The SOS1 transporter of Physcomitrella patens mediates sodium efflux in planta


Author for correspondence:
Begoña Benito
Tel: +34 91 336 4535


  • SOS1 is an Na+/H+ antiporter that plays a central role in Na+ tolerance in land plants. SOS1 mediation of Na+ efflux has been studied in plasma-membrane vesicles and deduced from the SOS1 suppression of the Na+ sensitivity of yeast mutants defective in Na+-efflux. However, SOS1-mediated Na+ efflux has not been characterized in either plant or yeast cells. Here, we use Physcomitrella patens to investigate the function of SOS1 in planta.
  • In P. patens, a nonvascular plant in which the study of ion cellular fluxes is technically simple, the existence of two SOS1 genes suggests that the Na+ efflux remaining after the deletion of the ENA1 ATPase is mediated by a SOS1 system. Therefore, we cloned the P. patens SOS1 and SOS1B genes (PpSOS1 and PpSOS1B, respectively) and complementary DNAs, and constructed the PpΔsos1 and PpΔena1/PpΔsos1 deletion lines by gene targeting.
  • Comparison of wild-type, and PpΔsos1 and PpΔena1/PpΔsos1 mutant lines revealed that PpSOS1 is crucial for Na+ efflux and that the PpΔsos1 line, and especially the PpΔena1/PpΔsos1 lines, showed excessive Na+ accumulation and Na+-triggered cell death. The PpΔsos1 and PpΔena1/PpΔsos1 lines showed impaired high-affinity K+ uptake.
  • Our data support the hypothesis that PpSOS1 mediates cellular Na+ efflux and that PpSOS1 enhances K+ uptake by an indirect effect.


Potassium is the most abundant cation in living cells, while Na+ is less abundant and may be toxic. Although the toxic levels of Na differ, even among taxonomically close species (Benito et al., 2009), it is widely accepted that the substitution of Na+ for a significant part of the cytoplasmic K+ produces toxicity. In contrast to this toxicity, all living cells take up Na+ and require Na+-efflux systems that control the Na+ content, which is especially important in environments in which Na+ is much more abundant than K+. Flowering plants apparently have a single Na+ efflux system, SOS1, which is essential because the Na+ content of soils may be high (Munns & Tester, 2008). Two genes encoding SOS1-like proteins exist in Arabidopsis, namely AtSOS1 (Shi et al., 2000) and AtNHX8 (An et al., 2007), and in Cymodocea nodosa, namely CnSOS1A and CnSOS1B (Garciadeblás et al., 2007b), but apparently only one gene encoding an SOS1-like protein occurs in rice (Oryza sativa) (Martínez-Atienza et al., 2007). Bryophytes are different from flowering plants in that in addition to SOS1 transporters they have Na+ efflux ATPases (Benito & Rodríguez-Navarro, 2003) that are required when the external pH is alkaline (Fraile-Escanciano et al., 2009).

The SOS1 gene of Arabidopsis was initially identified as a genetic locus required for salt tolerance (Wu et al., 1996) and the gene was cloned a few years later (Shi et al., 2000). Consistent with the notion of the absence of other Na+ efflux systems, Arabidopsis sos1 mutant plants are extremely sensitive to the presence of Na+ and show defective root-to-shoot Na+ translocation and partitioning (Shi et al., 2002). Since the cloning of the AtSOS1 gene, significant progress has been made in understanding the function of the SOS1 system, especially regarding its transcriptional and functional regulation by the SOS2 Ser–Thr kinase and the SOS3 Ca2+ sensor, which belongs to the SOS3-like Ca2+ sensor/binding proteins (SCaBPs)/calcineurin B-like (CBL) protein family (Qiu et al., 2002; Zhu, 2003; Kolukisaoglu et al., 2004; Batistic & Kudla, 2009; Li et al., 2009). It is also known that the SOS1 messenger RNA (mRNA) in Arabidopsis is unstable under normal growth conditions but is stabilized under stress conditions through the action of reactive oxygen species (Chung et al., 2008). In addition to the SOS1 from Arabidopsis, the SOS1 transporters from C. nodosa (Garciadeblás et al., 2007b), rice (Martínez-Atienza et al., 2007), wheat (Triticum aestivum) (Xu et al., 2008), Populus (Wu et al., 2007), tomato (Solanum lycopersicum) (Olías et al., 2009) and Thellungiella (Oh et al., 2009) have been characterized.

Although progress has been made in understanding the functional regulation of SOS1 transporters, knowledge of their inherent functional mechanism is still incomplete. As a member of the bacterial NhaP Na+/H+ antiporter group in the CAP1 (cation proton antiporter 1) family of transporters, SOS1 can be expected to mediate Na+ efflux by exchanging Na+ and H+ (Shi et al., 2000; Brett et al., 2005; Pardo et al., 2006). However, although Arabidopsis sos1 mutant plants are extremely sensitive to the presence of Na+, defective Na+ efflux in sos1 plant cells has not been reported. Furthermore, sos1 mutants do not have a higher Na+ content than wild-type plants (Ding & Zhu, 1997; Zhu et al., 1998), which would be expected if SOS1 dominated Na+ efflux in roots. Indeed, in a Thsos1-RNAi (RNA interference) line of Thellungiella, the Na+ content of roots was higher than in the control wild-type line (Oh et al., 2009) but the reasons for this are not clear because the differences are small and the motilities of Na+ from shoots to roots in wild-type and Thsos1-RNAi plants are different (Oh et al., 2009). Indirect support for SOS1-mediated Na+ efflux might be the decreased Na+ efflux (Elphick et al., 2001) and increased Na+ content (Liu & Zhu, 1997; Zhu et al., 1998) exhibited by an Arabidopsis sos3 mutant. However, because SOS3 participates in complex signaling networks (Zhu, 2003; Kolukisaoglu et al., 2004; Batistic & Kudla, 2009; Li et al., 2009) the decreased Na+ efflux of the sos3 mutant might be an indirect effect, occurring, for example, through an increased accumulation of Na+ into the vacuole. In the absence of physical evidence of SOS1-mediated Na+ efflux in plant cells this function has been inferred from SOS1 functional expression in Na+-efflux defective yeast mutants (Quintero et al., 2002; Garciadeblás et al., 2007b; Martínez-Atienza et al., 2007; Xu et al., 2008; Oh et al., 2009; Olías et al., 2009) or bacterial mutants (Garciadeblás et al., 2007b; Wu et al., 2007), and also from Na+-dependent H+ movements in plasma membrane vesicles from plant roots, when the wild-type and sos1 plants were compared (Qiu et al., 2002, 2003; Vera-Estrella et al., 2005; Olías et al., 2009).

