Tonoplast-located GmCLC1 and GmNHX1 from soybean enhance NaCl tolerance in transgenic bright yellow (BY)-2 cells

ABSTRACT

Genes encoding ion transporters that regulate ion homeostasis in soybean have not been carefully investigated. Using degenerate primers, we cloned a putative chloride channel gene (GmCLC1) and a putative Na⁺/H⁺ antiporter gene (GmNHX1) from soybean. Confocal microscopic studies using yellow fluorescent fusion proteins revealed that GmCLC1 and GmNHX1 were both localized on tonoplast. The expressions of GmCLC1 and GmNHX1 were both induced by NaCl or dehydration stress imposed by polyethylene glycol (PEG). Using mitochondrial integrity and cell death as the damage indicators, a clear alleviation under NaCl stress (but not PEG stress) was observed in both GmCLC1 and GmNHX1 transgenic cells. Using fluorescent dye staining and quenching, respectively, a higher concentration of chloride ion (Cl^–) or sodium ion (Na⁺) was observed in isolated vacuoles in the cells of GmCLC1 and of GmNHX1 transgenic lines. Our result suggested that these vacuolar-located ion transporters function to sequester ions from cytoplasm into vacuole to reduce its toxic effects.

INTRODUCTION

Soybean, one of the most important cash crops, is classified as a moderately NaCl-tolerant plant (Maas & Hoffman 1977). Understanding the NaCl tolerance mechanism in this crop plant may ultimately help improve its yield on saline lands, and the NaCl tolerance conferring genes in soybean may also be applicable to crops that are more sensitive to NaCl (e.g. carrot, orange and rice). Homeostasis of sodium ion (Na⁺) and chloride ion (Cl^–) are important mechanisms to reduce NaCl stress in higher plants. However, very few studies on genes regulate ion homeostasis in soybean, although an old physiological study implicated that Cl^– homeostasis is a major mechanism of soybean to achieve NaCl tolerance (Abel 1969).

Relatively little is known about the ion transporters that regulate Cl^– homeostasis in plants. Chloride channels (CLCs) are a group of voltage-gated Cl^– channels originally reported in animals (Chen 2005). There are a few reports on the study of CLCs in plants, including a systemic search for CLC genes in Arabidopsis thaliana (Hechenberger et al. 1996). Plant CLCs are characterized by their 11 transmembrane segments, cytosolic N- and C-termini, and an extracellular hydrophobic region S4 (Czempinski et al. 1999). These Cl^– channels may play diverse cellular functions, such as stabilizing transplasma membrane electrical potential, regulating cell volume and transcellular chloride transport (Hechenberger et al. 1996; Barbier-Brygoo et al. 2000). The subcellular localizations of most plant CLC proteins are still unclear. A CLC gene cloned in tobacco has been found to encode a mitochondrial membrane-located CLC protein (Lurin et al. 2000), whereas a recent report on the characterization of genes encoding tonoplast-localized CLC channels in rice has also been found (Nakamura 2005♯1534).

By contrast, regulation of ion homeostasis of Na⁺ by Na⁺/H⁺ antiporters (NHX) has been studied extensively (Blumwald 2000), although genes encoding such antiporters in soybean have not been reported.

Two major types of plant NHX have been identified, which are located on plasma membrane (the salt overly sensitive (SOS)1 type) (Shi et al. 2002) or tonoplast (the NHX type) (Blumwald 2000). The Arabidopsis SOS1 has been involved in transporting Na⁺ across the plasma membrane and may play an important role in long distance Na⁺ transport in plants (Shi et al. 2002).

However, the activity of tonoplast NHX is much higher in the NaCl-tolerant Plantago maritima than the NaCl-sensitive Plantago media (Staal et al. 1991), suggesting that some plants may sequester Na⁺ into vacuoles to enhance tolerance to NaCl stress. The important role of NHX has been further confirmed by the observations that overexpressing NHX proteins confer NaCl tolerance in transgenic A. thaliana (Apse et al. 1999) and transgenic Lycopersicon esculentum (Zhang & Blumwald 2001). Systematic studies of the NHX family in A. thaliana also support the notion that NHX proteins are NaCl tolerance determinants and have a major function in vacuolar compartmentalization of Na⁺ (Yokoi et al. 2002).

To investigate the role of ion transporter genes in NaCl tolerance of soybean, we have cloned a CLC (GmCLC1) and a NHX (GmNHX1) gene from soybean using degenerate primers. By comparing the localization of these two putative ion transporters using yellow fluorescence protein (YFP) fusions, we have shown that they are located in similar subcellular compartments (tonoplast). Both GmCLC1 and GmNHX1 genes are induced by NaCl and dehydration stresses. When ectopically expressing these genes in tobacco bright yellow 2 (BY-2) cells, the possible roles of these genes in relation to NaCl tolerance are demonstrated (1) by their protection effects upon NaCl treatment; and (2) by the increase of vacuolar ion concentrations.

MATERIALS AND METHODS

Cloning of GmCLC1 and GmNHX1

To collect RNA samples for cloning the target genes, soybean (Glycine max L. Merr.) plants were grown in a greenhouse with temperature and humidity ranging from 24 to 28 °C and 60–80%, respectively. The seeds were first germinated in sand irrigated with water. After opening the first trifoliate, the seedlings were irrigated with modified Hoagland’s solution [in mm: KNO₃, 5.0; Ca(NO₃)₂, 5.0; MgSO₄, 2.0; KH₂PO₄, 1.0; H₃BO₃, 0.05; and in µm: Fe-EDTA, 13.5; MnSO₄, 9.0; ZnSO₄, 0.8; CuSO₄, 0.3; Na₂MoO₄(2H₂O), 0.08] supplemented with 150 mm NaCl for 3 d. Total RNA samples were extracted using a modified phenol: chloroform:isoamylalcohol (v/v/v; 25/24/1) protocol (Ausubel et al. 1995). The cDNA samples for cloning steps subsequently described were generated from total RNA samples (0.1 µg µL⁻¹ reaction mixture) in reverse transcription reactions using Moloney murine leukemia virus (MMLV)-reverse transcriptase (Gibco BRL, Grand Island, N.Y., USA) at 42 °C for 1 h, according to the manufacturer’s protocol.

