Genes and salt tolerance: bringing them together




  • Summary 1

  • I. Introduction 1
  • II. Physiological mechanisms of salt tolerance 2
  • III. Candidate genes for salt tolerance and results of transformation experiments 6
  • IV. Gene activity expected in roots, leaves and growing tissues of plants exposed to salinity, and results of gene expression studies 11
  • V. Conclusions 16
  • Acknowledgements 16

  • References 16


Salinity tolerance comes from genes that limit the rate of salt uptake from the soil and the transport of salt throughout the plant, adjust the ionic and osmotic balance of cells in roots and shoots, and regulate leaf development and the onset of senescence. This review lists some candidate genes for salinity tolerance, and draws together hypotheses about the functions of these genes and the specific tissues in which they might operate. Little has been revealed by gene expression studies so far, perhaps because the studies are not tissue-specific, and because the treatments are often traumatic and unnatural. Suggestions are made to increase the value of molecular studies in identifying genes that are important for salinity tolerance.

I. Introduction

Over 800 million hectares of land throughout the world are salt-affected, either by salinity (397 million ha) or the associated condition of sodicity (434 million ha) (FAO, 2005). This is over 6% of the world's total land area. Most of this salinity, and all of the sodicity, is natural. However, a significant proportion of recently cultivated agricultural land has become saline because of land clearing or irrigation. Of the 1500 million ha of land farmed by dryland agriculture, 32 million (2%) are affected by secondary salinity to varying degrees. Of the current 230 million ha of irrigated land, 45 million ha are salt-affected (20%) (FAO, 2005). Irrigated land is only 15% of total cultivated land, but as irrigated land has at least twice the productivity of rainfed land, it produces one-third of the world's food.

Definitions of salinity and sodicity are given in Table 1. The types of salinity and their causes, and the formation of the associated sodicity, are described in Appendix S1, available online as supplementary material. Natural or primary salinity is more widespread than first realized (Rengasamy, 2002), while secondary salinity, caused by land clearing or irrigation, continues to grow. Hence increased salt tolerance of crops and horticultural species is needed to sustain increases in food production in many regions in the world. Increased salt tolerance of perennial species used for fodder or fuel production is a key component in reducing the spread of secondary salinity, while increased salt tolerance of crops will directly improve production in soils with primary salinity (Appendix S1).

Table 1. Definition of saline and sodic soils, classed according to the USDA Salinity Laboratory (USSL, 2005)
TermDefinitionDescriptionEffect on plant growthComments
  1. Causes of salinity, and a fuller description of the origins and nature of salinity and sodicity, are given in Appendix S1.

SalinitySaline soils have a high concentration of soluble salts. They are classed as saline when the ECe is ≥ 4 dS m−1.This definition of salinity derives from the ECe that would reduce yield of most crops. However, many crops are affected by an ECe < 4 dS m−1.Osmotic and salt-specific components inhibit root and shoot growthECe is the electrical conductivity of the saturated paste extract, and reflects the concentration of salts in saturated soil. A conductivity of 4 dS m−1 is equivalent to 40 mm NaCl.
SodicitySodic soils have a low concentration of soluble salts, but a high exchangeable Na+ percentage (ESP). They are classed as sodic when the ESP is ≥ 15.This definition of sodicity derives from the ESP that causes degradation of the structure of clay soils, caused by Na+ displacing divalent cations bound to negative charges on the clay particles.Poor soil structure inhibits root growthAt high ESP, the clay particles separate. The soil drains poorly and becomes waterlogged when wet. It also becomes very hard when dry.
AlkalinityAlkaline soils are a type of sodic soil with a high pH. They are defined as having an ESP ≥ 15 with a pH of 8.5–10.The high pH is caused by carbonate salts in parent material.High pH affects nutrient uptake 

Increased salt tolerance requires new genetic sources of this tolerance, and more efficient techniques for identifying salt-tolerant germplasm. Powerful new molecular tools for manipulating genetic resources are becoming available, but the applications of the new technologies are not yet fully utilized to introduce new genes for tolerance into current cultivars.

This review outlines the major adaptive mechanisms of salt tolerance at the physiological and molecular levels. The main mechanisms for salt tolerance are to minimize the salt taken up by the roots, and to partition it at the tissue and cellular level so that it does not build up to toxic concentrations in the cytosol of transpiring leaves. Candidate genes are listed for ion transport, osmoprotection, and making plants grow more quickly in saline soil. Studies on gene expression in roots and leaves are reviewed, and suggestions made for increasing the value of these studies in improving plant performance on saline soil.

II. Physiological mechanisms of salt tolerance

1. Osmotic vs salt-specific effects on growth: the two phases of the growth response

Salt in soil water inhibits plant growth for two reasons. First, it reduces the plant's ability to take up water, and this leads to slower growth. This is the osmotic or water-deficit effect of salinity. Second, it may enter the transpiration stream and eventually injure cells in the transpiring leaves, further reducing growth. This is the salt-specific or ion-excess effect of salinity.

The two effects give rise to a two-phase growth response to salinity (Fig. 1). The diagram shows the growth response to salt that is added gradually.

Figure 1.

Schematic illustration of the two-phase growth response to salinity for genotypes that differ in the rate at which salt reaches toxic levels in leaves. With annual species, the timescale is d or wk, depending on species and salinity level. With perennial species, the timescale is months or yr. During phase 1, growth of both genotypes is reduced because of the osmotic effect of the saline solution outside the roots. During phase 2, leaves in the more sensitive genotype die and reduce the photosynthetic capacity of the plant. This exerts an additional effect on growth. Adapted from Munns (1993). If salt is added in one step, the growth rate plummets to zero or below and takes 1–24 h to recover to the new steady rate, depending on the degree of the osmotic shock (Munns, 2002).

Phase 1  The first phase of the growth response results from the effect of salt outside the plant. The salt in the soil solution reduces leaf growth and, to a lesser extent, root growth (Munns, 1993). The cellular and metabolic processes involved are in common to drought-affected plants. Neither Na+ nor Cl builds up in growing tissues at concentrations that inhibit growth: meristematic tissues are fed largely in the phloem, from which salt is effectively excluded; and rapidly elongating cells can accommodate the salt that arrives in the xylem within their expanding vacuoles.

Phase 2  The second phase of the growth response results from the toxic effect of salt inside the plant. The salt taken up by the plant concentrates in old leaves: continued transport into transpiring leaves over a long period eventually results in very high Na+ and Cl concentrations, and the leaves die. The cause of injury is probably the salt load exceeding the ability of cells to compartmentalize salts in the vacuole. Salts would then build up rapidly in the cytoplasm and inhibit enzyme activity. Alternatively, they might build up in the cell walls and dehydrate the cell.

