A conserved primary salt tolerance mechanism mediated by HKT transporters: a mechanism for sodium exclusion and maintenance of high K+/Na+ ratio in leaves during salinity stress



    1. Division of Biological Sciences, Cell and Developmental Biology Section, and Center for Molecular Genetics, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0116, USA and
    Search for more papers by this author

    Corresponding author
    1. Group of Molecular and Functional Plant Biology, Research Institute for Bioresources, Okayama University, 20-1, Chuo-2-chome, Kurashiki, Okayama 710-0046, Japan
      T. Horie. Fax: +086 434 1249; e-mail: horie@rib.okayama-u.ac.jp
    Search for more papers by this author

  • This manuscript is part of the special issue on Drought & Salt Stress. There should be no charge for the colour figures.

T. Horie. Fax: +086 434 1249; e-mail: horie@rib.okayama-u.ac.jp


Increasing soil salinity is a serious threat to agricultural productions worldwide in the 21st century. Several essential Na+ transporters such as AtNHX1 and AtSOS1 function in Na+ tolerance under salinity stress in plants. Recently, evidence for a new primary salt tolerance mechanism has been reported, which is mediated by a class of HKT transporters both in dicots such as Arabidopsis and monocot crops such as rice and wheat. Here we present a review on vital physiological functions of HKT transporters including AtHKT1;1 and OsHKT1;5 in preventing shoot Na+ over-accumulation by mediating Na+ exclusion from xylem vessels in the presence of a large amount of Na+ thereby protecting leaves from salinity stress. Findings of the HKT2 transporter sub-family are also updated in this review. Subjects regarding function and regulation of HKT transporters, which need to be elucidated in future research, are discussed.


Soil salinity is one of the major environmental stresses causing a significant loss of productivity in world agriculture, especially in irrigated lands (Flowers 1999; Zhu 2001; Horie & Schroeder 2004). Harmful effects on plant growth and yield due to salinity stress are caused mainly by two factors: (1) osmotic stress; and (2) ion toxicity (Katsuhara & Kawasaki 1996; Blumwald 2000; Hasegawa et al. 2000; Munns & Tester 2008). In normal soil, water potential in root cells is lower than in the outer environment, and water influx into roots occurs via water channel proteins called aquaporins (Tournaire-Roux et al. 2003; Luu & Maurel 2005; Katsuhara et al. 2008). In saline environments, however, the difference in the water potential between soil and root cells is reduced or even inverted, leading to a reduction in water uptake or loss of water (Boursiac et al. 2005). Growth inhibition, and ultimately, serious tissue damages, are the consequences. When the internal water balance is disturbed by environmental stresses such as salinity stress, plants synthesize and accumulate organic compounds so-called compatible solutes, such as polyols, non-reducing sugars, and amino acids (Gorham, Wyn Jones & Mcdonnell 1985). Many attempts towards molecular breeding of water stress tolerant plants have been made by introducing genes that encode key enzymes for biosynthesis of compatible solutes (Tarczynski, Jensen & Bohnert 1993; Kishor et al. 1995; Nakayama et al. 2000).

Ion toxicity is a general term for various impairments in diverse cellular processes due to elevated ion concentrations. Cellular Na+ toxicity is the predominant ion toxicity caused by salinity stress, resulting in the inhibition of a variety of processes such as K+ absorption (Rains & Epstein 1965), vital enzyme reactions (Flowers & Läuchli 1983; Murguía, Bellés & Serrano 1995), protein synthesis (Hall & Flowers 1973) and photosynthesis (Tsugane et al. 1999). Photosynthetic processes are likely to be the most important cellular reaction to be protected from Na+ toxicity as the disruption of the processes is directly connected to reductions in the carbon fixation and biomass production in plants (Greenway & Munns 1980; Flowers 1999). During salt stress, Na+ is excluded from shoots and K+ is accumulated in shoots, thereby stabilizing the high cytosolic K+/Na+ ratio especially in leaves (Flowers & Läuchli 1983; Gorham et al. 1987; Schachtman, Bloom & Dvorak 1989; Dubcovsky et al. 1996; Apse et al. 1999; Serrano & Rodríguez-Navarro 2001; Shi et al. 2002; Ren et al. 2005; Sunarpi et al. 2005).

Several essential Na+ transporters that detoxify elevated Na+ concentrations have been identified in the model plant Arabidopsis thaliana, including AtNHX1, AtSOS1 and AtHKT1;1 (Wu, Ding & Zhu 1996; Apse et al. 1999; Gaxiola et al. 1999; Shi et al. 2000; Uozumi et al. 2000; Mäser et al. 2002a; Horie & Schroeder 2004; Apse & Blumwald 2007). The AtNHX1 gene was originally identified by sequence homology to the Nhx1 gene in Saccharomyces cerevisiae and to members of the NHX family in Caenorhabditis elegans and Homo sapiens. The AtNHX1 transporter was later shown to function in Na+ sequestration into the vacuole during salinity stress, thereby maintaining a high K+/Na+ ratio in the cytosol (Apse et al. 1999; Gaxiola et al. 1999). The Salt Overly Sensitive 1 (SOS1) gene, which has been identified in a genetic screen (Wu et al. 1996), encodes a putative membrane protein that is homologous to plasma membrane Na+/H+ antiporters from bacteria and fungi (Shi et al. 2000). Analyses of sos1 mutant plants have led to a working model of the SOS1 transporter under salinity stress, in which SOS1 controls Na+ concentrations in xylem sap by unloading Na+ from or loading Na+ into xylem vessels depending on the strength of salinity stress, thus influencing long-distance Na+ transport from roots to shoots (Shi et al. 2002). SOS1 was also suggested to function in direct Na+ extrusion to outer environment at the epidermis of the root tip where meristematic cells lacking large vacuoles for Na+ sequestration exist (Shi et al. 2002). Direct Na+ extrusion by SOS1 from mature epidermal zones of Arabidopsis roots was also suggested by another group of investigators (Shabala et al. 2005). AtHKT1;1 was originally identified as an Arabidopsis homologue of the wheat TaHKT2;1 (Schachtman & Schroeder 1994; Uozumi et al. 2000). When expressed in heterologous systems such as S. cerevisiae and Xenopus laevis oocytes, AtHKT1;1 shows a strong preference for Na+ selective transport. This selectivity contrasts the Na+/K+ co-transport function of TaHKT2;1 in such heterologous systems (Rubio, Gassmann & Schroeder 1995; Gassmann, Rubio & Schroeder 1996; Uozumi et al. 2000). The disruption of the AtHKT1;1 gene renders plants hypersensitive to Na+ accompanied by severe leaf chlorosis and Na+ over-accumulation in shoots in saline conditions (Mäser et al. 2002a; Berthomieu et al. 2003; Gong et al. 2004; Sunarpi et al. 2005; Horie et al. 2006) (Fig. 1). Two detailed working models explaining the protective function of AtHKT1;1 from Na+ over-accumulation in aerial parts under salinity stress have been proposed: Na+ loading into phloem (Berthomieu et al. 2003) and Na+ exclusion from xylem (Sunarpi et al. 2005; Horie et al. 2006; Davenport et al. 2007) (Fig. 2).