Degenerate primer PCRs were used to obtain partial clones of GmCLC1 and GmNHX1. The following PCR protocol was used: 94 °C 1 min; 50 cycles of 94 °C 1 min, 54 °C 1 min and 72 °C 1 min; and 72 °C 5 min in a 25 µL reaction mixture [5.0 µL of first strand cDNA, 5.0 mm MgCl₂, 0.2 mm dNTPs, 0.8 µm of each primer, 0.5 U Taq DNA polymerase (Roche, Mannheim, Germany) and 1× PCR buffer]. For cloning GmCLC1, the primer pair HMOL615 (5′-GAYTAYGARATHAAYGARAA-3′) and HMOL617 (5′-GAIGCNCCNGTRTGNACCAT-3′) were used to generate a ∼500 bp fragment, followed by a nested PCR reaction using HMOL616 (5′-GARAAYATHGCIG GNTAYAA-3′) and HMOL617 to obtain a ∼200 bp fragment. For cloning GmNHX1, the primer pair HMOL612 (5′-TNTTYGGN GARGGNGTIGTIA-3′) and HMOL613 (5′-WSCATNACIATNCCRCARAA-3′) were used to generate a ∼300 bp fragment (N = A/C/G/T; Y = C/T; R = A/G; W = A/T; S = C/G; H = A/C/T).

DNA sequences of full-length coding regions of GmNHX1 and GmCLC1 were obtained by 5′ and 3′ rapid amplification of cDNA ends (RACE) using the Smart RACE cDNA ampification kit (Clontech K1811-1, Hiedelberg, Germany), according to the manufacturer’s protocol. Five mL of 5′ or 3′ RACE-Ready cDNA was used as the starting template for the PCR reactions [50.0 µL reaction mixture containing 0.2 mm deoxynucleotide triphosphate (dNTPs) mix and 0.2 µm of each primer, 1 U Advantage 2 Taq DNA polymerase (Clontech) and 1× PCR Advantage 2 buffer (Clontech)]. In nested PCR reactions, 1 µL of the primary RACE PCR products was used as template.

For GmCLC1, the gene specific primers (GSPs) for 5′-and 3′ RACE were HMOL793 (5′-GCAAAACACACA CACAGAATAGAAGC-3′) and HMOL792 (5′-ATGC GGGGTACAAGTTTCTTGCTG-3′), respectively; the nested GSPs for 5′-and 3′ RACE were HMOL969 (5′-CGC CCGTGGAAATATAGGAAACC-3′) and HMOL970 (5′-GGCCTGGAATACCTGAAATCAAAGC-3′), respectively. For GmNHX1, the GSPs for 5′-and 3′ RACE were HMOL800 (5′-CTCACGATCTGTAGAGTGCCTGC-3′) and HMOL799 (5′-AGTGTTGACAGGTCTACTTAG TGC-3′), respectively; the nested GSPs for 5′-and 3′ RACE were HMOL798 (5′-TGAAGGGTCGATTTGGTTGAG GTC-3′) and HMOL901 (5′-GCATACCTGTCCTACAT GCTTGCT-3′), respectively. The 5′-and 3′ RACE products were isolated by agarose gel electrophoresis and were purified by High Pure PCR product purification kit (Boehringer 1732 668, Ingelheim, Germany) before being subcloned into pBluescript II KS(+). DNA sequences of the clones were determined by the Genetic Analyzer ABI prism 3100 system after being subjected to thermal cycling reactions using the ABI PRISM dRhodamine Terminator Cycle Sequencing Ready Reaction kit (Perkin Elmer 402078, Wellesley, MA, USA), according to the manufacturer’s protocol.

Clones containing the full-length coding regions were amplified from cDNA samples. The PCR reaction mixture composition was similar to RACE reactions previously described. The primer pairs used for amplifying full-length coding regions of GmCLC1 and GmNHX1 were HMOL1155 (5′-GAGTACGCGGGGAATCCAAAAC-3′) and HMOL1156 (5′-CCAGCTGAATAGAGAGAC-3′), and HMOL1151 (5′-GGCTTTGAACTTGACTCAG-3′) and HMOL1152 (5′-GCTCACTCCACAAAACTTATGC-3′), respectively. The PCR products were then cloned into pBluescriptII KS(+). The clone identities were confirmed by DNA sequencing on both strands using T3 primer, T7 primer and GSPs used in RACE reactions.

The gene-encoding soybean β-amylase (identical to GenBank accession D50866) was obtained in a separate experiment using the Clontech PCR-Select^TM cDNA Subtraction Kit (K1804-1) (data not shown) and was cloned into pBluescriptII KS(+).

DNA sequence and phylogenetic analysis

The amino acid sequence encoding by the corresponding full-length coding regions of GmCLC1 and GmNHX1 was predicted using DNA translation programs (http://www.expasy.org/). Homology searches were performed with the Position-Specific Iterated basic local alignment search tool (PSI-BLAST) program (http://www.ncbi.nlm.nih.gov/BLAST/). Conserved domain search (CDS) was used to identify conserved domains (CDs) that are present in the GmCLC1 and GmNHX1. The possible structural models and topologies of the proteins were predicted by the hydropathy profiles using transmembrane prediction (TMpred) (Hofmann & Stoffel 1993) and transmembrane topology prediction (TopPred) (von Heijne 1992; Claros & von Heijne 1994) programs. Putative signal peptides or organelle targeting peptides were searched by Prediction of Protein Sorting Signals and Localization Sites in Amino Acid Sequences (PSORT) (Nakai 2000) and iPSORT (Bannai et al. 2002) programs (http://www.psort.org/).