The rate of leaf death is crucial for survival of the plant. If new leaves are continually produced at a rate greater than that at which old leaves die, there will be enough photosynthesizing leaves for the plant to produce flowers and seeds, although in reduced numbers. However, if old leaves die more quickly than new ones develop, the plant may not survive to produce seed. For an annual plant there is a race against time to initiate flowers and form seeds, while the leaf area is still adequate to supply the necessary photosynthate. For perennial species, there is an opportunity to enter a state akin to dormancy and thus survive the stress.

In summary, the initial growth reduction is caused by the osmotic effect of salt outside the roots, and the subsequent growth reduction is caused by the inability to prevent salt from reaching toxic levels in transpiring leaves.

2. Mechanisms controlling leaf and root growth (phase 1 response)

The mechanisms controlling this phase of the growth response are not specific to salinity; they are caused by factors associated with water stress. This is supported by evidence that Na+ and Cl are below toxic concentrations in the growing cells themselves, in leaves (Hu & Schmidhalter, 1998; Fricke, 2004) and roots (Jeschke, 1984; Jeschke et al., 1986).

Whether water status, hormonal regulation or supply of photosynthate exerts the dominant control over growth of plants in dry or saline soil is an issue that has been hotly debated. Over the timescale of days, there is much evidence to suggest that hormonal signals, rather than water relations, are controlling growth in saline soils. The evidence for this is that leaf expansion in saline soil on the timescale of days does not respond to an increase in leaf water status (summarized by Munns, 2002). These experiments indicated that there are chemical signals coming from roots in dry or saline soil that reduce leaf growth. These are commonly referred to as root signals. Abscisic acid is the obvious candidate for this signal, as it is found in xylem sap and increases after drought and salinity stress (reviewed by Munns & Cramer, 1996). However, there is still no conclusive proof that abscisic acid (ABA) is the only signal from the roots (reviewed by Dodd, 2005). Moreover, the origin of the ABA in the xylem sap is not known, for it moves readily in the phloem and recirculates from leaves to roots (reviewed by Munns & Cramer, 1996).

Hormonal control of cell division and differentiation is clear from the appearance of leaves, which are smaller in area but often thicker, indicating that cell size and shape has changed (James et al., 2002). Leaves from salt-treated plants have a higher weight : area ratio, which means that their transpiration efficiency is higher (more carbon fixed per water lost), a feature that is common in plants adapted to dry and to saline soil.

Hormonal control of cell division and elongation is also evident in roots. Several studies have shown that salinity has differential effects on root elongation rates and lateral root initiation (Rubinigg et al., 2004).

3. Mechanisms controlling the salt-specific effects of salinity (phase 2 response)

Mechanisms for tolerance of the salt-specific effects of salinity are of two main types: those minimizing the entry of salt into the plant; and those minimizing the concentration of salt in the cytoplasm. Root cytosolic Na+ concentrations are probably in the order of 10–30 mm (summarized by Tester & Davenport, 2003). Leaf Na+ cytosolic concentrations are unknown, but are considered to be much less than 100 mm (Wyn Jones & Gorham, 2002). The concentration at which Cl becomes toxic is even less defined.

Low salt accumulation in leaves: the mechanism known as salt exclusion  Roots must exclude most of the Na+ and Cl dissolved in the soil solution, or the salt in the shoot will gradually build up with time to toxic levels. Plants transpire about 50 times more water than they retain in their leaves (Munns, 2005). If a plant lets in only 1/50 of the salt in the soil solution (i.e. it excludes 98%), the concentration of salt in the shoot as a whole would never increase over that in the soil, and the plant could survive indefinitely in saline soil (Munns, 2005).

Most plants do exclude about 98% of the salt in the soil solution, allowing only 2% to be transported in the xylem to the shoots. Table 2 shows the differences between cereal genotypes with contrasting rates of Na+ uptake when grown in 50 mm NaCl. Bread wheat excludes > 98% of the Na+ in the soil solution, and the concentrations do not build up in leaves to > 50 mm (Husain et al., 2004). Barley, on the other hand, excludes < 98% of the Na+ in the soil solution, and concentrations can reach very high levels, up to 500 mm (Rawson et al., 1988). Cl exclusion also differs between species (Table 2).

Table 2. Na+ and Cl concentrations in the xylem, and percentage exclusion of plants grown at 50 mm NaCl
Xylem Na+ (mm)Exclusion (%)Xylem Cl (mm)Exclusion (%)
  • Genotypes represent a wide range of Na+ exclusion, with bread wheat typically having low rates of Na+ transport to shoots, and barley having high rates.

  • *

    Calculated from rates of Na+ transport from root to shoot, and from transpiration rates.

  • Sap collected from the xylem-feeding insect Philaenus spumarius.

  • Sap collected by root pressurization to induce flow equivalent to transpiration rate.

  • §

    Line 149 is a novel Na+-excluding durum wheat landrace (Munns et al., 2003).

  • R. Munns, S. Husain and R.A. James (unpublished data).

Bread wheatJanz0.3991.597Calculated*Unpublished
Chinese Spring1.198Insect feeding Watson et al. (2001)
Kharchia-651.497Calculated* Garcia et al. (1997)
Punjab-851.697Calculated* Garcia et al. (1997)
Durum wheatLine 149§0.6991.298Calculated*Unpublished
Langdon3.094Insect feeding Watson et al. (2001)
RiceIR362.894Calculated* Garcia et al. (1997)
BarleyClipper3.2944.791Collected sap Munns (1985)

Export from leaves in the phloem could conceivably help to maintain low salt concentrations. However there appears to be relatively little retranslocation of salt from leaves, in relation to the import in the transpiration stream. This can be seen in the continued presence of salt in leaves long after the salt around the roots is removed. Measurements of ions in phloem sap have indicated that the more salt-tolerant species exclude Na+ and Cl from the phloem to a large extent, while less tolerant ones do not have the same restriction (reviewed by Munns et al., 1986, 1988). Estimates of xylem and phloem fluxes indicated that, in barley, phloem export from a leaf was only ≈ 10% of the import in the xylem (Munns et al., 1986). Exclusion of salt from the phloem ensures that salt is not redirected to growing tissues of the shoot. Although salt that is loaded into the phloem in lower leaves may be translocated down to the roots, the phloem from younger leaves may go up to the meristematic and elongating tissues in the shoot. As shown by 14C urea-feeding studies, lower leaves feed carbon to the root, upper leaves feed the shoot apex, and mid-position leaves feed both shoot apex and root (Layzell et al., 1981).

There is a strong correlation between salt exclusion and salt tolerance in many species (reviewed by Greenway & Munns, 1980; Läuchli, 1984; Yeo & Flowers, 1986; Munns & James, 2003; Tester & Davenport, 2003). Figure 2 shows the relationship between Na+ exclusion and salt tolerance for durum wheat. The following section shows how natural variation in Na+ exclusion can be used to introduce new sources of salt tolerance into commercial cultivars of wheat.