In addition to the experimental difficulties entailed in studying SOS1-mediated Na+ efflux, the functional understanding of SOS1 is complicated by other activities of SOS1, either direct or indirect, some of which are revealed by the sos1 mutant of Arabidopsis. In this mutant, high-affinity K+ uptake and K+ content are reduced (Wu et al., 1996; Zhu et al., 1998), the Na+ uptake rate in plants treated with low concentrations of K+ is slower than in wild-type plants (Ding & Zhu, 1997), the K+ efflux produced by the addition of Na+ is higher (Shabala et al., 2005), and membrane traffic and vacuolar functions are affected (Oh et al., 2010). Some of these defects might result from a defective pH homeostasis (Oh et al., 2010). In line with the notion of its functional complexity, SOS1 mediates or stimulates K+ uptake in Escherichia coli (Garciadeblás et al., 2007b). Populus euphratica SOS1 also improves the K+/Na+ ratio in E. coli (Wu et al., 2007) but the flux that is primarily enhanced – K+ influx or Na+ efflux – has not been determined.

To overcome the difficulties that complex vascular plants impose on investigating the cellular functions of plant transporters, the moss Physcomitrella patens has been proposed as a useful tool (Garciadeblás et al., 2007a). This plant is anatomically simple and has an SOS1 gene that encodes a typical SOS1 transporter (Benito & Rodríguez-Navarro, 2003). P. patens also has ENA ATPases that mediate Na+ efflux, and disruption mutants of the ENA genes are available (Fraile-Escanciano et al., 2009). Therefore, we undertook a study of the SOS1 transporter of P. patens (PpSOS1) and report data which strongly support the hypothesis that PpSOS1 mediates Na+ efflux and that the involved mechanism is an electroneutral Na+/H+ exchange, as previously proposed for AtSOS1 (Qiu et al., 2003). PpΔena1/PpΔsos1 double-mutant plants showed excessive accumulation of Na+ at all external pH values, and both the single PpΔsos1 mutant and the double-mutant plants died at relatively low Na+ concentrations. The system encoded by a second SOS1 gene in P. patens, PpSOS1B, is apparently not involved in plasma-membrane Na+ efflux.

Materials and Methods

Plants and growth conditions

The moss P. patens (Hedw.) Brunch & Schimp (Ashton et al., 1979) was maintained axenically in BCDAT medium (Nishiyama et al., 2000), which was supplemented with 5 g l−1 of agar when required. Physiological tests were performed in modified Knop medium (Frank et al., 2005) adjusted to pH 4.5 or 9.0. For pH 9.0, the Knop medium was supplemented with 20 mM TAPS (N-Tris (hydroxymethyl)methyl-3-aminopropanesulfonic acid) and the Ca2+ content was decreased to 0.5 mM. KFM is a K+- and Na+-free medium (Garciadeblás et al., 2007a) that was used for K+ starvation and K+- or Rb+-uptake tests. Plants were grown on agar medium and in either biofermenters or jars with air bubbling in a phytochamber with a discontinuous white light (16 h light at 25°C : 8 h dark at 18°C) at quantum irradiance of 200 μmol m−2 s−1. All media were inoculated with aliquots of moss suspensions that were fragmented using a Polytron PT2100 homogenizer (Kinematica AG, Lucerne, Switzerland). Cultures prepared for growth and flux experiments were approximately 70% gametophores and 30% protonema.

Bacterial and yeast strains, growth conditions and plasmids

E. coli DH5α was routinely used for plasmid DNA propagation. The Saccharomyces cerevisiae strains B31 (Mataade2 ura3 trp1 ena1-4Δ::HIS3 nha1Δ::LEU2) (Bañuelos et al., 2002b), in which the Na+ efflux systems ENA1–4 and NHA1 are absent, and AXT3K (ena1Δ::HIS3::ena4 nha1Δ::LEU2 nhx1Δ::KanMX4), which is deficient in the ENA1–4 and NHA1 Na+ efflux systems, and in the NHX1 vacuolar Na+/H+ antiporter (Quintero et al., 2002), were used for expressing the P. patens SOS1 complementary DNAs (cDNAs). Yeast strains were grown in either complex YPD (1% yeast extract, 2% peptone, 2% dextrose) or minimal SD (synthetic dextrose) medium (Sherman, 1991) media. Growth at variable K+ and Na+ concentrations was tested in arginine phosphate (AP) medium (Rodríguez-Navarro & Ramos, 1984) supplemented with the required K+ and Na+ concentrations. For functional expression tests in yeast cells, the cDNAs were cloned into plasmid pYPGE15 (Brunelli & Pall, 1993) in which expression is under control of the PGK1 gene promoter which exhibits constitutive expression.