To study the phylogenetic relationship of GmCLC1 and GmNHX1 with other CLC and NHX homologues, multiple sequence alignment was performed using the ClustalW program (Thompson, Higgins & Gibson 1994) (built-in in Bioedit), and the phylogenetic trees were constructed with the MEGA2 program (version 2.1) (Kumar et al. 2001). Tree topology was calculated by neighbor-joining method (Saitou & Nei 1987) and 1000 replicates were used for bootstrap test.

Northern blot analysis

For NaCl treatments, the surface-sterilized seeds were germinated in filter papers containing modified Hoagland’s solution. After germination, 1-week-old seedlings of uniform growth stage were transferred to a hydroponic system containing the same culture medium. After the opening of the first trifoliate, the seedlings were treated with Hoagland’s solution supplemented with 125 mm NaCl. The youngest fully expanded trifoliate of treated plants were collected for total RNA extraction after 0-, 1-, 4-, 8-, 72- and 144-h treatment. For polyethylene glycol (PEG) treatment, the soybean plants were germinated and grown similarly, except that the resulting seedlings were treated with Hoagland’s solution (control) or Hoagland’s solution supplemented with 5% PEG 6000 or 125 mm NaCl (as positive control). The youngest fully expanded trifoliate was collected for total RNA extraction after 48-h treatment.

Northern blot analysis was performed as previously described (Sambrook & Russell 2001). Hybridization was performed using ultrasensitive hybridization buffer (ULTRAhyb, Ambion, Austin, TX, USA) at 42 °C (probe concentration = 25 mg mL⁻¹). Antisense single-stranded DNA probes were labelled with digoxygenin (DIG) (Roche) according to the manufacturer’s manual. To produce probes, the following forward and reverse primers were used, respectively: GmCLC1 probes, HMOL1155 (5′-GAGTACGCGGGGAATCCAAAAC-3′) and HMOL1156 (5′-CCAGCTGAATAGAGAGAC-3′); GmNHX1 probes, HMOL1151 (5′-GGCTTTGAACT TGACTCAG-3′) and HMOL1152 (5′-GCTCACTCCA CAAAACTTATGC-3′); β-amylase, nested PCR primer 1 (5′-TCGAGCGGCCGCCCGGGCAGGT-3′) and nested PCR primer 2R (5′-AGCGTGGTCGCGGCCGAGGT-3′), provided in the Clontech PCR-Select cDNA Subtraction Kit (K1804-1).

Subcellular localization

To study the subcellular localization of our target proteins, fusion proteins were constructed in which a YFP was fused to the C-terminal of GmCLC1 or GmNHX1. The recombinant constructs were expressed under the control of the cauliflower mosaic virus 35S promoter in a binary vector (Brears et al. 1993) and were transferred into the tobacco BY-2 cells (Nagata, Nemoto & Hasezawa 1992) by Agrobacterium-mediated (strain LBA4404) transformation. After selection of positive transformants on MS medium (Sigma, St. Louis, Missouri, USA) containing kanamycin (50 µg mL⁻¹), the subcellular localization of YFP signals was observed under a confocal laser scanning microscope. The YFP signal was excited by Argon laser at 514 nm. A 545/40 filter set was used and the confocal images were collected using the Bio-Rad Radiance 2100 system (Hertfordshire, England).

Analysis of mitochondrial integrity

Cells were treated with 100 mm NaCl or 2% PEG for 1 h before staining with 10 µg mL⁻¹ rhodamine123 (Rh123) for 1 h. The signal of Rh123 was excited by green HeNe laser (Bio-Rad, Hertfordshire, England) at 543 nm. The filter set HQ590/70 (Bio-Rad, Hertfordshire, England) was used and confocal images were collected by the Bio-Rad Radiance 2100 system. Twenty to 25 cells were counted for each sample for statistical analysis.

Cell viability assay

Cells were treated with 100 mm NaCl or 2% PEG for 24 h before staining with 0.4% trypan blue (T 8154, Sigma). Stained cells were observed under light microscope. Around 300 cells were counted for each sample.

Isolation of protoplasts and vacuoles

Protoplasts and vacuoles were isolated from suspension BY-2 cell cultures using a protocol described in Mettler & Leonard (1979).

Qualitative detection of changes in Cl^– and Na⁺

Lucigenin (N,N′-dimethyl-9,9′-bisacridinium dinitrate) (catalogue number (cat♯) L 6868, Invitrogen, Carlshad, CA, USA) fluorescence quenching was used to detect the changes in Cl^– concentration qualitatively. Higher quenching (lower lucigenin signal) suggests a higher concentration of Cl^–. Vacuoles used in this experiment were pre-treated with 100 mm NaCl in 0.4 m mannitol for 1 h before incubation with 10 µm lucigenin. The lucigenin signal was excited by argon laser (Bio-Rad, Hertfordshire, England) at 457 nm. A 530/60 filter set was used and the confocal images were collected using the Bio-Rad Radiance 2100 system.

Sodium Green indicator (cat♯ S6901, Invitrogen) was used to stain the intracellular Na⁺. The BY-2 cells were pre-treated in 100 mm NaCl for 1 h before staining with 10 µm Sodium Green indicator for 30 min. The Sodium Green signal was excited by argon laser at 514 nm. A 545/40 filter set was used and the confocal images were collected using the Bio-Rad Radiance 2100 system.