Figure 2.

Relationship between salinity tolerance (shoot dry biomass in salt treatment as percentage of biomass in control treatment) and leaf Na+ concentration in durum wheat genotypes (Triticum turgidum ssp. durum). NaCl (150 mm) was added gradually before leaf 3 appeared. Na+ concentrations were measured on leaf 3 after 10 d, and shoot biomass after 24 d. Values are expressed as percentage of shoot biomass in control conditions. All values are means (n = 5). Data from Munns & James (2003).

Genetic improvement in salt tolerance of durum wheat using the trait for sodium exclusion  Cultivated durum (pasta) wheat (Triticum turgidum ssp. durum) is more salt-sensitive than bread wheat (Triticum aestivum), a feature that restricts its production in farms with sodic or saline soils. To increase the salt tolerance of durum wheat, we proposed that lowering its sodium-uptake rate would make it more tolerant of saline soil, based on the pioneering work of John Gorham and colleagues (Gorham et al., 1990). In order to introduce the trait of Na+ exclusion into current durum wheat, we searched for genetic variation in salt tolerance across a wide range of ancient durum-related accessions and landraces representing five Triticum turgidum subspecies. Selections were screened nondestructively for low Na+ concentration in leaves and the associated enhanced K+/Na+ discrimination (Munns et al., 2000). Wide genetic variation was found in Na+ accumulation and K+/Na+ discrimination (Fig. 3).

Figure 3.

Sodium concentration after 10 d in leaves of bread and durum wheat genotypes grown in 150 mm NaCl with supplemental Ca2+. Tamaroi and Wollaroi are Australian durum wheats; Janz is a bread wheat. Line 149 is the best of the tetraploids screened for low Na+ by Munns et al. (2000). Line 141 is one of the highest Na+ accumulators, and has Na+ concentrations similar to barley.

Two durum wheat genotypes with contrasting rates of Na+ transport to leaves were used to assess the effects of the Na+ exclusion trait on preventing leaf injury and enhancing yield (Husain et al., 2003). Older leaves of the high-Na+ lines lost chlorophyll more rapidly and died earlier than the low-Na+ lines. The low-Na+ trait improved yield by > 20% in saline soil in glasshouse trials at moderate salinity (Husain et al., 2003). However, yield was not improved at high salinity (Fig. 4). This indicates that traits other than Na+ exclusion are important at high salinity, where the osmotic effect of the NaCl outweighs its salt-specific effect on growth and yield.

Figure 4.

Effect of three different salinity levels (1, 75 and 150 mm NaCl) on grain yield (g per plant) of low and high Na+ genotypes. Shaded bars, low Na+ genotype (Line 149); open bars, high Na+ genotype (Line 141). Asterisks indicate significant differences from the control and from each other at P = 0.05. Data from Husain et al. (2003).

Genetic analysis of the segregating populations developed from crosses between genotypes with low and high Na+ accumulation indicated just two major genes that interacted, in a classically Mendelian way, to give very low rates of Na+ accumulation (Munns et al., 2003). The identity of the two genes is not yet known. It is likely that the effect of one gene is on the loading of Na+ in the xylem in the roots, while the other is on retrieving Na+ from the xylem in the upper part of the root system and the lower part of the leaves (Davenport et al., 2005). These genes would work together to produce very low Na+ concentrations in the leaf blades, and so fit the Mendelian analysis of two dominant and interacting genes.

A locus for the low-Na+ trait was mapped to the long arm of chromosome 2A using a quantitative trait locus (QTL) approach (Lindsay et al., 2004), using amplified fragment length polymorphism (AFLP), restriction fragment length polymorphism (RFLP) and microsatellite markers. This approach identified several markers linked to a gene at a QTL designated Nax1 (Na+ exclusion). The microsatellite markers identified were also associated with the low-Na+ trait in four other populations with different genetic backgrounds, and are being used to select low-Na+ progeny in a durum-breeding programme (Lindsay et al., 2004).

Other traits for salt tolerance  Some plants have ways of partitioning the salt arriving in the shoot, either by retaining it in the leaf base or stem, or directing salt away from younger leaves and towards older ones. This is particularly important for species that cannot exclude more than 98% of the salt from the transpiration stream. Other traits are needed to improve tolerance to the osmotic effect of the salt outside the roots which, as mentioned above, can have a greater effect than the salt-specific component on growth and yield. Such traits include water-use efficiency, osmotic adjustment, and morphological or developmental patterns that conserve water and advance the flowering date. These agronomic traits are discussed by Colmer et al. (2005) for improvements in salt tolerance of wheat and barley.

Species that cannot exclude 98% of the salt from the transpiration stream (e.g. barley; Table 2) must also be able to compartmentalize the salt in vacuoles, thereby protecting the cytoplasm from ion toxicity and avoiding buildup in the cell wall which would cause dehydration (Flowers & Yeo, 1986). Otherwise, in older leaves the salt concentration would eventually become high enough to kill the cells. If Na+ and Cl are sequestered in the vacuole of the cell, K+ and organic solutes should accumulate in the cytoplasm and organelles to balance the osmotic pressure of the Na+ and Cl in the vacuole. The genes responsible for this are considered in the next section.

III. Candidate genes for salt tolerance and results of transformation experiments

Genes that could increase salt tolerance fall into three main functional groups: (1) those that control salt uptake and transport; (2) those that have an osmotic or protective function; and (3) those that could make a plant grow more quickly in saline soil.

1. Genes that control salt uptake and transport

This section focuses on the transport of Na+ rather than Cl. The control of Na+ uptake is probably more difficult and energy-demanding than Cl because plant cells have a negative electrical potential (about −180 mV). The prevention of Na+ uptake from the apoplast is more demanding in terms of ion selectivity and energy costs than the prevention of Cl uptake: Cl would not enter a cell passively unless the cytosolic Cl concentration was very low (Appendix S2, available online as supplementary material). Furthermore, genetic variation in salt tolerance of the whole plant correlates with the degree to which the plant can limit the rate of Na+ transport to leaves. However, in some dicotyledonous species salinity tolerance correlates with reductions in Cl accumulation in the leaves. For these species, Cl uptake and transport mechanisms may be particularly relevant. Reviews by Tyerman & Skerrett (1999) and White & Broadley (2001) cover Cl transport mechanisms.

Membrane proteins controlling Na+ transport  Proteins controlling the uptake of Na+ from the soil and the subsequent transport of Na+ within the plant are embedded in the lipid bilayer of the membrane, which otherwise is relatively impermeable to ions. Na+ probably enters cells through nonselective cation channels, and at high salinity, possibly through K+ channels or transporters that are incompletely selective for K+ (Amtmann & Sanders, 1999; Schachtman & Liu, 1999; Tyerman & Skerrett, 1999; Tester & Davenport, 2003). These are summarized in Table 3.