Recombinant DNA techniques

Manipulation of nucleic acids was performed using standard protocols or, when appropriate, according to the manufacturers’ instructions. PCR was performed in a Perkin-Elmer thermocycler using the Expand-High-Fidelity PCR System (Roche Diagnostics GmbH, Mannheim, Germany). DNA sequencing was performed in an automated ABI PRISM 3730 DNA analyzer (Applied Biosystems, Foster City, CA, USA). Full-length PpSOS1 and PpSOS1B cDNAs were amplified from P. patens total RNA by standard reverse transcription–polymerase chain reaction (RT-PCR) methods using specific forward and reverse primers (PpSOS1: forward, GTCTAGATACACAATGGAGTCCATGAGCACCG and reverse, GAAGGCTGTCACAACTCTCGGC; PpSOS1B: forward, GTCTAGATACACAATGAATGTCCAGAAGAGCTCC and reverse, GGGGTACCGTTACGAGTGAGGATGGGA), including the ATG and STOP codon triplets that were designed from the NCBI sequence CAD91921 and the JGI genome sequence estExt_fgenesh1_pg.C_150054, respectively. The resulting PCR fragments were cloned into the PCR2.1-Topo vector using the TOPO TA Cloning Kit (Invitrogen). For expression in yeast cells, the full-length cDNAs from P. patens were cloned into the vector pYPGE15 (Brunelli & Pall, 1993). In all cases most of the polylinker sequences preceding the translation initiation codon were eliminated and a sequence environment was created around it that was as similar as possible to (A/U)A(A/C)A(A/C)A AUGUC(U/C) (Hamilton et al., 1987). Mutant SOS1B-1 cDNA, in which a large part of the long C-terminal tail is deleted, was constructed by digesting the pYPGE15 construct containing the PpSOS1B cDNA with XbaI and SspI, which includes the first ATG triplet and all the putative transmembrane fragments. For functional testing in yeast mutants, the product of the digestion was then ligated into the XbaI and SmaI sites of the pYPGE15 polylinker.

Localization of PpSOS1–GFP and PpSOS1B–GFP in P. patens protoplasts

The PpSOS1–GFP and PpSOS1B–GFP constructs were in-frame fusions of the 3′-end of the PpSOS1 and PpSOS1B open reading frames (ORFs) to the green fluorescent protein (GFP) gene of plasmid 35S–AdhI::GFP (Rubio-Somoza et al., 2006). To generate these constructs, the PpSOS1 full-length cDNA was amplified using the BamHI–SOS1 primers (which include the BamHI restriction site) and then the PpSOS1B full-length cDNA was amplified using the BglII–SOS1B primers, which include the BglII restriction site. The PpSOS1 and PpSOS1B PCR cDNAs were cloned into the BamHI site of plasmid 35S–AdhI::GFP at the 5′-end of the GFP gene. The resulting constructs were then used for transient expression in P. patens protoplasts, which were transformed following the polyethylene-glycol method (Hohe et al., 2004). For large-scale protoplast isolation, we used pH-controlled bioreactor cultivation in modified Knop medium with a reduced calcium concentration, according to the protocol described by (Hohe et al., 2004). After transformation, the protoplasts were kept in the dark for 24 h in BCDAT medium supplemented with 6% mannitol and 5% glucose, followed by cultivation in the same medium for 3–4 d under normal growth conditions. The GFP fluorescence signal in P. patens protoplasts was visualized using a confocal ultraspectral Leica microscope TCS-Sp2-AOBS-UV (Leica Microsystems, Mannheim, Germany).

Evans Blue staining and light microscopy observation

For detection of cell death, Physcomitrella samples were incubated for 2 h with 0.05% Evans Blue and washed four times with deionized water to remove excess and unbound dye. Material was then mounted on a slide and examined using a light microscope.

Generation of the PpΔsos1 and PpΔena1/PpΔsos1 knockout lines

The PpSOS1 knockout fragment was constructed in the pTN186 vector by inserting two fragments of the 5′ and 3′ noncoding regions of the SOS1 gene. These two fragments, which extended from the −1073 to the −24 positions and 1167 bp downstream of the STOP codon, were amplified by PCR. The 5′ fragment was inserted between the XhoI and ClaI restriction sites, and the 3′ fragment was inserted between the SphI and NotI restriction sites of the polylinker of the pTN186 vector. In the disrupted fragment, the two PpSOS1 fragments flanked the hygromicin resistance gene, which was under the control of the promoter and terminator of the CaMV (cauliflower mosaic virus) 35S gene. Knockout mutants were generated by transforming P. patens protoplasts with 25 μg of the linear DNA fragment obtained by digesting the knockout vector with the XhoI and NotI restriction enzymes. Stable antibiotic-resistant clones were selected after two rounds of incubation in BCDAT medium supplemented with 30 μg ml−1 of hygromicin (Roche Applied Science, Mannheim, Germany). The first screening of putative disrupted plants was carried out by three independent PCR reactions on genomic DNA purified from transformant plants, one spanning the complete targeted regions and two amplifying part of the marker cassette and part of the 5′- and 3′-gene regions outside the knockout construct. In clones in which these amplifications produced the expected fragments, the fragments were sequenced to check that integration occurred as designed. The basic defects of sos1 mutants were studied in five independent lines and were identical. The results reported in this study were obtained with lines sos1-1 and ena1-1/sos1-1. Real-time PCR showed that the SOS1 mRNA was completely absent in these lines.