RESULTS

Cloning and phylogenetic analysis of GmCLC1 and GmNHX1

The cDNA clones including full-length coding regions of CLC and NHX-type ion transporters in soybean were yet to be reported. We designed degenerate primers according to consensus regions of the homologous members of these proteins (as previously mentioned in Materials and Methods) to clone partial cDNA fragments of CLC and NHX from soybean. Subsequently, 5′- and 3′-RACE experiments were conducted (see Materials and methods) to reveal DNA sequences of the full-length coding regions of two ion transporter genes in soybean, namely GmCLC1 and GmNHX1.

PSI BLAST searches were performed on predicted protein products of GmCLC1 and GmNHX1. Most proteins showing strong homology to GmCLC1 and GmNHX1 belonged to the CLC and NHX families, respectively. Through the CDS associated with PSI BLAST, we identified a voltage-gated CLC domain (deduced from the database pfam00654; % of alignment = 100.0%; CD length = 372.0) at residues 143–565 of the GmCLC1 protein and a Na⁺/H⁺ exchanger family domain (deduced from the database pfam00999; % of alignment = 97.2%; CD length = 397.0) at residues 28–435 of the GmNHX1 protein.

We also performed phylogenetic analysis (see Materials and methods) to determine the possible relationship of GmCLC1 and GmNHX1 to the published eukaryotic CLC and NHX proteins, respectively (Fig. 1). In the CLC-rooted phylogenetic tree (Fig. 1a), three main groups were formed including plant CLC, animal CLC and yeast CLC. Within the plant CLC group, GmCLC1 belonged to a different branch other than the branch containing the mitochondrial membrane-located tobacco CLC (NtCAA64829) (Lurin et al. 2000). In the NHX-rooted phylogenetic tree (Fig. 1b), GmNHX1 together with other plant NHX proteins (tonoplast localized) formed a tight group which is distinct from the plant SOS1 group (plasma membrane localized) and the animal Na⁺/H⁺ exchanger group.

Figure 1. Phylogenetic analysis of GmCLC1 and GmNHX1. Phylogenetic relationship of GmCLC1 (a) and GmNHX1 (b) in relation to other anion/CLCs and NHX, respectively, were performed using the ClustalW and MEGA (version 2.1) programs, as described in Materials and Methods. One thousand replicates were used for all possible tree topology calculation. The first two letters of each protein label represent the abbreviation of genus and species names. Except for GmCLC1 (GenBank accession number AY972079) and GmNHX1 (GenBank accession number AY972078), the two-letter abbreviation was followed by the GenBank accession number of the corresponding protein. * indicates that NtCAA64829 is the only CLC gene cloned in plants that have a known subcellular localization (located in mitochondria) (Lurin et al. 2000). ** and ***, respectively, denote NHX and SOS1 NHX previously shown to locate on tonoplast or on plasma membrane (Apse et al. 1999; Shi et al. 2002; Yokoi et al. 2002). The scale bar indicates the estimated number of amino acid substitutions per site. At, Arabidopsis thaliana; Ce, Caenorhabditis elegans; Cn, Cymodocea nodosa; Dd, Dictyostelium discoideum; Dm, Drosophila melanogaster; Ec, Escherichia coli; Hs, Homo sapiens; Mm, Mus musculus; Ms, Medicago sativa; Nc, Neurospora crassa; Nt, Nicotiana tabacum; Os, Oryza sativa (japonica cultivar-group); Rn, Rattus norvegicus; Rr, Rattus rattus; Sc, Saccharomyces cerevisiae; Se, Salicornia europaea (for AAN08157); Sp, Schizosaccharomyces pombe; Ss, Synechococcus sp. PCC 7942 (for AAK37770) or Suaeda maritima ssp. salsa (for AAK53432 and AAP15178); St, Solanum tuberosum (for CAA71369); Tm, Torpedo marmorata; Tt, Tetragonia tetragonioides; Zm, Zea mays. CLC, chloride channel; NHX, Na⁺/H⁺ antiporters; SOS, salt overly sensitive.

PSORT, TMPred and TopPred analyses were performed to determine the possible location of transmembrane segments based on hydropathy plots (see Materials and methods). GmCLC1 and GmNHX1 have 10–12 and 11–12 transmembrane segments, respectively (Fig. 2). The number of transmembrane segments is similar to previously identified plant CLC and NHX proteins (Barbier-Brygoo et al. 2000; Putney, Denker & Barker 2002). Predicting from the animal model, the hydrophobic domain IV of GmCLC1 was presumably extracellularly located (Fahlke et al. 1997). Furthermore, three conserved amino acid sequence motifs GxGxPE, GKxGPxxH and PxxGxLF (Barbier-Brygoo et al. 2000) were located between hydrophobic domains II-III, III-IV and V-VI of GmCLC1, respectively (Fig. 2a). The GKxGPxxH motif present on the extracellular loop between hydrophobic domains III-IV was involved in isoform-specific anion selectivity in yeast and animal (Fahlke et al. 1997).