Table 3. Cloned genes with likely relevance for controlling K+ or Na+ uptake by roots and transport within the plant, and which are candidates for overexpression studies
Type of transporterGene familyCandidate genes for salt toleranceProbable function in higher plants
  1. Information from reviews by Mäser et al. (2002); Reinhold & Guy (2002); Demidchik et al. (2002); Véry & Sentenac (2003); and transcriptome analysis by Maathuis et al. (2003).

K+ channelShaker type (single pore, tetramer) inward channel AKT1, AKT2, KAT1 AKT1 (Arabidopsis K+transporter) is an inward-rectifying K+ channel expressed in roots. It is highly selective for K+ over Na+ (PNa/PK = 0.05). However, at high salinities this channel could transport Na+. AKT2 and KAT1 are related. These are expressed in leaf phloem tissue and guard cells, respectively, in Arabidopsis, but may function in other cell types in other species.
K+ channelShaker type, outward channel SKOR SKOR (stelar K+outward rectifier) is important in maintaining K+ homeostasis in both roots and shoots. SKOR mutants have lower K+ concentrations in shoots but not roots, indicating that SKOR influences xylem loading of K+. SKOR is probably located on the plasma membrane.
K+ channel KCO family (two pore channel) KCO1 KCO1 (K+channel outward) rectifier is expressed in leaf cells, probably on the tonoplast.
Nonselective cation channel CNGC and GLR families CNGC1–20, GLR1–20Some members of the CNGCs (cyclic nucleotide-gated channels) and GLRs (glutamate receptors) families are predicted to have a similar permeability to Na+ and K+, and to be regulated by Ca2+ (Demidchik et al., 2002).
K+ antiporterK/H antiporter KEA or CPA (CHA) familyK+ antiporters may be important in K+ homeostasis by loading K+ into vacuoles. KEA (K+exchange antiporter) is present in the plant genome, but its function is unknown. It is possible that these could carry Na+, just as Na+/H+ antiporters can carry K+.
K+ transporter KUP/HAK/KT family HAK1–10, KUP1–4A very large family of K+ transporters, similar to the K+ uptake family in bacteria and high-affinity K+ transporters in fungi. There are many variants in all higher plants, and they are likely to be very important in control of K+ homeostasis. Their selectivity for K+ over Na+ in higher plants is not known.
K+ transporter HKT family HKT1 K+ starvation induces HKT1 (high-affinity K+transporter) expression in wheat, indicating that it functions in high-affinity K+ uptake, but it also transports Na+. In Arabidopsis it is likely that AtHKT1 is important in regulating Na+ and K+ homeostasis, as mutations that disrupt its function alter the transport of Na+ from root to shoot, and the K+/Na+ ratio in the root (Rus et al., 2004).
Cation antiporter CHX family CHX10, 15 Cation hydrogen exchangers such as CHX10 and CHX15 regulate K+ uptake by vacuoles. They may carry Na+, and their expression is down- regulated under salt stress (Maathuis et al., 2003).
Na+ antiporter NHX family NHX1 AtNHX1 (Na+/H+exchanger) is an Na+/H+ antiporter on the tonoplast membrane. It is expressed in roots and leaves, and selectively transports Na+ into the vacuole, as well as K+ in nonsaline conditions (Apse et al., 2003).
NHX2–5 AtNHX2–5 are expressed in specific cell types, transport Na+ or K+ into the vacuole, and have a likely role in K+ or pH regulation.
SOS1 SOS1 (AtNHX7) is an Na+/H+ antiporter on the plasma membrane. SOS1 (salt overly sensitive) is expressed in root cells (Shi et al., 2002). SOS1 would efflux Na+ from cells and may be important in Na+ extrusion from roots into the external medium.
Proton pumpAHA P-type H ± ATPase AHA2 H+ transport across plasma membrane.
Proton pumpH ± PPase AVP1 H+ transport across tonoplast.

Passive ion transport occurs through channels, which are membrane proteins with ion-selective pores that allow ion movement down an electrochemical gradient. They are usually highly selective for a specific ion, but there is a class of nonselective cation channels that transport Na+, K+, Ca2+ and NH4+ (Demidchik et al., 2002). Some nonchannel transporters, such as HKT1, can allow passive transport under certain conditions (Tester & Davenport, 2003).

Active ion transport occurs through symporters and antiporters that can transport ions against an electrochemical potential gradient. Transport is driven by the electrochemical potential difference of a coupled solute, usually H+. Transporters undergo a cycle of conformational changes while transporting ions, the rate of transport being much lower than that of channels. Most are highly selective for an ion other than Na+, but the NHX family of antiporters (Na+/H+ exchangers) are selective for Na+.

Proton pumps are essential to provide the electrical potential difference that drives Na+ symporters or antiporters. Active Na+ uptake into the cell across the plasma membrane is driven by P-type H+-ATPases, which hydrolyse ATP to pump H+ into the cell wall (Reinhold & Guy, 2002); active ion transport into the vacuole is driven by V-type H+-ATPases, which hydrolyse ATP to pump H+ into the vacuole (Binzel & Ratajczak, 2002). Also important is the H+-PPiase on the tonoplast, which hydrolyses pyrophosphate to pump H+ into the vacuole (Gaxiola et al., 2001; Binzel & Ratajczak, 2002).

There are no classical Na+ pumps in higher plants. In fungi and mosses there is an Na+ pump, ENA1, which hydrolyses ATP to pump Na+ out of the cell (Benito & Rodriguez-Navarro, 2003). Also in yeast are ‘HAL’ genes (Serrano et al., 1999), with uncertain function other than the control of K+ or Na+ transport. Many of these do not have orthologues in higher plants. The expression of these in plants may introduce new mechanisms for control of Na+ or K+ transport: for example Rus et al. (2001) have shown that HAL1, which controls K+ transport, improves salt tolerance of tomato.

Na+ that enters a cell will have one of three fates (presuming it cannot be retained in the cytoplasm, where it would be toxic). Na+ can move symplastically into an adjacent cell via plasmadesmata; it can be effluxed back into the cell wall; or it can be ‘compartmentalized’ (i.e. transported into the vacuole). Efflux occurs through Na+/H+ antiporters on the plasma membrane such as SOS1 (Zhu, 2003). Compartmentalization occurs through vacuolar Na+/H+ antiporters such as NHX1 (Blumwald et al., 2000).

The processes that maintain low Na+ in organelles such as chloroplasts and mitochondria are less well known, although the maintenance of pH in organelles, as well as their cation composition, is important. A cation/proton antiporter, CHX23, which is specific to the chloroplast envelope membrane, has been identified by Song et al. (2004). Such proteins are probably essential for pH control of chloroplasts.