Figure 1.

AtHKT1;1 protects leaves from salinity stress. A typical example of the Na+ hypersensitive phenotype due to the loss-of-function mutation in the AtHKT1;1 gene of A. thaliana ecotype Wassilewskija. Mature athkt1-3 T-DNA mutants (Mäser et al. 2002a) and wild-type plants were treated with 100 mm NaCl for a week as described in Horie et al. (2006). The plants were further grown for another week without stress and the photograph was taken afterward. All plants tested (20 to 25 independent plants, respectively) showed similar phenotypes and representative plants were chosen.

Figure 2.

Schematic summaries of the two different working models for the physiological function of AtHKT1;1 under salinity stress. (a) The Na+ recirculation model proposed by Berthomieu et al. (2003) prior to the xylem Na+ exclusion model of AtHKT1;1 (Sunarpi et al. 2005; Horie et al. 2006). In this model, AtHKT1;1 mediates Na+ loading into phloem sieve elements and Na+ ions are transferred to roots, reducing Na+ accumulation in shoots. (b) The xylem Na+ exclusion model of the primary function of AtHKT1;1 and also the OsHKT1;5 transporter in rice plants (Ren et al. 2005) in xylem parenchyma cells, which results in maintaining high K+/Na+ ratio in aerial parts during salinity stress. AtHKT1;1 and OsHKT1;5 absorb Na+ ions into the xylem parenchyma cell from the xylem vessel. The removal of Na+ by AtHKT1;1 and OsHKT1;5 possibly causes membrane depolarization of the xylem parenchyma cell and triggers K+ secretion into the xylem vessel via membrane-depolarization-induced K+ efflux channels such as K+ outward-rectifying (KOR) channel and the non-selective outward-rectifying (NOR) channel (for details see: Wegner & Raschke 1994; Wegner & De Boer 1997). HKT: AtHKT1;1 and OsHKT1;5 transporters (Note that OsHKT1;5 is applicable only for the xylem Na+ exclusion model); Kout Ch: K+ efflux channels.

Recent QTL analyses using rice (Oryza sativa) plants led to the identification of a gene contributing to salt tolerance of an indica rice cultivar Nona Bokra. The gene was found to encode a Na+ transporter, named SKC1 or OsHKT1;5 (Ren et al. 2005), which is an orthologue of AtHKT1;1. OsHKT1;5 was found to function in Na+ exclusion from xylem (Ren et al. 2005). Based on QTL mapping analyses using wheat plants, two major QTL controlling Na+ exclusion from leaves, thereby conferring salt tolerance to durum wheat, have been suggested to encode the AtHKT1;1 orthologues, HKT1;4 and HKT1;5 (Huang et al. 2006; Byrt et al. 2007). The common requirement for such a HKT-mediated Na+ exclusion from the xylem in Arabidopsis, rice and wheat suggests that this is a widely conserved primary salt tolerance mechanism in glycophytes (Horie, Hauser & Schroeder 2009). In this review, we focus on recent findings on a class of HKT transporters, including AtHKT1;1 and OsHKT1;5. We also briefly review findings on other classes of HKT transporters in the mechanism of Na+/K+ transport and homeostasis in plants.


The AtHKT1;1 gene was found to be the only gene encoding a HKT transporter in the genome of A. thaliana (Uozumi et al. 2000). Heterologous expression analyses using yeast and X. laevis oocytes revealed that AtHKT1;1 primarily mediates Na+ selective transport (Uozumi et al. 2000). Dysfunctional mutations in the AtHKT1;1 gene rendered the plants hypersensitive to elevated Na+ concentrations so that leaves and other aerial parts showed severe chlorosis under salinity stress (Mäser et al. 2002a; Berthomieu et al. 2003; Gong et al. 2004; Sunarpi et al. 2005; Horie et al. 2006) (Fig. 1). The lack of a functional AtHKT1;1 leads to Na+ hyper-accumulation in shoots and a reduction in Na+ accumulation in roots compared to wild-type plants (Mäser et al. 2002a; Berthomieu et al. 2003; Gong et al. 2004; Horie et al. 2006), indicating that AtHKT1;1 is a key component in vital Na+ homeostasis in saline conditions.

A detailed working model for AtHKT1;1 in salt-stressed Arabidopsis plants has been proposed by Berthomieu et al. (2003). athkt1;1 EMS mutant plants were found to accumulate less Na+ in the phloem sap with either a minor or an insignificant difference in the Na+ concentration of xylem sap compared to wild-type plants under salinity stress (Berthomieu et al. 2003). Based on these findings, together with a phloem-specific expression pattern of AtHKT1;1 mRNA, AtHKT1;1 was hypothesized to function in the elimination of accumulated Na+ from shoots by loading Na+ into phloem, which is eventually transferred to roots by long-distance transport in the phloem (‘Recirculation’ model; Fig. 2a) (Berthomieu et al. 2003). In this model, AtHKT1;1 would mediate outward Na+ transport from inside of the companion cells to phloem (Berthomieu et al. 2003).

