,000 g for 20 min, the pellets were washed with 1 ml 75% alcohol, dried, and resuspended in RNase free water. The total RNA was quantified with a NanoDrop 2000C spectrophotometer (Thermo Scientific, USA) at wavelengths of 230, 260, and 280 nm and the 260/280 nm ratio was determined. RNA integrity was confirmed by running samples through 1% agarose gels. After that samples were stored at −80°C.

First-strand cDNA was synthesized from approximately 3 μg of total RNA using the PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa, Dalian, China) according to the manufacturer's instructions. Quantitative PCR was carried out on a real-time PCR system (Bio-Rad iQ5) using the SYBR Green Kit (TaKaRa, Dalian, China). Thermal cycling conditions were 95°C for 10 min followed by 95°C for 10 s, 56°C for 10 s for 40 cycles, and a melting cycle was then performed from 65° to 95°C. A house-keeping gene actin was used as an internal control for qPCR (Reid et al., 2006), and it was previously used as a reliable internal reference gene in grape under salt stress (Ma et al., 2015). VvPMAs transcripts were quantified after normalization to actin. Results are reported as 2^−ΔCT or 2^−ΔΔCT, where ΔCT is the number of PCR cycles required for the log phase of amplification for the experimental gene minus the same measure for actin and ΔΔCT is the ΔCT of treatment minus the ΔCT of control (0 h) for the same gene (Livak and Schmittgen, 2001). Primer sequences used for the qPCR analyses were shown in Supplementary Table 1.

RNA was extracted from the root of a Crimson seedless grape plant and cDNA synthesis was carried out as described above. The full-length cDNA encoding VvPMA1 was isolated through PCR using Pfu DNA polymerase (TaKaRa, Dalian, China) and the primers listed in Supplementary Table 2. The generated PCR fragment was purified and directly inserted into the pEASY-Blunt vector following the manufacturer's instructions (Transgen Biotech. Co., China). The plasmids were transformed into E. coli DH5α cells. Finally, the 9 recombinant plasmids were prepared and then sequenced, respectively.

Plasmid pMP625 containing the promoter and terminator of PMA1 (yeast endogenous PM H⁺-ATPase) was kindly provided by Dr. Palmgren (University of Copenhagen, Denmark). The full-length coding regions of VvPMA1α and VvPMA1β were amplified from pEASY-Blunt-VvPMA1α and pEASY-Blunt-VvPMA1β using the primers VvPMA1α-P1, 5′CTCGAGATGGGAGGCGACAAATCC 3′, which includes an underlined XhoI site, and VvPMA1-P2, 5′ GGGCCCTCATACTGTGTAATGCTGTTGGATT 3′, with an underlined ApaI site; VvPMA1β-P1, 5′CTCGAGATGCGCATTTCATTGAATT 3′ and VvPMA1-P2, 5′ GGGCCCTCATACTGTGTAATGCTGTTGGATT 3′. The PCR products were subcloned into XhoI and ApaI sites of pMP625 to generate plasmid pMP625-VvPMA1α and pMP625-VvPMA1β. All PCR amplifications were carried out using Pfu DNA polymerase, which exhibits the lowest error rate of any thermostable DNA polymerase, and all constructs were sequenced to confirm their identity.

The yeast strain RS72 (also gifted from Dr. Palmgren) was cultured as described previously (Ekberg et al., 2010). In RS72 (MATa ade1–100 his4–519 leu2–3, 112), the natural constitutive promoter of the PMA1 has been replaced by the galactose-dependent GAL1 promoter. As PMA1 is essential for yeast growth, RS72 is only able to grow on galactose-containing medium. Using this strain, plasmid-borne plants' PM H⁺-ATPases under the control of the constitutive PMA1 promoter can be tested for their ability to rescue pma1 mutants on glucose-containing medium.

Yeast was made competent for plasmid uptake by treatment with lithium acetate and polyethyleneglycol according to Gietz et al. (1995). Positive transformants were selected on synthetic medium containing 2% (w/v) galactose, 0.7% (w/v) yeast nitrogen base without amino acids (Difco), 0.2 mM adenine, 0.4 mM histidine, without leucine after 3 days of growth at 30°C. And transformants were grown for 3 days at 30°C in liquid synthetic medium containing 2% galactose. Then, 5 ml transformed yeast cells were transferred to solid at two different concentrations (OD₆₀₀ 0.1 and OD₆₀₀ 0.01) or 20 ml liquid synthetic medium at OD₆₀₀ 0.01 containing either 2% galactose at pH 5.5 or 2% glucose at pH-values of 5.5, 4.5, and 3.5. The media were buffered with 50 mM succinic acid adjusted to various pH-values with Tris. Growth was recorded after incubation for 3 days at 30°C. Each experiment was replicated independently three times.