Candidate genes for improving control of Na+ uptake and transport  Many genes are important in maintaining K+ or Na+ homeostasis in higher plants, and could be considered candidates for genetic manipulation. These are listed in Table 3. The K+ channels and transporters may regulate Na+ transport – either directly, because they may be incompletely selective for K+ and transport Na+ when presented with a high Na+ concentration or a high Na+/K+ ratio; or indirectly, because they may buffer the cell against Na+ uptake by maintaining rigorous K+ homeostasis. Single amino acid changes in the pore region of a channel or transporter can change the selectivity for cations, so a single base-pair mutation can have a large effect.

Many of these transporters have a large number of variants, both within and between species. In cereals there is a particularly large redundancy in the HKT and the HAK/KUP/KT families (Damien Platten, pers. comm.). These transporters control K+ uptake and K+/Na+ selectivity. The function of the redundancy is unclear. Are some variants active and others silent? Are some expressed under different Na+ or K+ concentrations, or in different tissues, or at different cell developmental stages? The function of individual genes in such large gene families will be difficult to fully understand.

Not listed in Table 3 are regulatory genes such as kinases and phosphatases that provide post-transcriptional regulation of ion transporters. The SOS pathway appears to be regulated in this way (Zhu, 2003). It is activated by a sodium-induced rise in Ca2+ which binds SOS3, leading to a cascade which activates the kinase SOS2, which phosphorylates and thereby activates SOS1 (Zhu, 2003). Other regulatory pathways are summarized by Tester & Davenport (2003, p. 512).

There is little unequivocal information on the cell or tissue specificity of these cation transporters, most information being based on in situ localization. Microscopic images can be difficult to interpret, and as reporter genes are expressed only in the cytoplasm, the intensity of the expression in different cell types may reflect differences in cytoplasmic volumes in different cell types. For example, the visual expression of SOS1 promoter-GUS expression pattern in root epidermal cells near the root tip only, and in the cells lining the xylem throughout the plant (Shi et al., 2003), may be influenced by the higher cytoplasmic volumes in these cell types. Furthermore, expression patterns from gene constructs may vary from one species to another, and it is difficult to know whether this is caused by real differences in expression in heterologous systems, or by different tissue anatomy.

How to evaluate the function of cation transporters in salt tolerance  It is not easy to measure the activity of individual ion transporters, particularly in specific cell types. Measurement of ion concentrations at a tissue or organ level is easy, and direct measurement or calculation of ion concentrations in xylem sap is simple although tedious (Table 1), but suprisingly few measurements are made of ion concentrations in tissues or xylem sap. More commonly, salt tolerance of the whole plant is measured, as either plant biomass or plant survival.

In some studies, rather than growing plants for a long time at moderate salinity, the severity of the stress is increased until the more sensitive plants die. This has the advantage of showing up individuals with high tolerance of the stress, but can wipe out the chance of finding the mechanism, as the experimental control (e.g. the wild-type genotype) may have died.

Results of transformation studies  Transformation experiments with the greatest successes have come from manipulating tonoplast Na+ transport. An antiporter cloned from Arabidopsis, AtNHX1, when overexpressed in that plant, increased its salt tolerance so that plants could then grow and set seed at 200 mm NaCl, whereas before they were limited to 100 mm NaCl (Apse et al., 1999). Similar results with AtNHX1 were presented for tomato and brassica (Zhang & Blumwald, 2001; Zhang et al., 2001). In those experiments the death of control plants (plants transformed with vector only) at high salinity does not enable us to understand the function of the gene at the whole-plant level. The death of tomato plants at 200 mm NaCl is puzzling, as in other studies the same tomato cultivar, Moneymaker, could grow and set seed at 200 mm NaCl (Flowers, 2004). When AtNHX1 was overexpressed in wheat in moderately saline soil, with an ECe of 10 dS m−1, the grain yield of the best transgenic line was reduced by only 50%, compared with 65% in the untransformed control plants (Xue et al., 2004). Na+ concentrations in the root were unchanged by the treatment, but Na+ in the shoot decreased. It is therefore difficult to know how the antiporter was acting, as one might expect higher tissue concentrations to be tolerated. When the cotton orthologue, GhNHX1, was overexpressed in tobacco there were more dramatic increases in salt tolerance (Wu et al., 2004).

Overexpression of the plasma membrane Na+/H+ antiporter SOS1 in Arabidopsis increased the tolerance of Arabidopsis (Shi et al., 2003). Roots grew more quickly at 100 mm NaCl, and the percentage of plants that survived 5 d at 200 mm NaCl was increased from 17 to 43%. This was probably the result of the 50% reduction in Na+ uptake by plants (Shi et al., 2003).

The activity of Na+/H+ antiporters could be limited by their number, or by the H+ difference across the membranes. In that case, increasing the capacity of a proton pump would increase the salt tolerance of the plant. The vacuolar H+-PPiase, AVP1, which may be important to energize the vacuole membrane under salt stress, was overexpressed in Arabidopsis and increased its salt tolerance (Gaxiola et al., 2001). Transformed plants were able to grow in 250 mm NaCl, whereas the wild type died. The enhanced performance of transformed plants may have been caused by their higher accumulation of Na+ and K+.

HKT transporters are probably very important in regulation of K+ and Na+ transport from root to shoots, but the mechanism is not known. Transgenic studies have yielded somewhat inconsistent results with TaHKT1 (Laurie et al., 2002) and AtHKT1 (Rus et al., 2004). However, a knockout mutant of AtHKT1 clearly increased salt sensitivity (Rus et al., 2004), which shows that the function of HKT1 in the control of K+ or Na+ transport is important.

The HAL1 gene from yeast controls K+/Na+ selectivity and salt tolerance of yeast cells. Expression in tomato increased fruit yield and enhanced K/Na+ selectivity in leaves (Rus et al., 2001). The exact function of this gene in higher plants is not known.

The experiments described above all measured plant growth; studies with cells or calluses are not included, nor are studies that present only qualitative data. A complete list of transformation experiments is given by Flowers (2004). He suggests that claims of improved salt tolerance are not substantiated in many of these cases because of poor experimental designs, for example a single high-salinity treatment for 3 d, and inappropriate choices of methods to evaluate tolerance, such as photographs or survival numbers. A checklist of methods is suggested at the end of this review.

2. Genes with an osmotic or unknown protective function

Molecules having a protective function include small organic compounds that are variously called osmolytes, osmoprotectants or compatible solutes. These have two functional roles: at high concentrations, osmotic adjustment; and at low concentrations, an unknown protective role. Other gene products include enzymes that ‘mop up’ free radicals, and proteins that protect the formation or stability of other proteins. None of these is confined to salinity: all occur under drought as well as salinity stress, and sometimes under other stresses that also reduce growth, such as low temperature.