An alternative working model of AtHKT1;1 has been later proposed by Sunarpi et al. (2005) (‘Exclusion’ model; Fig. 2b). Three independent athkt1;1 null mutant alleles generated by T-DNA and fast neutron mutagenesis have been shown to cause significant increases in the Na+ content of shoots and xylem sap compared to wild-type backgrounds (Sunarpi et al. 2005). A peptide anti-AtHKT1;1 antibody that recognizes a hydrophilic loop facing outside of the cell has been produced, and immune-electron microscopy was performed in an attempt to detect tissue specific and subcellular localizations of the AtHKT1;1 transporter protein in plants. As a result, AtHKT1;1 has been found to localize at the plasma membrane of the xylem parenchyma cell (Sunarpi et al. 2005). Longitudinal and cross-sections of beta-glucuronidase (GUS)-stained leaves and roots, derived from Arabidopsis plants transformed with the AtHKT1;1 promoter-GUS DNA construct, further showed strong GUS signals localized in the vicinity of the xylem – that is, xylem parenchyma cells (Sunarpi et al. 2005), supporting the results from the immune-electron microscopy. Based on these findings, a primary role for AtHKT1;1 in the Na+ detoxification mechanism was proposed to be Na+ extrusion from xylem vessels in roots by absorption of accumulated Na+ in the vessels into xylem parenchyma cells, reducing the amount of Na+ ions transported to aerial parts by long-distance Na+ transport, thus protecting leaves from salinity stress (Sunarpi et al. 2005) (Fig. 2b). The xylem Na+ exclusion model was supported by independent analyses with 22NaCl radioactive tracers (Davenport et al. 2007). In the same analyses no evidence for the Na+-recirculation model was found (Davenport et al. 2007). Moreover, Møller et al. (2009) recently demonstrated that overexpression of AtHKT1;1 in the root pericycle that includes xylem parenchyma cells by means of the enhancer trap system (Haseloff 1999) renders the Arabiodpsis plants more salt tolerant, due to increased Na+ influx activity into the targeted parenchyma cells. These findings also strongly support the xylem Na+ unloading function of AtHKT1;1 in response to elevated Na+ levels. By contrast, overexpressing AtHKT1;1 using its innate promoter increased Na+ sensitivity of the transgenic Arabidopsis plants, which led to the hypothesis that AtHKT1;1 is a primary Na+ entry in roots of Arabidopsis plants (Rus et al. 2004). Results from the recent enhancer trap-mediated AtHKT1;1 overexpression (Møller et al. 2009) and Na+ hypersensitive nature of independent athkt1;1 mutant alleles (Mäser et al. 2002a; Berthomieu et al. 2003; Gong et al. 2004; Horie et al. 2006) seem not to be consistent with the proposed Na+ uptake activity of AtHKT1;1 from the outer environment in Arabidopsis roots (Rus et al. 2001, 2004). Together, the primary function of AtHKT1;1 in salinity tolerance mechanisms is unloading Na+ from xylem vessels to prevent Na+ over-accumulation in aerial parts, particularly in leaves (Fig. 2b).

It should be noted, however, that the function of AtHKT1;1 is not restricted to roots, and the xylem Na+ exclusion model does not exclude the proposed function of AtHKT1;1 in phloem Na+ loading (Berthomieu et al. 2003) for the following reasons: (1) significant reductions in the Na+ content of phloem sap were also found in three independent athkt1;1 null mutants compared to wild-type plants under salinity stress; (2) weak GUS signals were detected in the vicinity of phloem of GUS-stained leaf-sections derived from transgenic AtHKT1;1 promoter-GUS plants; (3) expression of AtHKT1;1 in aerial parts was confirmed at both the transcriptional (histochemical GUS-staining and RT-PCR) and at the translational level (immune-electron microscopy); and (4) the transcript level of AtHKT1;1 was found to be up-regulated in response to relatively mild increases in the osmotic pressure caused by NaCl, KCl, sorbitol and mannitol treatments (Sunarpi et al. 2005), leading to the possibility that AtHKT1;1 expression in the phloem might be enhanced in osmotic stress conditions. Na+ exclusion from xylem vessels by AtHKT1;1 in aerial parts would also be required to circumvent Na+ entries into leaves as suggested in the case of its wheat orthologue (TmHKT1;4; see further discussion). Even if such role of AtHKT1;1 in aerial parts is taken into account, important questions remain to be solved: (1) Does AtHKT1;1 reduce the net Na+ concentration in aerial parts via Na+-recirculation?; (2) Do Na+ ions that have passed the barrier of AtHKT1;1 in the root xylem keep accumulating in aerial parts, that is, ending up in vacuoles of various cells (Blumwald & Poole 1985, 1987; Apse et al. 1999; Gaxiola et al. 1999)?; and (3) Otherwise, where do all Na+ ions that accumulated in aerial parts go? A plant physiological study has pointed out that Na+ transport in the phloem appears to be an important mechanism to maintain low Na+ levels in shoots of bean (Jacoby 1979), suggesting that Na+ recirculation could be an important mechanism against salinity stress in plants. Addressing these questions is essential to fully understand the Na+ ion toxicity and the salt tolerance mechanism mediated by AtHKT1;1.

It is also interesting to note that the primary function of AtHKT1;1 in xylem Na+ unloading in roots corresponds to a proposed role for the SOS1 transporter in Na+ extrusion from xylem vessels under severe saline conditions (Shi et al. 2002), suggesting that Na+ sequestration in the root tissue in possible combination with the activity of Na+ exclusion from roots might be a crucial strategy for viability of plants under severe salinity stress.


