Pioneering studies conducted more than 30 years ago by Epstein and co-workers (see [5]) demonstrated that Na⁺ competes with K⁺ for uptake by plant roots, implying that K⁺ transporters are also the gates for Na⁺ entry. Although the first K⁺ transporters were cloned a decade ago, however, the identity of the major pathway(s) for Na⁺ uptake has remained elusive. Transporters of the HAK-KT-KUP family (which includes the high-affinity K⁺-uptake transporter HAK, K⁺ transporter KT, and K⁺-uptake transporter KUP) are highly selective for K⁺ (K_m 10-50 μM) but can also transport Na⁺, albeit with much lower affinity. HAK-KT-KUP proteins are thought to be K⁺/H⁺ symporters (moving both molecules in the same direction), permitting a concentration of K⁺ against its electrochemical gradient, to ensure K⁺ nutrition [1]. Another family of transporters, the HKT-TRK proteins (which includes the high-affinity K⁺ transporter HKT and its homologs), shows mixed ion selectivity when expressed in heterologous systems. The wheat TaHKT1, for example, functions as a Na⁺/K⁺ symporter at micromolar Na⁺ concentrations but as a Na⁺ uniporter (transporting only this single type of ion) at millimolar Na⁺ concentrations [6], whereas the Arabidopsis AtHKT1 transports only Na⁺ [7]. Rice has both types of transporter: OsHKT1 is a Na⁺ transporter like AtHKT1 but OsHKT2 behaves as a symporter or uniporter as does TaHKT1 [8]. Transcripts of the OsHKT genes accumulated under low K⁺ concentrations and diminish in high external Na⁺ [8]; together with the nature of the transporters, these data suggest that HKT proteins might mediate substantial Na⁺ uptake [1,8].

Genetic evidence supporting a significant role for HKT proteins in Na⁺ uptake has been provided recently by Rus et al. [9]. The Arabidopsis SOS3 (salt overly-sensitive 3) gene product is a Ca²⁺-binding protein, deficiency in which elicits Na⁺ sensitivity and an inability to grow at low external K⁺ concentrations [10]. Theoretically, hypersensitivity of sos3 mutant plants to NaCl could arise either from increased Na⁺ entry or from reduced K⁺ uptake. Searching for mutations in other genes that suppressed the salt-sensitive phenotype of sos3 mutants, Rus et al. [9] isolated two independent loss-of-function mutants in AtHKT1, the only gene of the HKT-TRK family in Arabidopsis. The hkt1 mutation dramatically reduced the net Na⁺ content of hkt1 sos3 double-mutant plants under a saline regime, to levels even lower than those of wild-type plants. Interestingly, the hkt1 mutation also suppressed, instead of exacerbating, the low-K⁺ phenotype of sos3 plants. In fact, the K⁺ content of hkt1 sos3 double-mutant plants was higher than that of wild-type plants. Together, these results indicate that AtHKT1 is not a relevant K⁺-uptake system and provide evidence of substantial Na⁺ influx through AtHKT1. Because AtHKT1 is preferentially expressed in roots, it may mediate physiological Na⁺ uptake in conditions of poor K⁺ availability, in which Na⁺ could partially substitute for K⁺, for instance as osmoticum (a solute contributing to osmotic pressure) in the vacuole [1].

The recent study by Rus et al. [9] leaves open the question of whether or not the cation selectivity of AtHKT1 is regulated by a SOS3-dependent signaling pathway. Fungal TRK K⁺ transport proteins, which are structurally related to HKTs, modulate their Na⁺/K⁺ selectivity according to the ionic environment and the K⁺ status of the cell [1]. Because the hkt1 mutation was not segregated from the sos3 mutant background to assess the phenotype of a single hkt1 mutant [9], it is unclear whether AtHKT1 is an unconditional Na⁺ transporter or if its Na⁺/K⁺ selectivity is distorted in the sos3 mutant background, allowing unrestricted Na⁺ entry. Matters are further complicated by the likelihood that SOS3 regulates Na⁺ efflux [11]. Thus, it remains possible that Na⁺ uptake through AtHKT1 becomes detrimental to the plant only when other relevant Na⁺ fluxes are compromised by the sos3 mutation (Figure 1).

An interesting determination made by Rus et al. [9] is that the capacity of the hkt1 mutation to suppress the Na⁺ sensitivity of sos3 mutant plants disappeared in medium with low concentrations of Ca²⁺ (0.15 mM). This suggests the operation of an alternative Na⁺-influx system, different from AtHKT1, that is hampered by high Ca²⁺ (2 mM) but becomes the prevalent pathway for Na⁺ entry at low external Ca²⁺ concentrations. Electrophysiological studies have confirmed the presence of non-selective voltage-independent cation channels (known as NSCCs or VICs) in Arabidopsis root cells [2,12,13]. These channels lack significant selectivity between monovalent cations, their affinity for K⁺ being only slightly higher than their Na⁺ affinity (P_K/P_Na = 1.49), and they are strongly blocked by external Ca²⁺ at low concentrations, with half-maximal inhibition at 0.1 mM. The probability of the NSCC/VIC channel being open decreased 10-15-fold with a 10-fold increase in external Ca²⁺ concentration [13]. On the basis of similarities between NSCC/VIC transport behavior and Na⁺ uptake by intact roots, NSCC/VICs are thought to contribute substantially to total Na⁺ influx in saline conditions [2,12,13].