Protective molecules are referred to in this section as the gene product, rather than the gene, for brevity and familiarity.

Organic solutes and proteins with osmotic or other protective functions  Compatible solutes or osmolytes would be essential for coordinated regulation of vacuolar and cytoplasmic volumes. The term ‘compatible solute’ was advanced by Wyn Jones et al. (1977) to describe nontoxic solutes that could increase in high concentrations in the cytosol and be compatible with metabolic activity. They would be important to adapt plants to drought, as they could enhance osmotic adjustment and allow turgor maintenance of cells that would otherwise dehydrate. Tissue dehydration is less likely in salinity than drought, as the uptake of NaCl promotes osmotic adjustment, but compatible solutes would be essential to balance the osmotic pressure of the cytoplasm where salts accumulated preferentially in the vacuole.

In addition, certain solutes could have a metabolic protective role – they could stabilize soluble or membrane proteins, and so maintain growth at high salinity. The term ‘osmoprotectant’ has arisen for this function (reviewed by Rhodes et al., 2002). Specific functions are uncertain in higher plants; their putative function is based on that in microorganisms. The metabolic pathways and genes controlling their synthesis or accumulation have been summarized in comprehensive reviews by Hare et al. (1998); Nuccio et al. (1999); Chen & Murata (2002) and, with particular reference to salinity, Rhodes et al. (2002). There are four main classes of solute that could have an osmotic or protective role: N-containing solutes such as proline and glycine betaine; sugars such as sucrose and raffinose; straight-chain polyhydric alcohols (polyols) such as mannitol and sorbitol; and cyclic polyhydric alcohols (cyclic polyols) (Table 4). Many of these may be important as compatible solutes during freezing.

Table 4. Major categories of solutes with possible function as osmolytes or protectants, and results of key overexpression transformation studies, summarized from reviews by Hare et al. (1998); Nuccio et al. (1999); Chen & Murata (2002); Rhodes et al. (2002); Smirnoff (1998)
Solute typeFunctionNatural concentration rangeOverexpression studies with whole plants that resulted in increased salt tolerance
  1. All these studies used 100–200 mm NaCl for 2–4 wk in most cases, and increased the salt gradually at 25–50 mm d−1. P5CS (Δ-1-pyrroline-5-carboxylate synthetase modified from Vinca aconitifolia); codA (choline oxidase from Arthrobacter globiformis); otsA and otsB (trehalose-6-phosphate synthase and trehalose-6-phosphate phosphatase from Escherichia coli); mt1D (mannitol-1-phosphate dehydrogenase from E. coli); S6PDH (sorbitol-6-phosphate dehydrogenase modified from apple); imt1(d-myo-inositol methyltransferase from Mesembryanthemum crystallinum).

N-containing solutes such as proline, glycine betaine, trigonelline, proline betaineTurgor maintenance.1–50 mmproline P5CS modtobacco60 mm Hong et al. (2000)
Possibly redox regulation, to buffer cellular redox potential as metabolic activities slow down during stress, and so to promote recovery from stress. glycine betaine codA rice5 mm Sakamoto et al. (1998)
TrehaloseSignalling, possibly through hexose sensing.5–10 µmtrehalose otsA, otsBrice300 µm Garg et al. (2002)
Straight-chain polyhydric such as mannitol and sorbitolTurgor maintenance, particularly in the cytosol.1–50 mmmannitol mt1D wheat2 mm Abebe et al. (2003)
Mannitol may be involved in hexose sensing or in mopping up hydroxyl radicals. mannitol mt1D tobacco6 mm Tarczynski et al., 1993)
  sorbitolS6PDHpersimmon60 mm Gao et al. (2001)
Cyclic polyhydric alcohols such as myo-inositol, ononitol and pinitolTurgor maintenance in certain halophytes.1–200 mmononitol imt1 tobacco35 mm Sheveleva et al. (1997)
Might also mop up hydroxyl radicals.      
Sugars such as fructose, sucrose, glucose and raffinoseTurgor maintenance, carbohydrate storage.1–200 mm

If the concentrations of compatible solutes were high enough to contribute to turgor maintenance in the cell as a whole, production of these solutes would have an energy cost and hence a growth penalty. Their synthesis would divert substrates from metabolic pathways essential to growth, such as protein synthesis and cell-wall synthesis (Munns, 1988). However, they could contribute to leaf longevity when salt concentrations start to rise. Hence they may not make the plant grow more quickly when the salt is first applied, but would allow it to avoid wilting, to avoid toxicity in the presence of high internal concentrations of salt, and possibly to survive a period of stress.

Proteins with protective roles of unknown function include late-embryogenesis-abundant proteins (LEAs) and their close relatives, dehydrins. These are produced as cells lose water, and confer either tolerance to dehydration or recovery on subsequent hydration, but the biochemical mechanism is uncertain. They have properties similar to chaperonins, and may help to stabilize protein structure when cells have high salt concentrations or start to lose water. However, cell dehydration is not a common event in salt-treated plants, as leaf tissue is usually turgid (because of high salt content) and loses water only when the sodium rises to toxic levels. Many of these protective proteins are expressed only after osmotic shocks.

Enzymes that detoxify reactive oxygen species (ROS) may have an essential role in adapting plants to salinity stress. Free oxygen radicals (oxygen molecules or hydroxyl groups with an extra electron) are highly reactive and can oxidize essential compounds, particularly membrane lipids, unless detoxified by antioxidants. Important pathways of detoxifying ROS involve superoxide dismutase, catalase, and the ascorbate–glutathione cycle (Foyer et al., 1994). These enzymes are induced under drought and salinity, and during other stresses such as high light or extremes of temperature. The activities of the antioxidant defence systems are important in limiting photo-oxidative damage and in detoxifying ROS produced in excess of those normally required for signal transduction or metabolism (Foyer et al., 1994). Any environmental condition that suppresses photosynthesis (such as low temperature, salinity or drought) while plants are in high light could allow the production of excessive and potentially toxic ROS.

How to evaluate the function of protective compounds in salt tolerance  As the function of most of these compounds is unknown, their benefit can be measured only by their effect on either leaf injury or plant growth rate. Many solutes accumulate under abiotic stress simply because the supply of assimilate by source leaves has exceeded its demand by growing tissues (Munns, 1988), so an increase in any one solute may not make a plant grow more quickly. It might even make it grow more slowly if there is a high nitrogen or energy cost for its synthesis. Should growth rate be measured over the short or long term? That is, while osmotic stress dominates the growth response (phase 1), or when internal salt toxicity can occur in some genotypes (phase 2)?

