Total RNA was extracted from the leaves of switchgrass cv. Alamo using the TRIzol reagent method (Invitrogen, Carlsbad, CA, USA), and the cDNA was synthesized by a TIANScript RT Kit (Tiangen Biotechnol-ogy Co., Ltd., Beijing, China) with an oligo-dT primer. The cDNA amplification primers (NHX-F1 and NHX-R1) were designed based on multiple known sequences of NHX1-like genes from different species (Figure S1). For amplification of full length PvNHX1 cDNA, rapid amplification of cDNA ends (RACE) was performed by the SMARTTM RACE cDNA Amplification Kit (Clontech, Mountain View, CA, USA). Gene specific primers NHX-P1 and NHX-P2 were designed for amplification of 5′ and 3′ untranslated regions, respectively. These fragments were assembled by overlapping sequences to obtain the full length cDNA of PvNHX1. All primers used in this study were listed in Table S1.

The bioinformatics analysis of PvNHX1 were conducted using various bioinformatics tools: the basic local alignment search tool (BLAST) (http://blast.ncbi.nlm.nih.gov/Blast.cgi) was performed on the National Center for Biotechnology Information (NCBI) platform to determine sequence similarity; Multiple sequence alignment was performed for the amino acid sequence alignment using DNAMAN software; Prediction of transmembrane domains were performed with TMHMM Server2.0; Signal peptide prediction was checked using Signal P 4.1 (http://www.cbs.dtu.dk/services/SignalP/); To investigate the phylogenetic relationships of PvNHX1 with other Na⁺ (K⁺)/H⁺ antiporters, phylogenetic analysis was performed with Clustal X and MEGA software using the neighbor-joining method (Tamura et al., 2013).

The open reading frames (ORFs) of PvNHX1 gene (stop codon removed) was amplified using primers NHX-F2, NHX-R2 (Table S1), and resulting in a sequence with Bsa I and Eco31 I sites. The ORF was inserted between the constitutive CaMV 35S promoter and GFP gene in a pBWA(V)HS-GFP (provided by the BioRun Company) expression vector to generate an in-frame fusion of GFP to the C-terminus of PvNHX1. Then, the fusion constructs pBWA(V)HS-PvNHX1-GFP was introduced into rice protoplasts prepared from cell culture. As control proteins for tonoplast targeting, we used pBWA(V)HS-GFP and AtTPK1-RFP. For red fluorescence, mKATE, which was a far-red fluorescent protein, fused to AtTPK1 to stain vacuolar membranes. The protoplasts transformation was performed as following the procedure described by Voelker et al. (2006). Visualization of GFP and RFP in the transformed protoplasts were observed by a laser confocal scanning microscope (FV10-ASW, Olympus, Japan) at a wavelength of 488 and 588 nm, respectively.

To identify the tissue-specific expression of PvNHX1 gene, quantitative real-time PCR (qRT-PCR) assay was performed. Total RNA from root, stem and leaf of switchgrass cultivar “Alamo” were isolated for cDNA synthesis using PrimeScriptTM RT reagent kit with gDNA Eraser kit (TaKaRa, Shiga, Japan). The qRT-PCRs were performed using gene specific primers NHX-F3 and NHX-R3 (Table S1). The switchgrass ubiquitin-1 gene (PvUBQ1, genbank FL955474.1) was amplified as the internal control with specific primers NHX-F4 and NHX-R4 (Table S1).

We also quantified the development stage- specific expressions of PvNHX1 gene in leaves (collected and immediately frozen at −80°C until analysis), including the five elongation (E1, E2, E3, E4, and E5) and three reproductive (R1, R2, and R3) stages of switchgrass cultivar “Alamo” as described by Hardin et al. (2013). To assess the effect of salt on the expression pattern of PvNHX1, we analyzed the transcript levels of PvNHX1 exposed to different salt stress (0, 150, 250, and 350 mM NaCl) for different time interval (0, 10, 20, and 30 days) using qRT-PCR.

The ORFs of PvNHX1 gene was amplified by specific primers NHX-F5 and NHX-R5 (Table S1) and inserted into binary expression vector Ubi1301 (provided by the Sinogene Scientific Company) via the BamH I and Kpn I sites. The transformation, selection and regeneration of switchgrass cultivar “Alamo” plants were performed with Agrobacterium- mediated method as described by Liu et al. (2015). Finally, regenerated plantlets were transferred into soil and cultured in the greenhouse. The plants carrying the empty vector (EV) and wild-type (WT) originated from the same tissue culture process were both taken as controls.