Real-time PCR assays

Real-time PCR assays were performed as described previously (Garciadeblás et al., 2003), except that the standard DNA solutions corresponded to the genes studied in this report, namely the SOS1, SOS1B and actin5 genes of P. patens. Total RNA preparations were treated with RNase-free DNase I (40 U in 100 μl; Roche, Applied Science) for 1 h at 37°C. After treatment, RNA was purified using the method described in the RNeasy plant kit (Qiagen). Real-time quantitative PCR of derived cDNA was carried out using the Universal ProbeLibrary system (Roche Diagnostics) in triplicate. Primers and probes for each gene assay were designed using the Universal ProbeLibrary Assay Design Center ( For SOS1: forward primer, CACTCTATCCCCATTGCACA; reverse primer, CAGTGTATTTGTCCGGAACG; probe number, 43. For SOS1B: forward primer, CATGGATTTGTGGCAAGGA; reverse primer, GAATGGCAATCCTTGTGAGG; probe number, 72. For ACT 5: forward primer, GTACGTGGCGATCGACTTC; reverse primer, GGCAGCTCGTAGCTCTTCTC; probe number, 17. Quantitative PCR was performed with the FastStart TaqMan Probe Master (Rox) kit using an Applied Biosystems 7500 real-time PCR system, according to the manufacturer’s instructions.

Determination of Na+ accumulation and Na+ loss in P. patens

To determine cation contents, plant samples were washed, dried, weighed and extracted with 0.1 M HCl. In the extraction solutions the cation concentrations were determined using atomic emission spectrophotometry. Na+ loss was determined in P. patens plants transferred to Knop medium containing the required concentration of Na+ or K+ and the appropriate pH. At intervals, plant samples were transferred to Millipore membrane filters with a pore size of 0.8 μm and washed with 10 mM RbCl/1 mM MgCl2 for 1 min. In flux experiments in which many samples were handled, the DW was estimated from the FW using a previously constructed FW/DW table. The assessment of Na+ fluxes from analyses of Na+ accumulation has been fully described in yeast cells (Rodríguez-Navarro & Asensio, 1977; Rodríguez-Navarro & Ortega, 1982). The rationale of these experiments is that when Na+ is added to the external medium the Na+ content increases as a function of the interaction of two independent processes – a constant influx and an efflux that increases in parallel with the increase in the Na+ content. Eventually, a steady state is reached and from this moment on a constant Na+ content is maintained by plants. At the beginning of the experiment, the initial rate of Na+ uptake is equivalent to influx. If at any time during the course of Na+ accumulation the plants are transferred to an external Na+ concentration which is much lower than the Na+ influx Km, the influx is reduced to undetectable values and the measured rate of the net loss is equivalent to Na+ efflux. In this type of experiment, the concentration of external Na+ should not be decreased to concentrations at which the Na+ electrochemical potential gradient across the plasma membrane favors the passive transport of Na+ from the cytosol to the external medium. If this occurs, Na+ may be lost through many systems that in normal conditions do not mediate efflux. We checked that this problem did not arise in our experimental conditions. If plants incubated with Na+ and showing a constant Na+ content are transferred to a lower Na+ concentration, the steady state is disrupted by the imposition of a reduced influx, and a transitory Na+ loss takes place until a new steady state is reached. For all these experiments, plants grown in Knop medium, pH 4.5, were transferred to the same basal medium with 50 mM Na+ at pH values of 4.5 and 9.0, and their Na+ content was recorded as described in each case.

K+- and Rb+-uptake experiments in P. patens

K+-starved plants were prepared by growing the plants for 21 d in KFM medium. These plants were transferred to KFM containing the required concentration of the tested cation, K+ or Rb+, and the cation concentrations in the bathing medium were determined at intervals using atomic emission spectrophotometry. The sampling volume was small with regard to the total volume and, because of the decrease in volume of the testing medium was negligible, it was not necessary to introduce volume corrections. The time courses of cation depletion can be used for kinetic analyses, considering that the cation influx at any given concentration is the slope of the tangent to the depletion curve at that point. The methods for computing the Vmax and Km values, as well as possible errors of the method, have been described elsewhere (Bañuelos et al., 2002a).

Protein alignments and phylogenetic tree constructions

A phylogenetic tree was constructed using the TreeView program from multiple protein sequence alignments generated using clustal x (Thompson, 1997).

PpSOS1B sequence data from this article can be found in the GenBank/EMBL data libraries under accession number FN555709.


Physcomitrella has two SOS1-like transporters

The existence of an SOS1 gene in P. patens has been reported previously (Benito & Rodríguez-Navarro, 2003), and the genome sequence of P. patens ( revealed the existence of a second SOS1 gene, PpSOS1B. Both SOS1 genes encode very similar proteins of 1153 and 1121 amino acids, respectively, which show all the sequence characteristics of SOS1 antiporters. Unlike the AtNHX8 gene in Arabidopsis (An et al., 2007), which encodes an SOS1-like transporter shorter in the C-terminus, PpSOS1 and PpSOS1B are similar in length, sharing 68% sequence identity. PpSOS1 and PpSOS1B show 53 and 51% sequence identity, respectively, with AtSOS1. We also found two SOS1 genes, which could encode two proteins of similar length, in the genome of the clubmoss Selaginella moellendorffii (– SmSOS1A and SmSOS1B in Fig. 1b). The exon–intron structure of the SOS1 genes of P. patens was very similar to that of the Arabidopsis SOS1 gene but had one more (23 vs 22), and longer, introns. The introns in the PpSOS1 gene were generally 42% longer than in the PpSOS1B gene (Fig. 1a). The phylogenetic tree shown in Fig. 1b shows that the distance between PpSOS1 and PpSOS1B is similar to that between SmSOS1A and SmSOS1B, and both distances were slightly greater than that between AtSOS1 and AtNHX8. In all cases the pairwise distances between transporters from different genera are larger than those between the two transporters encoded by the two SOS1 genes in the same genome. These characteristics indicate that in species with two SOS1 genes, duplications probably occurred after the divergence of land-plant species.