Potassium (K+) is an essential macronutrient that is required for diverse cellular processes such as osmotic regulation, maintenance of membrane potential, enzyme activity, protein and starch synthesis, respiration and photosynthesis (Flowers & Läuchli 1983; Schroeder, Ward & Gassmann 1994). Various reports have indicated that increasing cytosolic K+ levels relative to Na+, thus increasing the K+/Na+ ratio, is crucial for Na+ tolerance in plants, and maintaining high K+/Na+ ratio in shoots is highly correlated with salinity tolerance in glycophytes (Flowers & Läuchli 1983; Gorham et al. 1987; Schachtman et al. 1989; Gorham, Wyn Jones & Bristol 1990; Shubert & Läuchli 1990; Munns 1993; Dubcovsky et al. 1996; Wu et al. 1996; Zhu, Liu & Xiong 1998; Shi et al. 2002; Ren et al. 2005; Sunarpi et al. 2005). Physiological experiments have shown that the K+ content in glycophytes can reach 1.5 to 5% of the dry weight, and the majority of K+ absorbed in roots is transported to shoots, which is a major limiting factor of the shoot growth and yields of crop plants (Flowers & Läuchli 1983). Since excessive Na+ ions inhibit various important cellular processes many of which are directly correlated with K+ transport and essential functions of K+, it is not surprising to see that K+ alleviates toxic effects of Na+, and that a high K+/Na+ ratio in shoots, especially in leaves, is preferred by glycophytes.

Interesting phenotypes in K+ accumulation were found in athkt1;1 mutant plants subjected to salinity stress. Three independent athkt1;1 null mutant alleles were shown to accumulate significantly less K+ in shoots and xylem sap under salinity stress, which is inverse to robust Na+ accumulation in the same tissues of athkt1;1 mutants. athkt1;1 mutant plants, therefore, show very low K+/Na+ ratios in shoots compared to wild-type plants under salinity stress (Table 1), which demonstrates that the AtHKT1;1 transporter controls not only Na+ concentrations but also K+ concentrations in shoots. AtHKT1;1 expressed in X. laevis oocytes has been found to be primarily Na+ selective in 100 mm alkali cation solutions (Uozumi et al. 2000). Neither inward K+ currents nor, more importantly in this case, outward K+ currents, were recorded in AtHKT1;1-expressing oocytes (Uozumi et al. 2000). Furthermore, degrees of K+ accumulation phenotypes found in shoots and xylem sap of athkt1;1 mutant plants that have been imposed to salinity stress were smaller than those of Na+ accumulations (Sunarpi et al. 2005; Horie et al. 2006). These findings imply that K+ accumulation phenotypes in athkt1;1 mutants are a secondary effect due to the loss of AtHKT1;1-mediated Na+ absorption from xylem vessels. One possible mechanism for an indirect role of AtHKT1;1 in promoting K+ loading into xylem vessels, and thus K+ accumulation in aerial parts, may be a coupling with K+ efflux activity of outward-rectifying K+ channels localized at xylem parenchyma cells (Wegner & De Boer 1997; Gaymard et al. 1998) (Fig. 2b). Since AtHKT1;1 expressed in X. laevis oocytes mediates the transport of a large amount of Na+ in a high NaCl-containing solution (Uozumi et al. 2000), it is likely that Na+ absorption by AtHKT1;1 from xylem vessels depolarizes the membrane potential of xylem parenchyma cells, which could instantly activate membrane-depolarization-induced K+ efflux channels (Fig. 2b).

Table 1.  The AtHKT1;1 transporter-mediated Na+ exclusion from xylem maintains high K+/Na+ ratios in aerial parts of Arabidopsis plants during salinity stress
GermplasmType of mutations in the AtHKT1;1 geneK/Na ratio in leavesa
Soil-grown plants treated with 50 mm NaCl for 6 days
K/Na ratio in shootsb
Hydroponic-cultured plant treated with 75 mm NaCl for 2 days
 Wild typeN/A2.71 ± 0.411.58 ± 0.21
 athkt1;1T-DNA insertion0.25 ± 0.030.55 ± 0.08
 FN1148FN deletionN/A0.74 ± 0.18
 Wild typeN/AN/A2.36 ± 0.39
 athkt1-3T-DNA insertionN/A1.16 ± 0.10

In summary, the AtHKT1;1 transporter plays a direct and vital role in excluding Na+ from xylem vessels under salinity stress, while indirectly stimulating K+ loading into xylem vessels. By controlling Na+ and K+ concentrations, two major Na+ tolerance-determining factors in shoots, AtHKT1;1 exerts a key function in protecting leaves from salinity stress (Figs 1 & 2; Table 1).


An essential QTL named SKC1 (Shoot K+Content) that plays an important role in the salt tolerance mechanism in rice has been mapped using crosses between a salt-tolerant indica variety Nona Bokra and a susceptible japonica variety Koshihikari (Ren et al. 2005). The SKC1 QTL was initially selected as the locus that maintains higher K+ content in shoots during salinity stress. However, measurements of K+ and Na+ contents of shoots and xylem sap of the susceptible parent line and the near-isogenic line (NIL) have revealed that the Nona Bokra-derived SKC1 locus has a larger influence on controlling Na+ contents in shoots and xylem sap than K+ contents (Ren et al. 2005). The SKC1 locus was found to encode a member of the HKT family, OsHKT1;5 (Ren et al. 2005). Electrophysiological analyses with OsHKT1;5 in X. laevis oocytes have further revealed that OsHKT1;5 primarily mediates Na+-selective transport, and that Nona Bokra-OsHKT1;5-mediated inward Na+ currents are significantly larger than the OsHKT1;5 transporter from Koshihikari. Taken together, this difference in the Na+ transport capacity of OsHKT1;5 might be the cause for differences in the sensitivity to salt stress (Ren et al. 2005). Interestingly, it turned out that the findings of the function of AtHKT1;1 reported by Sunarpi et al. (2005) are virtually identical to the reported function of the OsHKT1;5 transporter in K+ and Na+ accumulation in shoots and xylem sap of rice plants (Ren et al. 2005) (Fig. 2b). The genome of the japonica rice cultivar Nipponbare has been reported to contain nine OsHKT genes with two of them being non-functional (Garciadeblás et al. 2003). Among the remaining seven OsHKT transporters OsHKT1;5 was found to be also an AtHKT1;1 orthologue (Garciadeblás et al. 2003). These data indicate that the AtHKT1;1 and OsHKT1;5 transporters mediate a primary salt-tolerance mechanism to detoxify elevated Na+ concentrations in both the model plant Arabidopsis and the monocot crop plant rice in a similar manner (Fig. 2b).