The integration and expression of the transgenes was confirmed by PCR, Southern blot and qRT-PCR analyses. For PCR analysis, genomic DNA of putative transgenic lines and control plants were amplified using primers NHX-F6 and NHX-R6 (Table S1) specific to the expression vector Ubi1301. Southern blot analysis was conducted following the procedure as Liu et al. (2015). In brief, 25 ug of genomic DNA was digested with restriction endonucleases Kpn I, separated by 0.8% agarose gel and subsequently transferred to a nylon membrane (GE Healthcare Life Sciences, Indianapolis, IN). Hygromycin phosphotransferase (hyg) gene amplified by specific primers NHX-F7 and NHX-R7 (Table S1) was chosen as the probe. Probe labeling, prehybridization, hybridization and detection were determined according to instructions of the DIG Labeling and Detection starter kit II (Roche Applied Science, Mannheim, Germany). The purified Ubi1301-PvNHX1 plasmid and DNA from a non-transformed plant were used as a positive and negative control, respectively. For gene expression in PvNHX1 overexpressing transgenic switchgrass plants, leaves were collected from E3 stage of different transgenic lines.

Transgenic and control plants (WT and EV) were transplanted into plastic pots containing a compound medium of soil: vermiculite: humus [1:1:1 (v/v/v)]. Plant height, stem diameter, internode length, tiller number, leaf width and leaf length were recorded. Internode 3 (I3) was used for measuring stem diameter. The leaves of I3 were used to measure leaf blade length and leaf blade width. Ten individual tillers of the same transgenic line were randomly sampled for each parameter measurement.

To evaluate salt tolerance, three independent transgenic lines (L1, L3, and L8), with abundant transcripts of PvNHX1 and no obvious phenotypic changes, were selected as representatives and subjected to leaf senescence assay and salinity stress test. Mature leaves (third leaf from the top) from WT, EV and transgenic lines (L1, L3, and L8) were harvested for leaf senescence assay (Jha et al., 2013). Leaves were cut into 4 cm long pieces and floated in NaCl solution with concentration gradients of 0, 150, 250, 350, and 450 mM for 30 days. When the number of tillers were enough for samplings, the uniform tillers of both transgenic and control plants were chosen and trimmed from soil to sand culture for salinity stress test. They were watered with 1/2 × Hoagland nutrient solution supplemented with gradually increasing concentrations of NaCl (0, 150, 250, and 350 mM) every 2 days and continued for 30 days.

To determine the relative increase in growth difference of the transgenics in salt (150, 250, 350 mM NaCl) relative to control conditions, we measured growth parameters (plant height, leaf width, leaf length, stem diameter) on 0 and 30 days, respectively. The relative increase in growth parameters with and without salt was calculated using the formula: (TG₃₀-TG₀)-(WT₃₀-WT₀).

Leaves of transgenic and control plants were sampled every 10 days for determination of relative water content (RWC), electrolyte leakage (EL), malonaldehyde (MDA), and proline contents. RWC and EL were measured following the description in Bao et al. (2009) and Lutts et al. (1996), respectively; MDA concentration was estimated by the reaction of thiobarbituric acid (TBA) as described by Peever and Higgins (1989); For proline content, the samples were extracted in 3% sulfosalicylic and measured at 520 nm absorbance as described by Bates et al. (1973). In addition, shoots and roots were harvested for Na⁺ and K⁺ detection respectively, using the method described by (Wang et al., 2009).

To determine the molecular mechanism underlying differential phenotype and K⁺ accumulation in the transgenic lines and control plants, we searched for potential cell growth-, flowering-, and potassium transport-related genes in switchgrass genomics resource (https://phytozome.jgi.doe.gov/pz/portal.html#!info?alias=Org_Pvirgatum) (Paudel et al., 2016) using the rice (Oryza sativa L.) proteins as a BLAST query. Three cell growth-related (CNR2, CNR8, FTsZ), flowering-related (FLP3, MADS15, MADS6) and potassium transport-related (HKT4, HAK5, HAK27) genes were isolated from switchgrass. Moreover, three flowering-related genes (FT1, APL1, SL1) which have been functionally identified as key flowering regulators in switchgrass (Niu et al., 2016), were also chosen for gene expression analyses. Leaves from control plants (WT and EV) and three independent transgenic lines (L1, L3, and L8) were collected for analysis. PvUBQ1 was used for internal control. All primers used were listed in Table S1.

All experiments were independently performed three times and values were presented as the mean ± SE. Results were subjected to analysis of variance (ANOVA) using the SPSS software (Version 18.0, IBM, Armonk, NY, USA). Significance was defined as P <0.05 or P <0.01. Figures were created using SigmaPlot version 10 software (Systat Software, Point Richmond, CA, USA).