Figure 1.

SOS1 and SOS1B genes of Physcomitrella patens, intron–exon structure and phylogenetic position of the encoded proteins. (a) Scheme showing the position of the introns and comparison with the SOS1 gene of Arabidopsis. (b) Phylogenetic position of the SOS1 transporters of P. patens with regard to selected transporters. Abbreviations: At, Arabidopsis thaliana; Cn, Cymodocea nodosa; Os, Oryza sativa; Pp, P. patens; Sc, Saccharomyces cerevisiae; Sm, Selaginella moellendorffii; Th, Thellungiella halophila. AtNHX1 and ScNHA1 are the outgroups. Accession numbers: PpSOS1, CAD91921; PpSOS1B, FN555709; AtNHX1, NP_19067; ScNHA1, NP_013239.1; OsSOS1, AAW33875; CnSOS1A, CAD20320; CnSOS1B, CAL44986; AtNHX8, AAZ76246; AtSOS1, NP_178307; ThSOS1, ABN04857; SmSOS1A, jgi|Selmo1|75049|e_gw1.0.373.1; SmSOS1B, jgi|Selmo1|75049|e_gw1.0.373.1. The scale bar represents 0.1 substitutions per amino acid site.

Analysis of the transcript abundance of PpSOS1 and PpSOS1B revealed that expression of both genes was enhanced by NaCl and repressed by high pH (Table 1).

Table 1.   Real-time PCR determination of transcript expression from the SOS1 and SOS1B genes of Physcomitrella patens at different pH values and in the absence and presence of 100 mM NaCla
GenepH 4.5pH 9.0pH 4.5 + NaClpH 9.0 + NaCl
  1. aPlants were grown for 6 d under the recorded conditions.

  2. The values reported are the ratio between the gene transcript abundance and the actin transcript abundance.


To determine the cellular localization of PpSOS1 and PpSOS1B, we expressed PpSOS1–GFP and PpSOS1B–GFP fusion constructs in P. patens protoplasts. Cellular GFP fluorescence of P. patens protoplasts expressing the PpSOS1–GFP fusion could not be detected, and therefore the PpSOS1–GFP protein could not be localized. By contrast, the PpSOS1B–GFP fusion protein apparently localized to endomembranes (Fig. 2).

Figure 2.

 Comparison of the localization of Physcomitrella patens SOS1B–GFP fusion protein (PpSOS1B–GFP) (a–c) and the soluble green fluorescent protein (GFP) (d–f) in P. patens protoplasts. (a, d) Images of the GFP fluorescence. (b, e) Images of the chloroplast fluorescence. (c, f) Merged images of GFP and chloroplast fluorescence. In the PpSOS1B–GFP fusion the GFP signal was excluded from the nucleus, while soluble GFP accumulated in the nucleus. Bar, 30 μm.

Functional expression of PpSOS1 in yeast cells

The functions of PpSOS1 and PpSOS1B were studied in two Na+-sensitive yeast strains – B3.1 and AXT3K (Δena1–4 Δnha1 and Δena1–4 Δnha1 Δnhx1, respectively) – either alone or in co-expression with AtSOS2 and AtSOS3 cDNAs (Quintero et al., 2002). In all of these tests PpSOS1B failed to suppress the defect of the yeast mutants (Fig. 3). Expression of the SOS1A transporter of C. nodosa, which is not active in yeast cells, is strongly activated by eliminating the carboxy end of the transporter up to almost the last transmembrane fragment (B. Benito and A. Rodriguez-Navarro, unpublished results). A similar truncation of PpSOS1B was not functional (data not shown). However, PpSOS1 expression suppressed the defect in both yeast strains (we only show the results in the AXT3K strain) and the effect was appreciably enhanced when PpSOS1 was co-expressed with AtSOS2 and AtSOS3. Thus, the PpSOS1 transformant strains grew in up to 20 mM Na+ + 1 mM K+ while the strains co-expressing PpSOS1 with AtSOS2 and AtSOS3 grew in up to 75 mM Na+ + 1 mM K+ (Fig. 3). Consistent with previous expression of SOS1 transporters in yeast cells (Garciadeblás et al., 2007b) we did not find that PpSOS1 mediated Na+ efflux in yeast mutants.

Figure 3.

 Functional expression of Physcomitrella patens SOS1 and SOS1B (PpSOS1 and PpSOS1B, respectively) in yeast cells and activation by Arabidopsis thaliana SOS2 and SOS3 (AtSOS2 and AtSOS3, respectively). Suppression of the sensitivity of AXT3K (ena1Δ::HIS3::ena4 nha1Δ::LEU2 nhx1Δ::KanMX4), which is deficient in Na+ efflux, to high concentrations of Na+ in arginine phosphate (AP) minimal media was studied at the K+ and Na+ concentrations (in mM) shown along the top of the figure. Tests were carried out in parallel with the same strain transformed with the empty plasmid, pYPGE15, and expressing AtSOS2 and AtSOS3. Drops of cell suspensions of three serial 10-fold dilutions were inoculated on the indicated media.