High K+/Na+ ratios in shoots achieved by Na+ exclusion from shoots and selective K+ loading to shoots have also been found to be highly correlated with salinity tolerance of wheat plants (Schachtman et al. 1989; Gorham et al. 1990). Hexaploid bread wheat (Triticum aestivum) that contains three genomes, named A, B and D, was known to show higher salt tolerance than tetraploid durum wheat retaining A and B genomes (T. turgidum) (Gorham et al. 1990; Dubcovsky et al. 1996). The D genome has been found to carry an essential salt tolerance locus that maintains high K+/Na+ ratios in shoots during salinity stress, which was later named the Kna1 locus and mapped on a short region in the long arm of chromosome 4D (Gorham et al. 1987; Dubcovsky et al. 1996).

Independently, a new salt tolerant durum wheat line named Line 149 has been found in a crossed population between wild species of wheat T. monococcum and a cultivar of durum wheat, which showed higher salt tolerance accompanied with lower Na+ concentrations in the leaf blade than typical durum wheat (Munns et al. 2000). Genetic analyses using the Line 149 revealed that two independent loci named Nax1 and Nax2 contribute to the Na+ exclusion from the leaf blade (Munns et al. 2003; Lindsay et al. 2004). Characterization of near-isogenic lines that contain either the Nax1 or Nax2 locus has demonstrated that both loci significantly reduce Na+ transport from roots to shoots by excluding Na+ from xylem (James, Davenport & Munns 2006). The Nax1 locus, however, was found to play an exclusive role in preventing Na+ over-accumulation in the leaf blade by eliminating Na+ from the xylem of the leaf sheath (bottom of the leaf blade) (James et al. 2006). Interestingly, both Nax loci not only reduce Na+ accumulation but also increase K+ accumulation in the leaf blade (James et al. 2006). Mapping analyses of the Nax1 and Nax2 loci identified two candidate genes encoding HKT1;4-type transporters, TmHKT1;4-A1 and TmHKT1;4-A2 (originally named TmHKT7-A1 and TmHKT7-A2), for the Nax1 gene, and one gene encoding a HKT1;5-type transporter, TmHKT1;5-A, for the Nax2 gene (Huang et al. 2006; Byrt et al. 2007). Furthermore, the Nax2 locus on the chromosome 5AL in durum wheat has been found to be homologous to the chromosome 4DL region of bread wheat that includes the Kna1 locus (TaHKT1;5-D), which leads to the conclusion that the Kna1 gene and the Nax2 gene might have the same function (Byrt et al. 2007). Interestingly, these wheat HKT transporters are also orthologues of AtHKT1;1 and OsHKT1;5 (Fig. 3), indicating that all three major salt tolerant QTL phenotypes in wheat are controlled by a class of HKT transporters.

Figure 3.

Maximum likelihood phylogenetic tree of HKT proteins from higher plants. Saccharomyces cerevisiae channels (ScTrk1; ScTrk2) were included for representation of the out-group. Abbreviations: At: Arabidopsis thaliana; Ec: Eucalyptus camaldulensis; Hv: Hordeum vulgare; Mc: Mesembryanthemum crystallinum; Os: Oryza sativa (japonica cultivar-group); Pa: Phragmites australis; Pt: Populus trichocarpa; Sb: Sorghum bicolor; Sm: Selaginella moellendorffii; Ss: Suaeda salsa; Ta: Triticum aestivum; Th: Thellungiella halophila; Tm: T. monococcum; Tt: T. turgidum; Vv: Vitis vinifera; Pp: Physcomitrella patens. Accession numbers for gene names can be found in Supporting Information Table S1. The tree was constructed using phyml with 1000 bootstrap replicates on a gblocks-curated muscle alignment of all proteins (Castresana 2000; Guindon & Gascuel 2003; Edgar 2004). Nodes labelled with an asterisk have a 100% bootstrap support. The scale bar indicates 0.3 substitutions per site.

These important findings in rice and wheat indicate that Na+ exclusion from xylem is a widely used salt tolerance mechanism in monocot plants. Furthermore, findings from AtHKT1;1 analyses fit well with the proposed functions of rice and wheat HKT transporters in monocot plants, implying a possibility that re-absorption of Na+ from xylem vessels is a common and widely conserved mechanism against salinity stress in plants.