A putative vacuolar Na⁺ (K⁺)/H⁺ antiporter gene was isolated from switchgrass using homologous cloning and RACE method. The full-length cDNA was 1669-bp, which contained a complete open reading frame (ORF) of 1611-bp nucleotides coding for a polypeptide of 536 amino acids protein. The sequence data of PvNHX1 have been firstly submitted to the GenBank database (accession number: KJ739865). The deduced protein contains 12 putative hydrophobic peak domains and 10 strong transmembrane domains (Figures S2A,B). In the secondary structure, the percentage of alpha helix, extended strand, and random coil were 29, 26, and 45%, respectively (Figure S2C). Moreover, PvNHX1 protein harbors 5 completely conserved domains: NhaP, Na-H-Exchanger, b-cpa1, COG3263, and PRK05326 (Figure S2D). The phylogenetic analysis results showed that PvNHX1 was most closely related to proteins from Zoysia japomca and Oryza sativa (Figure S3).

To determine the subcellular location, a PvNHX1-GFP fusion protein was constructed under the control of the constitutive CaMV-35S promoter (Figure 1A). In the transient expression of the fusion protein in rice protoplasts, Green fluorescence was visible throughout the cytoplasm in protoplasts with control plasmid p35S::GFP (Figures 1F–H), and red fluorescence was visible predominantly along the vacuole membrane (Figure 1C). In protoplasts with p35S::PvNHX1-GFP, green fluorescence (Figure 1B) was consistent with the red fluorescence (Figures 1D,E), indicating the localization of PvNHX1 on the vacuole membrane. Subcellular localization of PvNHX1 was consistent with the phylogenetic analysis.

The expression of PvNHX1 was found in almost all of the tissues in switchgrass. As shown in Figure 2A, relatively higher levels were detected in the leaf and stem which showed 8.57 and 2.78-fold higher (P <0.01) compared with that in root, respectively, indicating PvNHX1 gene preferentially expressed in the leaf tissue.

Developmental stage-specific expression pattern of PvNHX1were also analyzed. PvNHX1 was expressed at a low level in the leaves at the early stages of switchgrass elongation development (E1) and expression started to increase gradually, reaching the maximum value at E3 (9.69-fold than E1, P <0.01); then displayed decreasing trends (Figure 2B). For the reproductive stages, the expression of PvNHX1 exhibited the highest level (15.83-fold) at R2, and start to decline at R3 (1.28-fold).

Significant increase in PvNHX1 transcript levels was observed in salt treated switchgrass. Under 150, 250, and 350 mM NaCl conditions, the expression levels of PvNHX1 in leaves increased 6.52- 10.98-, and 7.72-fold, respectively, relative to the control at normal condition (Figure 2C). Analysis of expression levels of PvNHX1 at various time points following 250 mM NaCl treatment showed 3.17-, 8.52-, and 6.57-fold increase at 10, 20, and 30 days, compared to 0 day, respectively (Figure 2D). Generally, PvNHX1 gene was induced by NaCl treatment, indicating the potent role of PvNHX1 in salt tolerance.

Transgenic switchgrass overexpressing PvNHX1 was generated by using a binary vectors Ubi1301- PvNHX1 through Agrobacterium-mediated transformation (Figure 3A). The expected sized fragments of 1,800 bp (Figure 3B) and 100 bp (Figure 3C) were amplified in Ubi1301- PvNHX1 and Ubi1301, respectively. A differential integration pattern of the transgene was observed in transgenic lines; a total of 2–4 copies of PvNHX1 gene were stably integrated into transgenic lines, whereas no signal was observed in WT (Figure 3D). The occurrence of two or more signals could be due to presence of multiple restriction sites in PvNHX1. Among the transgenic lines, L8 exhibited the highest expression, followed by L1 and L3 (Figure 3E).

We interestedly found that the transgenic lines exhibited better in growth-related phenotypes than control plants under the same greenhouse environment (Figure 4). The vertical height of WT and EV averaged 94.54 cm, whereas that of transgenic plants showed 1.24-, 1.14-, and 1.77-fold increase in L1, L3, and L8, respectively (Figure 4D). The increase of stem diameter in transgenic lines ranged from 18 to 28% (Figure 4E). The transgenic plants also showed an increase by 24 and 32% in leaf width and leaf length, respectively, whereas the corresponding level in control plants averaged only 11.57 and 63.89 cm, respectively (Figures 4F,G).

In leaf senescence assay, the damage caused by salt stress was visualized by the degree of bleaching in leaf tissues. The phenotype of the transgenic and the control plants was similar with normal condition (i.e., 0 mM NaCl); however, the leaves of control plants started to turn yellow at 10 days and exhibited excessive bleaching at 30 days under salt stress. In contrast, the leaves from transgenic lines continuously stayed green even after 30 days of 250 mM NaCl stress (Figure 5A). Similar patterns were also observed on salinity stress test carried out in sand cultures. After grown in normal and low salt conditions (150 mM NaCl) for 30 days, both transgenic and control plants grew vigorously. However, severe growth inhibition, chlorosis and wilting were observed in control plants under high salt conditions; whereas these symptoms were effectively alleviated in transgenic plants, indicating that transgenic plants were more tolerant to salt stress (Figure 5B).