Figure 2. Membrane topologies of GmCLC1 (a) and GmNHX1 (b). Possible membrane topologies of GmCLC1 and GmNHX1 were predicted by hydropathy profiles (see Materials and methods). *, **, and ***, respectively, denote conserved amino acid sequence motifs (GxGxPE, GKxGPxxH, PxxGxLF, as shown in a) (Barbier-Brygoo et al. 2000) located between hydrophobic domains II-III, III-IV and V-VI of GmCLC1. The hydrophobic domain IV in GmCLC1 is presumably extracellular, as deduced from the animal model (Barbier-Brygoo et al. 2000). Multiple alignment of GmCLC1 and GmNHX1 were performed by ClustalW built-in in the BioEdit program. The first two letters of each protein label represent the abbreviation of genus and species names. Except for GmCLC1 and GmNHX1, the two-letter abbreviation was followed by the GenBank accession number of the corresponding protein. At, Arabidopsis thaliana; Dd, Dictyostelium discoideum; Hs, Homo sapiens; Ms, Medicago sativa; Nt, Nicotiana tabacum; Os, Oryza sativa (japonica cultivar-group); Tm, Torpedo marmorata; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

Subcellular localization of the GmCLC1 and GmNHX1 proteins

To understand the physiological roles of GmCLC1 and GmNHX1 in soybean, it is important to determine the subcellular localization of these proteins. All previous studies concluded that NHX proteins are tonoplast localized (Apse et al. 1999; Yokoi et al. 2002); therefore, it is highly likely that GmNHX1 is also localized on tonoplast. PSORT and iPSORT programs, however, all predicted that GmCLC1 does not possess any N-terminal signal, and the subcellular localization of this protein is uncertain.

We experimentally investigated the subcellular localization of GmCLC1 and GmNHX1 by fusing them to the YFP (Haseloff 1999). In this study, YFP was fused to the C-terminal of the target proteins. The recombinant constructs were constitutively expressed under the control of the cauliflower mosaic virus 35S promoter. After transformation into the tobacco cell line BY-2 (Nagata et al. 1992) (see Materials and methods), the localization of YFP signals was studied via confocal laser scanning microscopy. In BY-2 cells, the vacuoles (v) occupied most of the cell volume with transvacuolar strands of cytoplasm (tvs) spanning the cell body (Fig. 3a–c). In both GmCLC1-YFP and GmNHX1-YFP transgenic lines, YFP fluorescence signals were concentrated on tonoplast (Fig. 3a, b) and were relatively weak at tvs regions. Using protoplasts generated from BY-2 cells, similar results were obtained (Fig. 3g, h). When YFP control was expressed in transgenic BY-2 cells, strong signals were obtained in all parts of the cell and protoplast except inside the vacuoles (Fig. 3c, i).

Figure 3. Subcellular localization of the GmCLC1-yellow fluorescence protein (YFP) and GmNHX1-YFP fusion proteins. The subcellular localization of YFP signals (in green) in cells (a–c) or in protoplasts (g–i) for GmCLC1-YFP (a and g), GmNHX1-YFP (b and h) and the YFP control (c and i) were visualized by confocal laser scanning microscopy (see Materials and methods), and the structures of the corresponding transgenic BY-2 cells were observed using differential interference contrast (DIC) (d and j, e and k, and f and l, respectively). Scale bar = 10 µm. v, vacuole; tvs, transvacuolar strands of cytoplasm; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

Induction of the GmCLC1 and GmNHX1 genes by NaCl and PEG treatments

The expression of GmCLC1 and GmNHX1 under NaCl stress was studied using Northern blot analysis. When soybean plants were subjected to a short-term NaCl treatment (up to 8 h), an initial rise of both GmCLC1 and GmNHX1 mRNA levels were observed after 1 h, followed by a drop at 8 h after treatment (Fig. 4). In parallel, the leaves of soybean under NaCl stress drooped within 1 h and recovered about 8 h after NaCl treatment. The stomatal conductance (gsw) to water vapour (mol H₂O m⁻² s⁻¹) dropped to less than 50% of the baseline within 10 min after NaCl treatment (data not shown). The accumulation of Na⁺ in leaves was minimal during the first 8 h after treatment (data not shown). Therefore, the initial rise of GmCLC1 and GmNHX1 mRNA levels may be a result of dehydration responses. To support this hypothesis, we also measured the mRNA level of the β-amylase gene, which encodes a dehydration-responsive enzyme in plants (Kaur, Gupta & Kaur 1998; Todaka, Matsushima & Morohashi 2000). A similar rise (4 h) and drop (8 h) of β-amylase mRNA levels was observed (Fig. 4). When soybean plants were subjected to long-term NaCl treatment (over 72 h), the steady-state mRNA levels of all the GmCLC1, GmNHX1 and β-amylase genes remained high (Fig. 4).

Figure 4. Expression of GmCLC1 and GmNHX1 in soybean leaves under NaCl stress. The short-term and long-term effects of NaCl treatments on the expression of GmCLC1, GmNHX1 and the β-amylase genes in soybean leaves were studied by Northern blot analysis (see Materials and Methods). Surface-sterilized seeds were germinated in filter papers containing modified Hoagland’s solution. After germination, 1-week-old seedlings of uniform growth stage were transferred to a hydroponic system containing the same culture medium. After opening of the first trifoliate, the seedlings were treated with Hoagland’s solution supplemented with 125 mm NaCl. The youngest fully expanded trifoliate leaves of treated plants were collected for total RNA extraction after 0-, 1-, 4-, 8-, 72- and 144-h treatment. Ethidium bromide staining of rRNA was used for total RNA quantitation. CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

To further explore the possible regulation of GmCLC1 and GmNHX1 gene expression by dehydration, soybean plants were subjected to PEG treatment to mimic the dehydration stress (Jia, Zhang & Liang 2001). The expression of GmCLC1, GmNHX1 and β-amylase genes were all induced by 5% (w/v) PEG (Fig. 5).