HKT transporters have been identified in many plant species including a halophyte ice plant Mesembryanthemum crystallinum (Schachtman & Schroeder 1994; Fairbairn et al. 2000; Uozumi et al. 2000; Horie et al. 2001; Golldack et al. 2002; Garciadeblás et al. 2003; Su et al. 2003; Haro et al. 2005). A phylogenetic analysis using the amino acid sequences of HKT transporters has shown that the HKT transporters are split into two major sub-families, class I and class II HKT transporters (Platten et al. 2006). Remarkable differences between these two classes of HKT transporters can be found in the structure of the putative selectivity pore-forming regions and the selectivity for K+ characterized in heterologous expression systems. Several class II HKT transporters including TaHKT2;1 and OsHKT2;2 show robust K+ permeability in addition to Na+ in yeast and X. laevis oocytes (Rubio et al. 1995; Gassmann et al. 1996; Horie et al. 2001). Biophysical and transport analyses in X. laevis oocytes and yeast using domain-swapped and point-mutated DNA constructs of class I and class II HKT transporters have highlighted a glycine residue that primarily determines robust K+ permeability of HKT transporters (Mäser et al. 2002b) (Fig. 4). HKT transporters have been suggested to be members of a large K+ transporter family, the HKT/Trk/Ktr family, which retains four selectivity pore-forming regions (p-loops) that show similarity to a bacterial K+ channel (Durell & Guy 1999; Durell et al. 1999; Kato et al. 2001). The pivotal glycine residue highlighted in ion selectivity studies of HKT transporters corresponds to one of the four glycine residues that are highly conserved in the p-loops of the HKT/Trk/Ktr family (Durell & Guy 1999; Durell et al. 1999; Mäser et al. 2002b) (Fig. 4). Class I HKT transporters were found to have a serine instead of the glycine at the corresponding position in the first p-loop region (SGGG-type) (Fig. 4). In contrast, class II HKT transporters retain all four glycines in the four p-loops (GGGG-type) with a rare exception of OsHKT2;1 (SGGG-type) (Horie et al. 2001; Mäser et al. 2002b) (Fig. 4). It has been shown that a serine to glycine replacement in the first p-loop region of AtHKT1;1 and OsHKT2;1 transporters that show a poor K+ permeability conferred a robust K+ permeability, while a glycine to serine replacement abolished the robust K+ permeability from typical class II transporters TaHKT2;1 and OsHKT2;2 (Mäser et al. 2002b). Note that an amino acid replacement of the glycine residue at the filter positions has been predominantly found in the first p-loop region thus far. However, our sequence analysis indicated that the replacement of the glycine also occurs in the second p-loop of the Vitis vinifera VvHKT1;2 class I transporter in addition to the replacement in the first p-loop (Fig. 4). Findings from the analysis of the bacterial KtrB transporter further demonstrated the importance of the conserved glycine residues in the putative p-loop regions for K+ permeability, which showed that replacements of each glycine in the p-loops with alanine have led to either decreases in the affinity for K+ or reduced activity of the KtrB transporter (Tholema et al. 2005). These studies indicated that the conserved glycine residues in the putative p-loop regions are crucial for K+ selectivity of the HKT/Trk/Ktr family. Note, however, that the glycine residue alone seems not to be the only relevant factor determining K+ permeability. Several SGGG-type class I HKT transporters were reported to show K+ transport properties in heterologous systems (Fairbairn et al. 2000; Su et al. 2003). Moreover, conserved lysine and arginine residues located at two positions within the eighth transmembrane domain of HKT/Trk/Ktr transporters have recently been reported to be necessary for K+ selectivity (Kato et al. 2007), providing evidence for additional K+ selectivity-relevant amino acids.

Figure 4.

A sequence alignment of the four putative selectivity pore-forming regions (p-loops: PA to PD) of plant HKT and yeast TRK proteins that are available in public databases. Arrow heads indicate the amino acid positions where a glycine residue is conserved in K+ permeable HKT/Trk/Ktr transporters. The structure of the HKT/Trk/Ktr transporter family has been suggested to be distantly related to the bacterial KcsA channel and the glycine residue in each p-loop has further been suggested to correspond to the first glycine of the glycine-tyrosine-glycine (GYG) motif that is highly conserved in the selectivity-pore-forming region of K+ channels (Durell & Guy 1999; Durell et al. 1999). More Na+ selective class I transporters retain a serine residue at the filter position of the PA region, whereas a glycine is conserved in the PB to PD with an exception in the case of VvHKT1;2 that retains an aspartic acid instead of the glycine in the PB as well. Typical class II transporters show robust K+ permeability and the glycine residues are conserved in all p-loops. Note, however, that OsHKT2;1 is a unique class II transporter that retains a serine in the PA region as class I transporters and shows a strong preference for Na+ selective transport in yeast and X. laevis oocytes (Horie et al. 2001; Mäser et al. 2002b; Garciadeblás et al. 2003). New HKT transporters found in primitive plants P. patens and S. moellendorffii, which could be considered to be class III transporters, retain glycine residues in the four p-loops as in typical class II transporters. Na+/K+ selectivities of these transporters have not yet been analysed. Class I, II and potential class III HKT transporters are coloured by sky blue, green and pink, respectively.


The OsHKT2;1 transporter is an unusual class II HKT transporter. It is clearly classified into the HKT2 transporter sub-family in sequence analyses (Platten et al. 2006) (Fig. 3). However, OsHKT2;1 is a SGGG-type, but not a class II-representative GGGG-type transporter, and shows a strong preference for Na+ selective transport in yeast and X. laevis oocytes comparable to typical class I HKT transporters (Horie et al. 2001; Mäser et al. 2002b; Garciadeblás et al. 2003) (Figs 3 & 4). It should be noted here that other groups of investigators reported a robust K+ permeability of OsHKT2;1 expressed in similar heterologous expression systems (Golldack et al. 2002; Jabnoune et al. 2009). The analysis of the ion selectivity of OsHKT2;1 using plant cells is, therefore, essential. Unlike in the case of class I HKT transporters such as AtHKT1;1 and OsHKT1;5, little evidence regarding the physiological roles of the class II HKT transporters in plants was available until recently, when a physiological role of the OsHKT2;1 transporter in rice plants has been elucidated (Horie et al. 2007). It has been reported that K+-starvation triggers significant increases in the OsHKT2;1 mRNA accumulation in roots (Horie et al. 2001; Garciadeblás et al. 2003). Growth experiments using three independent oshkt2;1 null mutant plants in a low Na+ with no added K+ condition revealed that oshkt2;1 mutant plants show meagre growth accompanied with significant less shoot biomass and chlorotic withering in a part of the leaves compared to wild-type plants (Horie et al. 2007). Furthermore, oshkt2;1 null mutant alleles have been shown to cause extreme reductions in the Na+ influx into roots of rice plants, demonstrating together with strong OsHKT2;1 expression at root cortex and endodermis that the OsHKT2;1 transporter directly mediates K+-starvation-induced Na+ influx into rice roots (Horie et al. 2007). These findings provided genetic evidence that Na+ compensates for some roles of K+ by acting as a substitute-nutrient in K+-depleted plants, which has been observed in various plant species and was a long-standing question in classical plant physiology (Mengel & Kirkby 1982; Flowers & Läuchli 1983; Horie et al. 2007, 2008; Mueller-Roeber & Dreyer 2007). Moreover, oshkt2;1 alleles were the first mutant alleles reported, which show robust reductions in the Na+ influx into plant roots (Horie et al. 2007). A special role in nutritional Na+ absorption for growth enhancement in K+-starved rice plants has been assigned to OsHKT2;1 (Horie et al. 2007). Particularly in low Na+ concentration ranges such as less than 0.2 mm, Na+ influx into K+-starved rice roots was primarily OsHKT2;1-dependent (Horie et al. 2007). At higher Na+ concentrations, however, Na+ influx was observed, indicating that Na+ influx into K+-starved rice roots particularly in high mm concentrations is composed of redundant Na+ entry points that are to be unravelled in future investigations (Horie et al. 2007).