PpSOS1 controls Na+ content at low pH

It was previously demonstrated (Fraile-Escanciano et al., 2009) that PpENA1 mediates Na+ efflux and that this system is functional at high pH, while another system is functional in acidic media. Therefore, to investigate the function of PpSOS1 in planta we disrupted the PpSOS1 gene in wild-type and PpΔena1 lines of P. patens. In a simple test to reveal the complementary functions of PpSOS1 and PpENA1 in controlling the Na+ content, wild-type, PpΔsos1 and PpΔena1/PpΔsos1 lines were grown in Knop medium (5 mM K+), containing 50 mM NaCl, at pH values of 4.5 and 9.0 for 6 d before harvesting the tissue and determining its Na+ and K+ contents (Fig. 4). At pH 4.5, the Na+ contents of both wild-type and PpΔena1 mutant lines were identically low, in contrast to the PpΔsos1 and PpΔena1/PpΔsos1 lines in which the Na+ contents were 2.5- and 4-fold higher, respectively. These results indicate that PpSOS1 was the system responsible for maintaining a low Na+ content at pH 4.5 and that PpENA1 only partially replaced this function. At pH 9.0, the Na+ content of the wild-type and PpΔsos1 mutant lines remained low, whereas in the PpΔena1 and PpΔena1/PpΔsos1 lines Na+ accumulated to a similar, high level. These results indicated that at pH 9.0, PpENA1, rather than PpSOS1, is necessary for maintaining a low Na+ content. The K+ content of the mutant plant lines showed a trend that was opposite to that of the Na+ content, namely the mutant lines with the highest Na+ content had the lowest K+ content.

Figure 4.

 Function of Physcomitrella patens SOS1 and ENA1 (PpSOS1 and PpENA1, respectively) in determining the cellular (a) Na+ and (b) K+ contents of P. patens. Wild-type, PpΔsos1, PpΔena1 and PpΔena1/PpΔsos1 lines of P. patens were grown in Knop medium, pH 4.5 and pH 9.0, supplemented with 50 mM NaCl. After 6 d in this medium the Na+ and K+ contents were determined. The means and SD of three independent experiments are shown. The same letters above the bars indicate that differences are not significant (unpaired t-test, < 0.001 for Na+ content and < 0.05 for K+ content).

To obtain more information about the characteristics of Na+ accumulation in the four lines, we followed the increase in the Na+ content for 12 d (Fig. 5). The initial rates of uptake at pH 4.5 and 9.0 were the same for the four lines, which suggested that the differences in long-term Na+ accumulation among the lines (Fig. 4) occurred because the ena1 and sos1 mutations decreased Na+ efflux: the ena1 mutation at both pH values and the sos1 mutation only at pH 4.5. At this pH the initial rate of Na+ uptake was 9 ± 4 nmol mg−1 DW min−1 (results obtained from five independent experiments).

Figure 5.

Physcomitrella patens SOS1 and ENA1 (PpSOS1 and PpENA1, respectively) reduce Na+ accumulation in P. patens. Wild-type, PpΔena1, PpΔsos1 and PpΔena1/PpΔsos1 lines of P. patens were grown in Knop medium, pH 4.5, and transferred to the same medium containing 50 mM Na+ at pH 4.5 (a) and pH 9.0 (b).

The PpΔsos1 mutation suppresses Na+ efflux in PpΔena1 plants

To demonstrate that the null mutation of PpSOS1 in Ppena1 plants abolishes Na+ efflux at pH 4.5, plants loaded with Na+ for 36 h were transferred to a medium without added Na+ (c. 100 μM Na+ in test medium after the transfer of the plants) in which Na+ influx is virtually zero. P. patens can take up Na+ from 100 μM concentrations but only after several weeks of K+ starvation, and therefore not under the conditions of this experiment (Haro et al., 2010). After transfer, Ppena1/PpSOS1 plants showed extensive Na+ loss, while Ppena1/Ppsos1 plants did not reduce their Na+ content during the same time-period (Fig. 6a). In a second experiment we tested the net loss that occurred when plants grown at 50 mM Na+ for 12 d, with their Na+ content at steady state (Fig. 5), were transferred to 25 mM Na+. Again, Ppena1/PpSOS1 plants reduced their Na+ content, reaching a new steady state in approx. 40 h, while the Na+ content of Ppena1/Ppsos1 plants remained unchanged (Fig. 6b).

Figure 6.

 The PpΔsos1 mutation suppresses Na+ efflux at pH 4.5 in PpΔena1 mutant plants. (a) Time courses of Na+ contents in PpΔena1 and PpΔena1/PpΔsos1 lines that had been incubated for 36 h in Knop medium, pH 4.5, containing 50 mM Na+ and then transferred to the same medium without Na+ added (c. 100 μM Na+ in the test conditions). (b) Time courses of Na+ contents in the same mutant lines that had been grown for 12 d in Knop medium, pH 4.5, containing 50 mM Na+ and then transferred to the same medium containing 25 mM Na+.

Na+ produces cell death in the Δsos1 mutant line

In the absence of Na+ neither PpΔsos1 nor PpΔena1/PpΔsos1 plants showed morphological abnormalities. By contrast, single-mutant and double-mutant plants showed necrosis and cell death when exposed to normally tolerable concentrations of Na+. The most significant signs were chloroplast swelling and browning, plasma membrane retraction, cytoplasmic shrinkage and cytoplasmic clearing. Protonemata and gametophores showed similar symptoms (Fig. 7). Cell death started earlier and was more extensive at increasing Na+ concentrations, and double-mutant plants were more sensitive than PpΔsos1 plants. Protonemata of PpΔena1/PpΔsos1 plants showed cell death at 25 mM Na+ in Knop medium (5 mM K+) but in gametophores the symptoms of Na+-triggered cell death appeared at higher Na+ concentrations (data not shown).

Figure 7.