Based on the extra-gain in growth difference of the transgenics in salt relative to control conditions, transgenic lines showed obvious advantages on growth under salt stress. As shown in Figure 5, the relative increase of growth parameters in transgenic lines under salt stress conditions were significantly higher (P <0.05 and P <0.01) than those under normal conditions. Comparatively, the relative increase of height, stem diameter, leaf width and leaf length in transgenic lines averaged 1.93-, 2.48-, 2.35-, and 1.53-fold higher in salt (250 mM NaCl) relative to normal conditions, respectively (Figures 5C–F). These results implied that overexpression of PvNHX1 triggered a significant increase in switchgrass growth and biomass under salt stress.

Transgenic lines and control plants exhibited approximately physiological status regarding RWC, EL, MDA, and proline under normal condition. However, under salt stress, significant differences were observed (Figure 6). Transgenic lines exhibited greater potential to maintain higher tissue water content than control plants. The decrease of RWC in control plants was 21% (P <0.01) and 28% (P <0.01) respectively under 250 and 350 mM NaCl conditions, whereas that of transgenic lines was only 10 and 13%, respectively (Figure 6A).

EL was measured to examine cell membrane stability under salinity stress. The EL in transgenic lines were significantly lower (P <0.01) than control plants under salt stress conditions (Figure 6B). Comparatively, EL of the control plants increased to 117 and 261% under 250 and 350 mM NaCl conditions, respectively, while those of transgenic lines was only 45 and 106%, respectively. Similarly, MDA, an indicator of membrane lipid peroxidation damage, increased gradually with the increase of salt concentrations. The increases in control plants was significantly higher (P <0.01) than those in transgenic lines, resulting in 1.68- and 1.99-fold increases under 250 and 350 mM NaCl conditions, respectively (Figure 6C).

With the increase of salt concentration, the proline content increased sharply in all plants but showed a more rapid increase in transgenic lines (Figure 6D). The increase in proline contents in transgenic lines were 2.08-, 2.22-, and 2.70- fold higher than those in control plants under 150, 250, and 350 mM NaCl stress, respectively.

There was no significant difference in these physiological parameters at the treatment beginning (Figure S4). With prolongation of treatment time, obvious differences were observed between transgenic lines and control plants. In general, transgenic lines exhibited significantly higher RWC, reduced malondialdehyde production, preserved cell membrane integrity and increased proline accumulation at every sampling point.

Transgenic lines and control plants exhibited approximately equal Na⁺ and K⁺ contents as well as K⁺/Na⁺ ratio under normal condition (Figure 7). In transgenic and control plants, salinity treatments led to increased Na⁺ content in all tissues examined, but the increases in control plants were significantly higher than that in transgenic lines (Figures 7A,B). For instance, under 350 mM NaCl stress, control plants accumulated 1.98- and 2.02-fold (P <0.01) higher Na⁺ in shoots and roots respectively, compared to transgenic lines. However, high salinity drastically decreased K⁺ concentration in all tissues (Figures 7C,D). Notably, transgenic lines accumulated a higher concentration of K⁺. Compared with control plants, the transgenic lines exhibited 1.24- and 2.06-fold in shoots and roots, respectively under 350 mM NaCl stress.

As a consequence, the ratio of K⁺/Na⁺ decreased significantly when plants were subjected to salinity stress, but higher ratios were observed in the transgenic lines (Figures 7E,F). For instance, the ratio in shoots and roots of the transgenic lines averaged 0.94 and 0.22, respectively, whereas those in the corresponding tissues of control plants were only 0.29 and 0.10, respectively under 350 mM NaCl conditions.

Having identified genes involved in switchgrass cell growth, flowering and potassium transport, we then analyzed their transcript levels by qRT-PCR in the transgenic and control plants. We found that cell number regulator genes (CNR2, CNR8) and cell division gene (FTsZ) were significantly up-regulated (P <0.01) in transgenic lines, with at least 2.66-fold increase in expression (Figures 8A–C). Transcript levels of three high-affinity potassium transporter genes (HKT4, HAK5, and HAK27) related to potassium transporters were found to be significantly higher (P <0.01) in transgenic lines overexpressing PvNHX1 than in the control pants (Figures 8D–F). Morever, genes involved in flowering, SUPPRESSION OF OVEREXPORESSION OF CONSTANS1 (SOC1)-like gene (SL1), FLOWERING PROMOTING FACTOR 3-like gene (FLP3) and MADS-Box gene (MADS15) were significantly up-regulated in transgenic lines (P <0.01), but expression of FLOWERING LOCUS T1 (FT1), APETALA 1 (AP1)-like gene (APL1), and MADS-Box gene (MADS6) decreased at least threefold compared to the control plants (Figures S5).

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.