The level of HKT2;1 transcripts has been also shown to be up-regulated in wheat and barley plants in response to K+-starvation, as in the case of OsHKT2;1 (Wang et al. 1998; Horie et al. 2001; Garciadeblás et al. 2003). TaHKT2;1 gene expression was found in root cortex cells of wheat plants similar to OsHKT2;1 in rice (Schachtman & Schroeder 1994; Horie et al. 2007). Patch clamp analyses using root cortex protoplasts from wheat demonstrated that K+-starvation induces large inward Na+ currents (Buschmann et al. 2000). Furthermore, knocking down of the TaHKT2;1 gene expression by an anti-sense TaHKT2;1 construct has been shown to reduce Na+ influx into roots of transgenic wheat plants (Laurie et al. 2002). A considerable difference between OsHKT2;1 and HKT2;1 transporters in wheat and barley is that both wheat TaHKT2;1 and barley HvHKT2;1 are typical GGGG-type class II HKT transporters (Fig. 4), which were found to show robust K+ permeability in yeast and X. laevis oocytes (TaHKT2;1: (Schachtman & Schroeder 1994; Rubio et al. 1995; Gassmann et al. 1996) or only in yeast (HvHKT2;1: Haro et al. 2005). Despite the difference in K+ permeability between OsHKT2;1 and wheat and barley HKT2;1 observed in heterologous systems, findings in wheat and barley plants were found to be fairly consistent with those from rice plants, which strongly suggest that the primary physiological role of HKT2;1 transporters are quite similar in planta. Whether the robust K+ permeability found in TaHKT2;1 and HvHKT2;1 can be considered as a special role in K+ uptake in addition to the role in nutritional Na+ uptake is an important question to be addressed. Note however, a group of investigators has raised the point that HKT-mediated K+ transport in heterologous systems was deemed to be an artefact, which does not occur in plant cells (Haro et al. 2005). It is therefore crucial to test whether ion selectivities of class I and class II HKT transporters characterized in heterologous systems can be reproduced in plant cells.

One of the class II HKT transporters in rice, OsHKT2;2, has been identified in a salt tolerant indica rice cultivar Pokkali (Horie et al. 2001). OsHKT2;2 was found to be a GGGG-type Na+-K+ co-transporter (Horie et al. 2001) (Fig. 4). The OsHKT2;2 gene was found to be a pseudogene in both japonica cv. Nipponbare and indica ssp. (Garciadeblás et al. 2003). The pattern of OsHKT2;2 mRNA accumulation was reported to be almost identical to that of OsHKT2;1 in Pokkali plants, including the K+-starvation-stimulated induction (Horie et al. 2001). Furthermore, phylogenetic analyses show that OsHKT2;2 is closely related to HKT2;1 transporters in rice, wheat and barley (Platten et al. 2006) (Fig. 3), suggesting that the physiological role of this transporter in Pokkali plants might be similar to other HKT2;1 transporters. Interestingly, however, the OsHKT2;2 transcript levels have been found to be up-regulated in response to 150 mm NaCl stress in Pokkali plants (Kader et al. 2006). Kader et al. (2006) further reported that expression of the OsHKT2;2 gene is detected in the phloem of leaves treated with 150 mm NaCl. These findings offer the possibility that K+ permeable-type class II HKT transporters including TaHKT2;1 and HvHKT2;1 could play an important K+/Na+ homeostatic role in salt-stressed plants (Fig. 3). Whether these transporters additionally mediate any yet unknown function in salt-stressed plants is an interesting subject to elucidate for a better understanding of roles of class II HKT transporters in plants.


There are indications that seven functional OsHKT genes are expressed in a japonica rice cultivar Nipponbare (Garciadeblás et al. 2003) (Fig. 3). Among them we have relatively little information regarding physiological roles of OsHKT1;1, 1;3, 2;3 and 2;4. OsHKT1;1 and OsHKT1;3 are SGGG-type class I HKT transporters (Fig. 4). OsHKT1;1 was characterized as a low affinity Na+-selective transporter in yeast (Garciadeblás et al. 2003) and more recently OsHKT1;1 and OsHKT1;3 were characterized as Na+ transporters that showed more strict Na+ selectivities than OsHKT2;1 in X. laevis oocytes (Jabnoune et al. 2009). OsHKT2;3 and OsHKT2;4 transporters are GGGG-type class II HKT transporters (Figs 3 & 4) and are highly identical to each other (approximately 93% identity at the amino acid sequence level). Neither detailed K+/Na+ selectivity nor physiological functions in K+ and/or Na+ homeostasis in rice plants of these HKT2 transporters have been described.

Recent genetic analyses using hexaploid wheat and barley suggested that such HKT1 and HKT2 genes also exist in these plant species (Huang et al. 2008). Physiological functions of all HKT transporters have been shown to be directly or indirectly associated with salt tolerance mechanisms in plants with no exception so far (Rus et al. 2001; Laurie et al. 2002; Mäser et al. 2002a; Berthomieu et al. 2003; Ren et al. 2005; Sunarpi et al. 2005; Huang et al. 2006, Byrt et al. 2007 & Horie et al. 2007). Gaining information of Na+/K+ transport mechanisms and physiological functions of these class I and class II HKT transporters will be crucial for drawing a complete picture of roles of HKT transporters in Na+ and K+ homeostasis during salinity stress in these important crop plants.