Figure 5. Expression of GmCLC1 and GmNHX1 in soybean leaves under polyethylene glycol (PEG) treatment. Seeds were germinated and grown as described in Fig. 4. After opening of the first trifoliate, seedlings were treated with Hoagland’s solution (untreated control) or with Hoagland’s solution supplemented with 5% PEG or 125 mm NaCl (positive control). The youngest fully expanded trifoliate leaves were collected for total RNA extraction after 48-h treatment. Northern blot analysis was performed using GmCLC1, GmNHX1 and the β-amylase genes as described in Materials and Methods. Ethidium bromide staining of rRNA was used as the reference for total RNA quantitation. CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

Effect of expressing the GmCLC1 and GmNHX1 genes on mitochondrial integrity and survival of BY-2 cells under NaCl and PEG treatments

While ectopic expression of the Arabidopsis NHX protein in transgenic A. thaliana (Apse et al. 1999) and L. esculentum (Zhang et al. 2001) led to enhancement of NaCl tolerance in the whole plant level, the role of CLC proteins on NaCl tolerance remained unproven. Moreover, there is no direct study of the protection effects of CLC and NHX proteins in individual plant cells under NaCl stress. The GmCLC1-YFP and GmNHX1-YFP constructs transformed into BY-2 cells (as previously mentioned) containing all transmembrane segments and conserved regions related to the functions of CLC and NHX (Barbier-Brygoo et al. 2000; Putney et al. 2002). A previous study showed that the addition of green fluorescent protein in the C-terminal of AtCLC-d did not affect its functional complementation of CLC-defective yeast mutants (Hechenberger et al. 1996). Consequently, we tested the response of our transgenic cell lines under NaCl and PEG treatments. Loss of mitochondrial integrity and cell death was used as the indicator of stress damage, because salinity and dehydration will cause membrane damage (peroxidation) in organelles probably because of accumulation of reactive oxygen species (Scandalios 1993). Rh123 is a fluorescent stain of mitochondria in living cells and is assumed to distribute itself electrophoretically into the mitochondrial matrix in response to transmembrane potential differences. This dye can only be taken up by active mitochondria while de-energized or depolarized mitochondria give weak and diffuse fluorescent signals (Wu 1987; Petit 1991).

When untreated wild-type BY-2 cells were stained with Rh123, discrete signals were observed (Fig. 6a). However, when the cells were pre-treated with 100 mm NaCl or 2% PEG before the Rh123 reaction, diffuse signals were obtained (Fig. 6b, c). These results suggest that NaCl stress will lead to damage of mitochondrial membrane in BY-2 cells. When subjected to the same treatments, GmCLC1-YFP and GmNHX1-YFP transgenic cells exhibited a clear protection effect under NaCl stress (Fig. 6h, k, respectively). Such alleviation was not observed in BY-2 cells expressing the YFP control (Fig. 6e). No obvious protection effects of GmCLC1-YFP and GmNHX1-YFP were found when the cells were subjected to PEG stress (Fig. 6i, l), supporting the roles in ion homeostasis of these genes.

Figure 6. Mitochondrial membrane integrity in transgenic BY-2 cells under NaCl stress and dehydration stress. Wild-type (a–c) BY-2 cells and transgenic BY-2 cells containing YFP (d–f), GmCLC1-YFP (G, H and I) and GmNHX1-YFP (j–l) gene constructs were pre-treated in a cell culture medium without stress (a, d, g and j), with 100 mm NaCl (b, e, h and k) or with 2% polyethylene glycol (PEG) (c, f, i and l) for 1 h before staining with 10 µg mL⁻¹ Rh123 for another hour. The signal of Rh123 was observed using a confocal laser scanning microscope (see Materials and methods). 20 to 25 cells were counted for each line, and the experiment was triplicated using different transgenic cell lines. Scale bar = 50 µm. YFP, yellow fluorescence protein; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

Cell death caused by NaCl stress was visualized by staining with trypan blue (Hou & Lin 1996) (Fig. 7). The results were similar to the mitochondrial integrity studies previously discussed. Trypan blue staining of NaCl that induced cell death was mainly found in wild-type BY-2 cells and in the YFP control (Fig. 7b, e, respectively). For both GmCLC1-YFP and GmNHX1-YFP transgenic constructs, protection effects were observed under NaCl (Fig. 7h, k, respectively), but not under PEG stress (Fig. 7i, l, respectively).

Figure 7. Cell viability of transgenic BY-2 cells under NaCl and polyethylene glycol (PEG) treatments. Wild-type (a–c) and transgenic BY-2 cells containing YFP (d–f), GmCLC1-YFP (g, h and i) and GmNHX1-YFP (j–l) gene constructs were pre-treated in a cell culture medium without stress (a, d, g and j), with 100 mm NaCl (b, e, h and k) or with 2% PEG (c, f, i and l) for 24 h. Treated cells were then stained with 0.4% trypan blue. Non-viable cells were stained blue. Around 300 cells were counted for each line, and the experiment was repeated twice using different transgenic cell lines. Scale bar = 50 µm. YFP, yellow fluorescence protein; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

Table 1 shows the numerical analysis of the data on mitochondrial integrity and cell death.

Table 1.  Effects of NaCl and PEG stresses on mitochondrial membrane integrity and survival of cells^a

Cells	% of cells with intact mitochondria			% of survived cells
	No treatment	NaCl treated	PEG treated	No treatment	NaCl treated	PEG treated
^a Number of cells with mitochondrial integrity and number of dead cells were estimated by the uptake of the fluorescent dye Rh123 and trypan blue staining, respectively (seeMaterials and methods). Cells were treated with 100 mm NaCl or 2% PEG for 1 h (mitochondrial integrity) or 24 h (cell death). Percentage of survived cells = (total number of cells–number of dead cells)/total number of cells. The percentage was presented as the mean value of two experiments (20–25 and 300 cells each for mitochondrial integrity and for cell death experiments, respectively) ± SD. PEG, polyethylene glycol; YFP, yellow fluorescence protein; CLC, chloride channel; NHX, Na^+/H^+ antiporters.
Wild type	100%	34.8 ± 3.79%	39.3 ± 2.56%	97.0 ± 0.54%	67.4 ± 2.92%	70.5 ± 2.38%
YFP control	100%	34.1 ± 3.95%	34.5 ± 6.29%	95.8 ± 1.10%	67.9 ± 2.16%	69.3 ± 1.22%
GmCLC1-YFP	100%	80.0 ± 3.88%	37.4 ± 0.75%	97.0 ± 0.90%	90.8 ± 0.35%	70.8 ± 0.47%
GmNHX1-YFP	100%	75.3 ± 2.76%	38.3 ± 4.67%	97.0 ± 1.52%	90.3 ± 1.11%	71.6 ± 0.92%