Compatible solutes would exert their effect in phase 1, but could also function in phase 2. The effect of higher proline accumulation, or of other compounds that enhance turgor maintenance, could be seen on the rate of elongation of new leaves and roots. High turgor does not necessarily make plants grow more quickly (Termaat et al., 1985; Munns, 2002), but they cannot grow at all if they are wilted. If compatible solutes function to maintain turgor as the leaves age, then longer-term measurements, taking in whole-plant growth, are appropriate.

The effect of osmoprotectants that accumulate to only low concentrations would be seen only in phase 2. These would protect metabolic function in cells in which salt concentrations are high. They are therefore most likely to function in photosynthetic tissues of older leaves, and to affect growth in the long term. Proteins with protective or chaperonin-like functions are also most likely to function in the long term.

Enzymes that detoxify ROS would feature in the long term to protect metabolic function in cells in which salt concentrations are high, and to prevent premature senescence. They could also be very important in the short term, particularly in tissues exposed to high light, as they would help photosynthesizing leaves to remain functional despite low stomatal conductance, which occurs soon after plants are exposed to salinity because of the osmotic effect of salt outside the roots (James et al., 2002).

Results of transformation studies  Many papers have reported positive results for salt tolerance when plants are transformed with genes for osmoprotectants or protective proteins, summarized by Hare et al. (1998); Nuccio et al. (1999); Chen & Murata (2002); Gorham & Wyn Jones (2002); Rhodes et al. (2002); Flowers (2004). A few cases are highlighted here and in Table 4.

Some species do not accumulate glycine betaine under stress, including the model species Arabidopsis, rice and tobacco. Overexpression of genes for glycine betaine synthesis in these species enhanced salt tolerance in terms of biomass production or survival, despite the fact that the glycine betaine level in the transgenic plants was much lower than in species which naturally accumulate this compound, and too low to have a significant effect on osmotic pressure even if confined to the cytoplasm (reviewed by Chen & Murata, 2002). Thus improved salt tolerance in these transgenics must have been via a protective effect of the low levels of glycine betaine, rather than an osmotic effect bringing about more favourable water relations. The level of glycine betaine found in transformed plants was no more than 5 mm in the leaves (Chen & Murata, 2002), much lower than in other solutes such as proline or sugars.

All species accumulate proline under stress, and the gene for its synthesis, P5CS (Δ-1-pyrroline-5-carboxylate synthetase), is induced rapidly by stress in all tissues (Hong et al., 2000). However this enzyme is subject to feedback control, and overexpression with this gene increases the concentration of proline in leaves only twofold (Hong et al., 2000). Tobacco carrying a modified P5CS gene from Vinca aconitifolia to avoid feedback inhibition had a fourfold increase in proline (up to 60 mm compared with the control, with vector only, of 15 mm) in 200 mm NaCl (Hong et al., 2000). These plants had a higher germination rate, a greater seedling weight at 2 wk, and produced fewer free radicals. That is, they were more tolerant to the osmotic stress.

Overexpression with the bacterial gene mt1D for mannitol synthesis has resulted in increased salinity tolerance, for example in wheat (Abebe et al., 2003). Shoot fresh weight in 150 mm NaCl was reduced by 77% of plants lacking the mt1D gene, but by only 50% in transformed plants. The transformed plants had only 2 mm mannitol, too little to influence their water relations, requiring some other explanation for the improved growth.

Overexpression with the bacterial genes for trehalose synthesis increased salt tolerance in rice: there was fourfold greater dry weight after 4 wk in 100 mm NaCl in transformed than in untransformed plants (Garg et al., 2002). Trehalose levels increased in transformed plants, but were still only 0.3 mm in leaves after salt treatment, so the trehalose may have altered growth through signalling pathways (Hare et al., 1998).

In summary, overproduction of many of these solutes through gene transformation has convincingly increased growth of plants in saline soil – but we do not know how.

3. Genes that control cell and tissue growth rates

Genes controlling cell growth and leaf function  Genes that could increase the growth rate of plants in saline soil would influence the rate of production of new leaves and roots (by controlling the rate of cell division; the development of new primordia for shoot or root branching; the rate of cell wall expansion; or the dimensions of differentiated cells), or they would influence the rate of photosynthesis (by controlling stomatal aperture; or the dimensions of mesophyll cells, which influences leaf morphology and transpiration efficiency).

Candidate genes controlling growth are probably involved in signalling pathways that start with a sensor and involve hormones, transcription factors, protein kinases, protein phosphatases and other signalling molecules such as calmodulin-binding proteins. It is highly likely that such genes are common to drought stress (reviewed by Chaves et al., 2003), and are common to other stresses such as cold, and soil conditions that reduce growth such as soil hardness caused by compaction or sodicity. Many are induced by treatment with ABA. The sensor may be a metabolite that changes in concentration, or a membrane protein that changes in configuration, as the cell shrinks in response to a salt, drought or cold stress, or a long-distance signal moving from roots to shoots in the transpiration stream.

Progress on the discovery of transcription factors and signalling pathways is fast, and is being covered by a series of comprehensive reviews and updates (Shinozaki & Yamaguchi-Shinozaki, 2000; Knight & Knight, 2001; Zhu, 2002; Wang et al., 2003; Zhang et al., 2004).

Few transcription factors have known physiological functions, and all are likely to be tissue- or cell-specific, but (apart from guard cells) no tissue or cell specificity has been investigated. This point is discussed further in the following section. Trehalose is considered to have signalling functions in growth, but again no tissue specificity has been investigated. Most of the transcription factors are induced by rapidly imposed and severe stress, so their function in adapting a plant to a long-term salinity stress is not yet known.

How to evaluate the function of genes controlling leaf growth in salt tolerance  Whether experiments are short or long term would depend on the function of the genes. Genes affecting the rate of cell division and elongation would be seen soon after the salinity is applied (in phase 1). Genes involved in leaf morphology, leading to enhanced transpiration efficiency, would take longer to show their effect, through enhanced carbon supply to growing regions. With overexpression studies of transcription factors, and without knowing exactly what a given transcription factor ‘does’, long-term growth experiments are appropriate. Comparisons with other stresses that reduce growth, such as water stress, low temperature and soil hardness, would indicate whether or not there are salt-specific factors involved.

Results of transformation studies  Studies on overexpression of 13 transcription factors are summarized by Zhang et al. (2004), although some of these studies are liable to the criticisms pointed out by Flowers (2004). So far no studies have proved that the enhanced responses to abiotic stresses are salt-specific.

IV. Gene activity expected in roots, leaves and growing tissues of plants exposed to salinity, and results of gene expression studies

This section is presented from a physiological perspective, and outlines the gene activity that might be expected in specific organs or tissues. It considers the target tissue and the processes that might be affected by salinity, and discusses the results of gene expression studies in those tissues.