Statistical analyses were performed using the software SigmaPlot 10.0 (Systat Software, Inc). Values were expressed as mean ± SD. All comparisons were done using Student's t-test for independent samples. In all cases, the confidence coefficient was set at 0.05.

The plants were subjected to salt stress (50 mM NaCl) for 0, 3, 12, 24, or 72 h. ATP hydrolytic activity in PM vesicles obtained from plant roots increased under the NaCl treatments. Moreover, the enzyme activity was up regulated within 3 h of treatment (P < 0.01). The maximum increase in enzyme activity resulted in an activity about 2.7-fold greater than that of the control at 24 h. Although the enzyme activity decreased between 24 and 72 h, the activity level remained above the control activity level (P < 0.01) (Figure 1).

To explore the prevailing isoforms of PM H⁺-ATPase genes in grape root treated with salt stress, the VvPMA genes were screened first from genome of the near-homozygous PN40024 genotype of V. vinifera cv Pinot Noir. According to the retrieved candidate VvPMA proteins and Pfam software verification, we obtained putative 14 VvPMAs in silico. Subsequently, we eliminated overlapping protein sequences and short fragments based on the functional domain sequence. Finally, a total of 7 candidate genes were identified as possible grape PM H⁺-ATPase genes (Table 1).

In order to determine the evolutionary relatedness among the PM H⁺-ATPases from grape and Arabidopsis, a phylogenetic tree was constructed using MEGA 5.0 software and the deduced amino acid sequences. The phylogenetic analysis showed PM H⁺-ATPase protein from each of the two species to fall within the same clades with a high degree of similarity. Additionally, VvPMAs were classified into four subfamilies according to Arango et al. (2003; Figure 2).

In seedling roots grown under normal conditions, expression of VvPMA1 and VvPMA3 was markedly higher than VvPMA4 and VvPMA6 expression (P < 0.01), but expression levels did not significantly differ between VvPMA1 and VvPMA3 (P > 0.05). Additionally, although expression of VvPMA2, VvPMA5, and VvPMA7 was found in other tissues (Supplementary Figure 1), their expression was undetectable in roots using either of two different primer pairs (Figure 3).

As the predominant genes expressed in root, VvPMA1 and VvPMA3 were greatly up-regulated under 50 mM NaCl stress (Figure 4). Moreover, VvPMA1 expression was not up-regulated significantly until 12 h of treatment, and then reached its maximum level at 48 h, when it was as much as 3.6 times greater than that of the control (P < 0.01; Figure 4). Expression remained greater than control levels until 72 h of treatment. For VvPMA3, expression was highest at 6 h of treatment, when it was 3.3-fold greater than at 0 h. After that, expression reduced, but remained above 0 h levels until 72 h of treatment (P < 0.01; Figure 4).

To study the function of VvPMA1, the full-length cDNA sequence of VvPMA1 was cloned and sequenced through reverse transcription PCR (RT-PCR). Interestingly, two different cDNA sequences of VvPMA1 were obtained. Comparison of the two VvPMA1 cDNA sequences (designated VvPMA1α and VvPMA1β) with the genomic sequence revealed that the second intron (93 bp) was retained in the VvPMA1β cDNA (Figure 5A, Supplementary Figure 2). Furthermore, RT-PCR using VvPMA1α and VvPMA1β specific primers (Figure 5A, Supplementary Table 1) detected both transcripts, respectively after 3 h of 50 mM NaCl stress (Figure 5B), proving that AS of VvPMA1 really occur.

The predicted VvPMA1α and VvPMA1β proteins had identical C domains, whereas a frame shift resulted in a loss of 35 amino acids at the N-terminus of VvPMA1β (Supplementary Figure 3).

To investigate whether both VvPMA1α and VvPMA1β encode functional proton pumps and test for their ability to complement pma1, we expressed them in mutant yeast RS72. The results were shown in Figure 6. The negative control (the strain MP625 with the yeast PMA1 under the GAL1 promoter) grew well in galactose medium only (Figure 6A). The strains with VvPMA1α and VvPMA1β both supported yeast growth on either glucose solid or liquid medium (Figures 6B,C), however, at pH 5.5 and more acidic pH levels, VvPMA1β increased yeast growth by a greater amount than VvPMA1α did (Figures 6B,C).

To ascertain the regulatory mechanisms controlling the two VvPMA1 splice variants under salt stress, the expression levels of the two variants were examined in root under salt treatments. The results showed expression of VvPMA1α to be much higher than VvPMA1β expression in salt and fresh water treated roots (Figures 7A,B). However, VvPMA1α expression had not too much change until 48 h of treatment (Figure 7A), whereas, VvPMA1β expression increased rapidly and peaked at 12 h under salt treatment (Figure 7B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.