Effect of expressing the GmCLC1 and GmNHX1 genes on vacuolar ion concentration

The effects of the GmCLC1 and the GmNHX1 genes on vacuolar ion concentration were studied qualitatively by the quenching effect of Cl^– on the fluorescent dye lucigenin (Wissing & Smith 2000) and by staining of Na⁺ with the fluorescent dye Sodium Green (Mazel et al. 2004).

Unlike Sodium Green, lucigenin could effectively enter vacuoles of BY-2 cells (data not shown). Therefore, we used isolated vacuoles for experiments involving lucigenin. When treated with NaCl, a much stronger quenching (i.e. higher concentration of Cl^–) was observed in vacuoles isolated from transgenic BY-2 cells expressing GmCLC1-YFP(Fig. 8e, f) compared with vacuoles obtained from untransformed BY-2 cells (Fig. 8a, b). Such differential signals were not observed in vacuoles of YFP control (Fig. 8c, d) or in GmNHX1-YFP constructs (Fig. 8g versus Fig. 8h).

Figure 8. Detection of chloride ions (Cl^–) in isolated vacuoles using lucigenin. Vacuoles isolated (see Materials and methods) from wild-type (a and b) and transgenic BY-2 cells containing YFP (c and d), GmCLC1-YFP (e and f) and GmNHX1-YFP (g and h) gene constructs were pre-treated in 0.4 m mannitol without (a, c, e and g) or with (b, d, f and h) 100 mm NaCl for 1 h. The treated vacuoles were then stained with 10 µm lucigenin and the signals were observed under a confocal microscope. Ten to 20 vacuoles were counted for each sample and the experiment was triplicated. YFP, yellow fluorescence protein; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

A much stronger Sodium Green fluorescent signal (i.e. higher concentration of Na⁺), however, was observed in the vacuoles of GmNHX1-YFP transgenic cells (Fig. 9g versus Fig. 9h) when subjected to NaCl treatments. Such enhancement in vacuolar Na⁺ accumulation was not observed in the vacuoles of untransformed BY-2 cell (Fig. 9a, b), the YFP control (Fig. 9c, d) and GmCLC1-YFP transgenic cells (Fig. 9e, f).

Figure 9. Detection of sodium ions (Na⁺) in cells using Sodium Green indicator. Wild-type (a and b) and transgenic BY-2 cells containing YFP (c and d), GmCLC1-YFP (e and f) and GmNHX1-YFP (g and h) gene constructs were pre-treated in a cell culture medium without (a, c, e and g) or with (b, d, f and h) 100 mm NaCl for 1 h. The treated cells were then stained with 10 µm Sodium Green indicator and the signal of Sodium Green indicator was observed under a confocal microscope. Twenty to 30 cells were counted for each sample and the experiment was triplicated. YFP, yellow fluorescence protein; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

Figure 10 shows the numerical analysis of changes in Cl^– and Na⁺ concentration.

Figure 10. Quantitative analysis of lucigenin and Sodium Green indicator signal. The experiment details were described in Figs 8 and 9. Signals of lucigenin (a) and Sodium Green indicator (b) in vacuoles were analysed using the ImageJ1.34n program (Sukumvanich et al. 2004). The total fluorescence intensity measured in pixels was divided by the area of the vacuole or the cell to obtain average pixel fluorescence intensity. In addition, background fluorescence intensity was measured in the same field and was subtracted. Data obtained were analysed by one-way analysis of variance (anova) test, in which case significant differences between individual treatments or lines were determined by the Tukey’s test. **indicates that the mean difference is significant at P < 0.01 level. Open and close bars denote samples with or without NaCl treatment, respectively. YFP, yellow fluorescence protein; CLC, chloride channel; NHX, Na⁺/H⁺ antiporters.

DISCUSSION

Maintaining ion homeostasis is an important strategy for plants to survive under NaCl stress. It helps prevent toxic effects of excessive ions that cause damage to membrane of cytoplasmic organelles (Blumwald 2000; Shi et al. 2000; Zhu 2001). Presumably, ion transporters located in plasma membrane and tonoplast may help in ion exclusion from cells and ion compartmentalization within cells, respectively.

Previous genetic study has tied Cl^– transport to NaCl tolerance in soybean (Abel 1969). In higher plants, Cl^– is permeable (K_m in mM range) to tonoplasts (Kastner & Sze 1987), and the majority of Cl^– may be located in vacuoles (Trebacz, Simonis & Schonknecht 1994). Cl^– transport in tonoplast was demonstrated in several plant species (Plant, Gelli & Blumwald 1994; Pei, et al. 1996; Wissing & Smith 2000).

In this paper, we reported the cloning and functional analysis of genes encoding a tonoplast-located CLC (GmCLC1) and a tonoplast-located NHX (GmNHX1) from soybean. Genes encoding these ion transporters have not been previously characterized. While there are very few reports on functional studies of CLC channels in plants, NHX in other plants were extensively investigated. A parallel study of both GmCLC1 and GmNHX1 may bring new insight, because transport of Cl^– and Na⁺ may be thermodynamically coupled (Lorenzen, Aberle & Plieth 2004).