NSCC/VICs appear to be widespread among plants but their physiological function and molecular identity have remained obscure. Recently, Maathuis and Sanders [12] have shown that some NSCC/VIC currents are rapidly and reversibly deactivated by cyclic nucleotides (cAMP and cGMP) on the cytosolic side of root cell membranes, suggesting a direct interaction of cyclic nucleotides with the channel. There are 20 candidate genes in the Arabidopsis genome that encode putative cyclic-nucleotide-gated channels (CNGCs), having hallmark cyclic-nucleotide- and calmodulin-binding domains [14]. In keeping with the idea that a nucleotide-sensitive NSCC/VIC forms a major pathway for Na⁺ influx, supplemental cAMP and cGMP were found to improve the tolerance of Arabidopsis seedlings to NaCl but not to equivalent concentrations (equiosmolar amounts) of sorbitol [12]. Salt tolerance induced by cyclic nucleotides correlated with less accumulation of Na⁺ after 5 days exposure to NaCl and reduced uptake of radioactively labeled ²²Na⁺ in short-term experiments [12]. It appears that only a subtype of NSCC/VICs are CNGCs, however, because nucleotide sensitivity was found in only a fraction of cells exhibiting NSCC/VIC currents [12]. Moreover, the Arabidopsis polypepetide CNGC2, which has homology to animal ion channels, is activated by cyclic nucleotides, rather than inactivated, and displays different transport properties from NSCC/VICs [15]. A mutant screen for improved tolerance to Na⁺ at low external Ca²⁺, perhaps in the sos3 mutant background, could confirm whether or not NSCC/VICs contribute to the Ca²⁺-sensitive pathway for Na⁺ influx unmasked by the hkt1 mutation, and could establish whether these channels belong to the CNGC family [9].

Ion compartmentation is one successful strategy used by plant cells to cope with saline stress, because it couples growth by cell expansion, mainly due to an enlarging vacuole, with ion sequestration and detoxification from the cytosol. Accrued ions provide energetically cheap osmoticum that helps balance the plant's osmotic potential with that of the hypertonic soil solution, thereby alleviating the water deficit that is inherent to plants in a saline environment. The Na⁺/H⁺ antiporters (which transport the two ions in opposite directions) at the tonoplast (vacuolar membrane), mediate the transport of Na⁺ into the vacuoles, with the energy for this coming from the electrochemical gradient of protons generated by the vacuolar H⁺-ATPase and H⁺-pyrophosphatase [4].

Gaxiola et al. [16] have reported inducing increased salt- and drought-tolerance in Arabidopsis plants by moderate overexpression of the endogenous Arabidopsis vacuolar H⁺-pyrophosphatase, AVP1. The stress tolerance of plants exposed to salinity was correlated with more negative solute potential (that is, greater concentration of total osmolytes), which permitted better water retention by desiccating plants, and with larger pools of Na⁺ and K⁺ (a 20-40% increase in concentration). Presumably, Na⁺ accumulation occurs because of a greater H⁺ gradient and enhanced Na⁺/H⁺ exchange across the tonoplast, although the activity of the Na⁺/H⁺ antiporter in vesicles of this membrane was not experimentally determined. Accumulation of K⁺ could also result from the activity of the vacuolar antiporters, one of which, NHX1 (Na⁺/H⁺ exchanger 1), has been shown to transport Na⁺ and K⁺ [17]. Consistent with this interpretation, overexpression of the vacuolar cation antiporter NHX1 also increased the salt tolerance of Arabidopsis, tomato and canola plants [18,19,20]. Vacuoles isolated from leaves of transgenic Arabidopsis sustained higher Na⁺/H⁺ exchange rates, which directed greater Na⁺ accumulation into aerial tissues than in wild-type plants. Drought tolerance of NHX1 transgenics was not examined, however, leaving the unanswered question of whether stress tolerance imparted by AVP1 can be explained solely by NHX1 upregulation.

It would be informative to compare AVP1 and NHX1 plants side by side and to test whether plants overexpressing both proteins are more stress-tolerant than either of the single-gene transgenics. Similarly, given that plants engineered to amass more organic solutes, such as proline or polyalcohols, are also stress-tolerant [3], super-plants combining a greater capacity for the accumulation of ions in the vacuole and for accumulation of compatible solutes in the cytosol hold great promise for substantially improved fitness in adverse environments. By looking in a variety of plant species, genomic tools promise to assist in furthering our understanding of the molecular basis for salt-tolerance, and our ability to engineer it, in the future.