 Na+-triggered cell death in the PpΔena1/PpΔsos1 line of Physcomitrella patens. Protonemata and gametophores treated with 100 mM NaCl. (a) Brown cells are intercalated between normal cells. (b) Closer view of protonema. (c) Evans Blue staining. (d–f) Closer view of protonemata cells: unaffected (d); slightly affected, showing a decreased number of chloroplasts and chloroplast swelling (e); greatly affected, showing cytoplasmic shrinkage and cytoplasmic clearing (f). (g–i) Closer view of gametophore leaf cells: unaffected (g); slightly affected, with some cells showing a decreased number of chloroplasts and chloroplast swelling (h); greatly affected, with almost all cells showing cytoplasmic shrinkage and cytoplasmic clearing. Bars: (a) 100 μm; (b–f) 50 μm; (g–i) 100 μm.

PpΔsos1 plants show a defective high-affinity K+ uptake

A remarkable effect of the sos1 mutation in Arabidopsis is the reduction of the high-affinity K+ influx (Wu et al., 1996). To test whether the PpΔsos1 mutation produced the same effect in P. patens, we prepared K+-starved plants of the wild-type, PpΔsos1, PpΔena1 and PpΔena1/PpΔsos1 lines by culture in a medium, KFM, without K+ and with a very low Na+ concentration (< 10 μM). The K+-starved plants had a very low Na+ content (10–20 nmol mg−1). Determination of the Na+ content was important because during the K+-starvation process, the K+ content of the plants decrease from 3000 to 900 nmol mg−1, and will take up Na+ if this cation is present. The consequently elevated cytoplasmic Na+ level could impair the K+ permeability, as proposed for sos1 Arabidopsis plants (Qi & Spalding, 2004). The high-affinity K+-uptake system of P. patens does not discriminate between K+ and Rb+, and either of these cations can be used to test high-affinity K+ uptake (Garciadeblás et al., 2007a). We show here the time course of Rb+ depletion because Rb+ influx, but not K+ influx, can be accurately calculated from the depletion curves (Bañuelos et al., 2002a). To test the effects of Na+, Rb+-uptake experiments were carried out in either the absence of Na+ or in the presence of 5 mM Na+. We found that high-affinity Rb+-influx was unaffected in PpΔena1, slightly impaired in PpΔsos1 and greatly impaired in PpΔena1/PpΔsos1 plants (Fig. 8). The presence of 5 mM Na+ did not affect the results (data not shown). Influx rates at 125 μM Rb+ were 0.46, 0.38 and 0.1 nmol Rb+ mg−1 min−1 for wild-type, PpΔsos1 and PpΔena1/PpΔsos1 plants, respectively. The deletion of PpSOS1 increased the Km (from c. 8 to c. 25 μM Rb+) and decreased the Vmax (from 0.46 to 0.38 nmol mg−1 min−1). In the double mutant the Rb+ influx was very slow and the kinetic constants could not be determined. However, the data suggest that the Vmax was decreased and the Km was increased.

Figure 8.

 Disruption of the SOS1 and ENA1 genes of Physcomitrella patens impairs high-affinity K+ uptake. K+-starved plants were prepared in KFM medium. Then, the time course of Rb+ depletion from the KFM medium, pH 4.5, was followed after the addition of 125 μM Rb+. Assays were carried out in 30 ml of medium with the following amounts of plants (DW): wild type, 200 mg; PpΔena1, 190 mg; PpΔsos1, 180 mg; and PpΔena1/PpΔsos1 190 mg.


The molecular cloning of the PpSOS1 gene and the construction of Δsos1 and Δena1/Δsos1 mutants of P. patens offers new opportunities for studying plant SOS1 systems at the molecular level. Our findings indicate that the SOS1 transport system is conserved from bryophytes to flowering plants, including the clubmoss S. moellendorffii, and sequence analysis of the proteins, including the complex C-terminus, does not reveal any significant difference between the most distant members (Fig. 1b). The Synechocystis NhaP and the plant SOS1 antiporters share an ancestral precursor (Hamada et al., 2001) but the shorter C-terminus of SynNhaP suggests that regulatory functions of the C-terminus of land-plant SOS1 transporters do not exist in cyanobacteria/NhaP antiporters. A feature of the regulation of AtSOS1 is its activation by AtSOS2 and AtSOS3 (Qiu et al., 2002; Zhu, 2003; Kolukisaoglu et al., 2004; Batistic & Kudla, 2009; Li et al., 2009) and we found that suppression of the Na+ sensitivity of an Na+ efflux yeast mutant by PpSOS1 was enhanced by the co-expression of AtSOS2 and AtSOS3. This heterologous activation of PpSOS1, which is similar to previous descriptions for the SOS1 transporters of Cymodocea (Garciadeblás et al., 2007b), rice (Martínez-Atienza et al., 2007), tomato (Olías et al., 2009) and Thellungiella (Oh et al., 2009), suggests that the regulation of PpSOS1 and AtSOS1 is similar and highly conserved in land-plant evolution, in that the components of the Arabidopsis pathway recognize and modify the activity of the P. patens transporter.

The function of the second SOS1 transporter, PpSOS1B, in P. patens was not examined in this study. Although the transcripts of PpSOS1 and PpSOS1B are abundant and show a parallel regulation, PpSOS1B may localize to endomembranes (Fig. 2), where it could not mediate cellular Na+ loss. Consistent with this notion, the suppression of Na+ efflux in the PpΔena1/PpΔsos1 line, in comparison with the PpΔena1/PpSOS1 line at pH 4.5, indicates that PpSOS1B does not cooperate with PpSOS1 in Na+ efflux.