The successful cloning of CLC and NHX homologues in soybean was supported by several lines of evidences (as previously mentioned in Results), including overall sequence similarities, phylogenetic analyses, predicted membrane topologies and consensus sequence motifs.

Phylogenetic analysis of CLC families (Fig. 2) gave further information on this group of Cl^– channels in plants. The plant CLC group is closer to a subfamily of animal CLCs that contains CLC-6 and CLC-7, putative organelle-located animal CLCs that regulate the vesicular pH (Chen 2005). Our observation of the tonoplast localization of the GmCLC1 protein in transgenic BY-2 cells (Fig. 3) may explain why GmCLC1 and CLC-Nt1 in the plant CLC group are seemingly on separate branches (Fig. 1a).

The patterns of YFP signals for GmCLC1-YFP and GmNHX1-YFP resemble each other (Fig. 3). Because GmNHX1 from soybean falls into the same class of tonoplast-localized plant NHX, it is highly likely that GmCLC1 and GmNHX1 are both localized on tonoplast.

Tonoplast-localized ion transporters may function to sequester cytoplasmic Na⁺ and Cl^– into vacuoles and may lead (1) to lowering the concentration of toxic ions in cytoplasm; and (2) to setting up an osmotic gradient as an adaptation to NaCl-induced physiological drought (Jain & Selvaraj 1997). The possible role of GmCLC1 and GmNHX1 in the adaptation to NaCl and dehydration stresses can be implicated by the expression patterns of the corresponding genes. When soybean plants were treated with NaCl, physiological drought was the immediate stress. It is supported by the observation that the leaves drooped within 1 h and the stomatal conductance dropped to less than 50% of the baseline within 10 min after NaCl treatment (data not shown). A prominent induction of GmCLC1 and GmNHX1 mRNA levels in leaves started to appear within 1 h after NaCl treatment (Fig. 4) when no significant accumulation of Na⁺ in leaves was established at this time. The recovery of leaves after 8 h of NaCl treatment suggest that soybean has acquired some osmotic adjustment mechanisms to resume transpiration during the 8-h period. Interestingly, the steady-state mRNA levels of both GmCLC1 and GmNHX1 dropped after 8 h of NaCl treatment. When the gene encoding the dehydration-responsive β-amylase (Kaur et al. 1998; Todaka et al. 2000) was analysed in the same samples, a similar initial rise (at 4 h) and subsequent drop (at 8 h) of mRNA levels was observed. When subjected to long-term NaCl treatments, all the GmCLC1, GmNHX1 and β-amylase genes remained at a high expression level (Fig. 4). Because dehydration symptoms were also observed in leaves when soybean plants were placed under such long-term treatments, we cannot distinguish whether the apparent NaCl induction effects are results of NaCl per se or whether the consequential drought is caused by stress. The same confusion may be faced when interpreting data from previous studies that showed NaCl induction of CLC and NHX gene expressions in other plants (Hamada et al. 2001; Shi et al. 2002).

The induction of GmCLC1 and GmNHX1 by PEG (Fig. 5) further supports the notion that the gene expression of GmCLC1 and GmNHX1 are at least in part coregulated by drought. This kind of drought-induced gene expression fits well with the physiological roles of the tonoplast-located ion transporters because they are important in setting up an osmotic gradient by keeping a high concentration of ions in vacuoles to maintain water uptake by the cell from the surrounding osmonegative environment (Jain & Selvaraj 1997).

We also performed gain-of-function tests using the transgenic BY-2 cells to investigate the possible relationship between vacuolar ion compartmentation and NaCl tolerance in the cellular level. Two parameters, mitochondrial activities (Fig. 6; Table 1) and cell death (Fig. 7; Table 1), were used in this study. NaCl- and PEG-induced mitochondrial damage likely are results of membrane peroxidation resulting from the accumulation of reactive oxygen species caused by NaCl accumulation or dehydration stress in cytoplasm (Gomez et al. 1999). A clear protection effect against NaCl, but not PEG stress was observed in both GmCLC1 and GmNHX1 transgenic BY-2 cell lines. Furthermore, the differential effects on ion accumulation in tonoplast because of GmCLC1 and GmNHX1 (Figs 8 & 9) expression confirms the specific roles of ion homeostasis of these genes in relation to NaCl tolerance in the cellular level.

In a separate study using tonoplasts isolated from soybean plants, we found that the activities of tonoplast H⁺-ATPase and H⁺-PPase both increased upon NaCl treatment (Yu et al. 2005). These vacuolar proton pumps will establish a membrane potential to drive the cross tonoplast movement of Cl^– (Martinoia, Massonneau & Frangne 2000) and Na⁺ (DuPont 1992).

Summarizing all our findings, we hypothesize that, at least, at the cell level, coordinated expression and function of GmCLC1 and GmNHX1 may control the sequestering of Na⁺ and Cl^– into vacuoles to reduce ionic toxicity and/or physiological drought experienced by the cytoplasm.

ACKNOWLEDGMENTS

This research was supported by the Hong Kong Research Grants Council (HK RGC) Earmarked Grant [Chinese University of Hong Kong (CUHK)-4180/99M to H.-M.L] and Hong Kong University Grants Committee Area of Excellence (HK UGC AoE) on Plant and Fungal Biotechnology Centre. The authors thank Dr L. Jiang for generously supplying the YFP clone and the wild-type BY-2 cells; Mr H.-H. Bai, Mr W.-K.K. Kwok, Mr S.-K. Lo, Ms. M.-Y. Lee and Ms. S.-W. Tong for their technical assistance; and Miss S.-M. Chow for critically reading the manuscript.