A phylogenetic analysis using publicly available HKT protein sequences has revealed that distinct HKT transporters that cannot be considered to be either class I or II subfamilies exist in primitive plants such as Physcomitrella patens and Selaginella moellendorffii (Fig. 3). These HKT sequences were found to be closer to the ancestral transporters TRK1 and TRK2 of S. cerevisiae than other HKT transporters (Fig. 3). These facts lead us to hypothesize that the HKT transporters in the primitive plants could possibly be classified into third clade. A protein sequence alignment analysis predicted that these new HKT transporters retain similar p-loop regions as other HKT transporters retaining four glycine residues in the filter positions like in class II transporters (Fig. 4). Detailed characterization of K+/Na+ selectivities of these HKT transporters will give further insights into the relation between the structure and ion selectivity of the plant HKT transporter family.


K+-starvation-induced robust Na+ influx into rice roots has been demonstrated to be OsHKT2;1-dependent (Horie et al. 2007). When high concentrations of NaCl were given to K+-starved oshkt2;1 null mutants and wild-type plants, no remarkable difference between mutants and wild-type plants was found despite the fact that OsHKT2;1 had been shown to mediate transport of a large amount of Na+ in yeast and X. laevis oocytes (Horie et al. 2001, 2007). Interestingly, the addition of a high concentration of NaCl triggered a rapid down-regulation of OsHKT2;1-dependent Na+ influx into K+-starved rice roots showing t1/2 of approximately 1.45 h, suggesting that OsHKT2;1 is inactivated in response to NaCl stress (Horie et al. 2007). These findings strongly suggest that OsHKT2;1-mediated nutritional Na+ influx in K+-starved rice plants is tightly regulated to avoid massive and ultimately toxic Na+-flow into rice plants. Possible posttranslational modification such as protein phosphorylation and mRNA degradation but not protein degradation events were found to be responsible for the rapid inactivation of OsHKT2;1 (Horie et al. 2007). Since Na+ transport via OsHKT2;1 seems to be controlled by cellular signalling pathways that restrict Na+ influx in addition to regulating K+ nutrition, the identification of components and molecular mechanisms will be crucial for dissecting signalling networks for salinity tolerance in rice plants. This could lead also to the identification of yet unknown Na+ influx transporters/channels that mediate toxic Na+ influx into plant roots under salinity stress (Rains & Epstein 1965, 1967; Amtmann et al. 1997; Roberts & Tester 1997; Tyerman et al. 1997; Davenport & Tester 2000; Volkov & Amtmann 2006).

AtHKT1;1 transcript levels were reported to increase in response to high concentrations of Na+ in both shoots and roots, suggesting up-regulation of AtHKT1;1 in xylem parenchyma cells for Na+ removal from xylem vessels (Sunarpi et al. 2005). The xylem parenchyma cell is known to contain abundant membrane systems for controlling secretion/absorption of water and various inorganic ions including K+ and Na+ across the wall of the xylem vessel (Läuchli et al. 1974). It is hard to imagine that vital Na+ unloading function of AtHKT1;1 under salinity stress is regulated merely at the transcriptional level in such a dense set of important regulations. Whether HKT transporters including AtHKT1;1 in Arabidopsis and HKT1;4/1;5 in rice and wheat are subjected to post-translational regulations by cellular components such as interacting regulatory proteins will be an important field to be examined.


In summary, recent findings on physiological functions of specific HKT transporters under salinity stress have revealed that Na+ removal from xylem is a primary and conserved salt tolerance mechanism maintaining high K+/Na+ ratio in leaves in dicot Arabidopsis and important monocot crop plants such as rice and wheat. This is a good indication that biological phenomena found in Arabidopsis plants can be also relevant in crop plants knowing that species specific differences may exist. It should be emphasized that the complete set of physiological functions of the essential class I HKT transporters such as AtHKT1;1, OsHKT1;5, TmHKT1;4 and TmHKT1;5/TaHKT1;5 in Na+ and K+ homeostasis during salinity stress have not been completely defined yet. For instance, whether AtHKT1;1 transporters mediate Na+ loading into phloem to reduce Na+ accumulation in aerial parts under salinity stress is a question to be addressed. In wheat, two independent class I HKT transporters TmHKT1;4 and TmHKT1;5/TaHKT1;5 have been found or suggested to function in protecting leaves from salinity stress and TmHKT1;4 seems to play a specialized role in the xylem of the leaf sheath in addition to the root xylem. Whether such specialized role is assigned to OsHKT1;4 in rice plants, and also whether such a system does not exist in Arabidopsis are interesting questions to be addressed.

Until now, no gene knock out data on HKT transporters in monocot plants are available with only one exception of a class II HKT transporter OsHKT2;1 (Horie et al. 2007). To gain further evidence and draw a complete picture of the HKT transporter-mediated salinity tolerance mechanism, loss-of-function alleles of these genes will be indispensable materials. Substantial information and materials from rice plants, including transposon and T-DNA insertion lines in rice (Hirochika et al. 1996; Jeon et al. 2000; Jeong et al. 2002; Miyao et al. 2003) are potential resources to address the functions of these HKT transporters.

Prolonged drought periods and desertification caused by global climate changes impose serious threats to agriculture and thus life of mankind in the 21st century. Although artificial irrigation could partially compensate for these losses, elevated salt concentrations in irrigated soils become a serious problem. Further research analysing the physiological roles and regulation mechanisms of not only AtHKT1;1 and HKT1;4/1;5 transporters but also the remaining HKT transporters will be necessary to elucidate vital mechanisms controlling Na+ and K+ homeostasis under salinity stress in glycophytes. These findings could contribute to engineer salt resistant plants necessary to cover the increasing demand for food and renewable biomass production in increasing worldwide salinization.


We thank Dr. Julian I. Schroeder (UCSD) for the invitation to write a review on HKT transporters and for comments on the manuscript. We also thank Dr. Maki Katsuhara (RIB, Okayama Univ.) for comments on the manuscript. Research in the author's laboratory was supported by grants from the US Department of Energy (DOE-DE-FG02-03ER15449 to Julian I. Schroeder).