It has been previously shown that PpENA1 mediates Na+ efflux (Fraile-Escanciano et al., 2009). Therefore, the function of PpSOS1 was established by comparing Na+ uptake and loss in PpΔena1 and PpΔena1/PpΔsos1 lines. In these comparisons we found a progressive reduction of Na+ accumulation and rapid loss of Na+ at pH 4.5 in PpΔena1 vs PpΔena1/PpΔsos1 plants (Figs 5,6). These results reveal that both PpENA1 and PpSOS1 reduce Na+ accumulation at pH 4.5 with reference to PpΔena1/PpΔsos1 plants, and that PpSOS1 was more effective than PpENA1. Therefore, the disruption of the SOS1 gene in PpΔena1 plants increased Na+ accumulation substantially. By contrast, at pH 9.0 the presence or the absence of the SOS1 gene did not show any effect, which also occurred when the PpENA1 ATPase was active. These results support the fact that Na+ efflux mediated by PpSOS1 was driven by the ΔpH, which is consistent with the notion that PpSOS1 mediates electroneutral Na+/H+ exchange, as proposed for AtSOS1 (Qiu et al., 2003).

Our previous (Fraile-Escanciano et al., 2009) and present results demonstrate that PpENA1 and PpSOS1 are complementary systems. PpENA1 is relevant exclusively at high pH values, where PpSOS1 is inactive, but is unnecessary at low pH values where PpSOS1 is active. The model, an Na+-ATPase for high pH values and an electroneutral Na+/H+ antiporter for low pH values, has been previously proposed in fungi (Benito et al., 2009) and, interestingly, it does not apply to vascular plants. Whilst Na+-ATPases exist in mosses and liverworts, they are not present in clubmosses (Fraile-Escanciano et al., 2009) or in flowering plants (Garciadeblas et al., 2001), indicating the evolutionary loss of this system in the tracheophytes.

The most important consequence of the elimination of the PpSOS1 system is the enhanced Na+ sensitivity of the resulting PpΔsos1 mutant, which suffered Na+-triggered cell death. Consistent with the partial redundancy of the SOS1 and ENA1 functions, the PpΔena1/PpΔsos1 line was hypersensitive to Na+. The Na+-triggered cell death of the PpΔsos1 and PpΔena1/PpΔsos1 lines is consistent with previous reports that link salt toxicity with programmed cell death in plants (Shabala, 2009, and references therein) and in sos1 Arabidopsis mutants (Huh et al., 2002). Furthermore, the cell injuries (Fig. 7) were very similar to those produced by cell-free culture filtrates from Erwinia carotovora or by treatments with Botrytis cinerea spores (Ponce-de-León et al., 2007). Because the deletions of PpSOS1 and PpENA1 affect both Na+ efflux and K+ uptake we have not clearly determined whether cell death was produced exclusively by the high amounts of Na+ or by the interaction of several effects. Furthermore, the multiple effects of the sos1 mutation on membrane traffic and vacuolar functions in Arabidopsis (Oh et al., 2010) strongly suggest that the cell death observed in the PpΔsos1 and PpΔena1/PpΔsos1 lines may not occur exclusively as a result of the increase in Na+ content.

Our analysis of K+ and Rb+ uptake in K+-starved plants of the PpΔsos1 and PpΔena1/PpΔsos1 lines revealed that high-affinity K+ uptake is impaired in these mutants, as it is in the sos1 mutant of Arabidopsis (Wu et al., 1996). In our experiments, plants were prepared in a growth medium that was practically Na+ free (< 10 μM) and the absence or presence of 5 mM Na+ during the uptake tests did not affect the results. Under these conditions we can rule out that Na+ is involved in the impairment of K+ uptake, as proposed for Arabidopsis (Qi & Spalding, 2004). The direct participation of PpENA1 or PpSOS1 on high-affinity K+ uptake can also be ruled out because in the experimental conditions of Fig. 8 high-affinity Rb+ uptake is mediated by PpHAK1 (Garciadeblás et al., 2007a). Therefore, the relevant question is how PpSOS1 enhances K+ uptake without involving Na+. The first observation that sheds light on this question is that the PpΔena1 mutation did not produce any effect on Rb+ influx, but substantially increased the negative effect of the PpΔsos1 mutation (Fig. 8). Therefore, PpSOS1 and PpENA1 are complementary systems for maintaining a high Rb+ influx, as described for Na+ efflux at acidic pH values. Although PpSOS1 and PpENA1 may fulfil the same function of Na+ efflux in the plasma membrane using different driving forces –ΔpH and ATP, respectively – the impairment of K+(Rb+) influx by the ena1 and sos1 mutations may not be caused by the loss of the plasma membrane functions of PpENA1 and PpSOS1. Further research is necessary to establish the roles of these systems in organelle functions (Benito et al., 2009; Oh et al., 2010).

In summary, our results strongly support the fact that SOS1 mediates Na+ efflux in plant cells and confirm that SOS1 controls K+ uptake. This opens a line of research for functional studies of SOS1 transporters from different plant species by expressing them in the PpΔsos1/PpΔena1 line. The substitution of P. patens for yeast cells provides the substantial technical advantage of using a homologous system (Bassham & Raikhel, 2000). Future lines of research include analysis of the activation of SOS1 transporters by the SOS2–SOS3 pathway in P. patens given the activation of PpSOS1 by AtSOS2–AtSOS3 in yeast cells, and amenability of P. patens for targeted mutagenesis of the corresponding members of the CBL and CIPK (CBL-interacting protein kinases) gene families identified in the P. patens genome (Batistic & Kudla, 2009).


We would like to thank Mitsuyasu Hasebe and Tomoaki Nishiyama for their gift of plasmid pTN186, Pablo González-Melendi for his help in the use of the confocal microscope, and Marcel Velduizen for his skilful technical assistance. The financial support for this work was provided by the Spanish Ministerio de Ciencia e Innovación and by the ERDF European program through grant no. AGL2007-61075, which also funded a fellowship to A.F.-E. Additional financial support was provided by the DGUI-UPM Research Group Program.