One factor common to all gene expression studies is that the salt is applied as a ‘shock’: that is, the plants are transferred to saline solution in one step. Sometimes the solution concentration is as high as 250 mm NaCl, which profoundly disturbs the water relations of all cells in the plant. This causes rapid, although transient, reductions in cell volume and turgor, and affects all metabolic and physiological processes that depend on substrate concentration or cell volume, including stomatal conductance and therefore photosynthesis. Recovery occurs within hours, the time taken for recovery depending on the degree of osmotic shock (Munns, 2002). If the osmotic shock is too great, cells slowly die.

Most results of gene expression studies show nonspecific effects of treatment. A number of reports describe changes in the transcriptome under the influence of salinity, cold and drought (Table 5), and all show a large proportion of changes induced by multiple treatments. For example, of 167 Arabidopsis genes that were strongly induced by salt stress (250 mm NaCl shock), only 15 were specific to salt, and 101 were equally induced by ‘drought’ (rapid dehydration) treatment (Seki et al., 2002). These included genes involved in carbohydrate metabolism. A significant proportion of genes that are induced by salt stress are induced by cold as well as dehydration stress (Kreps et al., 2002; Rabbani et al., 2003).

Table 5. Gene expression studies by microarray analysis
Tissue analysedSpeciesReferenceSalt treatment (NaCl)Other treatmentsComments
RootsRice (two cvs) Kawasaki et al. (2001)150 mm, 0.5–12 h, 7 dEarly times showed ABA-induced transcripts, later times showed defence-related genes, water channels induced at all times.
Roots Arabidopsis Maathuis et al. (2003)80 mm, 2–96 hK+ starvation, Ca2+ starvationOligonucleotides representing 1096 Arabidopsis transporter genes.
Roots, leaves separately Arabidopsis Kreps et al. (2002)100 mm, 3–27 h200 mm mannitol, coldLargest category of salt-specific genes were for oxidative stress resistance.
Root, leaves separatelyBarley Ueda et al. (2004)200 mm, 1, 24 hPEGSalt-responsive ESTs showed differential expression with salt and osmotic stress.
Roots, leaves, floral organs separately Arabidopsis Chen et al. (2002)100 mm200 mm mannitol, cold, wounding, pathogen attack; jasmonic acid402 transcription factors tested at different stages of development; many associated with senescence and pathogen attack.
LeavesBarley Atienza et al. (2004)175 mm, 24 hCold, dehydration, high light, copper toxicityAll stresses induced dehydrins and proline synthetic enzymes, and repressed chlorophyll-binding proteins.
Whole plant Arabidopsis Seki et al. (2002)250 mm, 1–24 hCold, dehydration40 of induced genes (11%) were transcription factors.
Whole plant Arabidopsis; Thelungiella halophila Taji et al. (2004)250 mm, 2–24 hSalt-related genes were expressed constitutively in T. halophila.
Whole plantRice Rabbani et al. (2003)250 mm, 1–24 hCold, dehydration, ABAOf the 57 genes induced by salt, none was specific to salt.
Whole plantBarley Ozturk et al. (2002)150 mm, 24 hDehydrationMany ABA-responsive genes were induced.
Whole plantPoplar Gu et al. (2004)300 mm, 0.5–72 h; 1–48 h after salt removedDetailed time course; RT–PCR on specific transcripts.

Many genes that are identified by expression studies include those in common with pathogen infection, such as those involved in salicylic acid-, jasmonic acid- and ethylene-signalling pathways. Others concern general abiotic stress responses, such as dehydrins, sugar transporters, chaperonins and heat-shock proteins. Many transcription factors are induced by other stresses, in particular dehydration and cold (Chen et al., 2002). If genes are induced by cold, as well as salinity and dehydration, these can be eliminated as having a role in adaptation to salinity, as they would be concerned only with adjusting to sudden changes in growth rate or changes in volume of cells or compartments, or water uptake. Low temperatures can inhibit water uptake.

Changes in expression of aquaporins (water-channel proteins) are common to all studies, and may be a response to the sudden shrinkage of all cells and organelles after the osmotic shock. Rapid volume adjustment of essential organelles such as chloroplasts, via aquaporins, would be an essential adaptation to the changed water status. However, some long-term experiments also show changes in aquaporin expression: for example, aquaporin transcripts were still upregulated after 1 wk salt treatment in rice (Kawasaki et al., 2001). This indicates an important adaptive role of these proteins.

1. Roots

Gene functions expected in roots  Genes for salt tolerance expected to function in roots would be involved in the control of ion uptake and transport to the shoots, and the regulation of root ion homeostasis and water status.

Roots have a remarkable ability to control their own Na+ and Cl concentrations. Their ion concentrations do not increase with time, and at high salinities they can have a much lower Na+ and Cl concentration than the external solution. For example, in a number of bread wheat genotypes growing in 150 mm NaCl, Na+ in the roots was only about 30 mm, with a small variation about that value with different genotypes (Gorham et al., 1990). Similarly low values for both Na+ and Cl with different durum and bread wheat genotypes were found in salinities from 50 to 150 mm NaCl (Husain et al., 2004; one genotype shown in Fig. 5). Barley also has a remarkably low Na+ concentration in roots: for example, plants grown at 250 mm NaCl had only 120 mm NaCl in their roots (1.8 mmol g−1 d. wt; Storey & Wyn Jones, 1978). Roots would therefore be accumulating organic solutes to maintain turgor, so enzymes that synthesize or transport organic solutes such as proline or sucrose should increase in activity.

Figure 5.

Effect of a range of NaCl concentrations on Na+ concentrations in the roots of durum landrace Line 141 after 14 d treatment, taken from two experiments with overlapping ranges of NaCl (Husain et al., 2004). Values are means (n = 4) ± SE. Cl concentrations were similar to Na+ (Husain et al., 2004). Dotted line, 1 : 1 relation between internal and external Na+ concentration.

The activity of transporters that control the net uptake of Na+ or Cl from the external solution, the compartmentation of these ions in vacuoles, and the loading of the xylem should be increased by NaCl treatments. As Na+ will enter root cells passively from a saline soil (Appendix S2), Na+ efflux might increase dramatically, requiring an antiporter such as SOS1 and a plasma membrane proton pump. K+ channels and symporters such as HKT1 may also be upregulated to maintain K+ homeostasis. Channels that are likely to carry Na+, such as nonselective cation channels, may be downregulated. Cation/H+ antiporters that control the ion homeostasis of organelles may be upregulated. Na+ will not enter the xylem passively unless there is a large membrane depolarization (Appendix S2), but may enter through K+ transporters, as K+ loading of the xylem could occur through active rather than passive transport (Appendix S2). The types of K+ transporter with the greatest selectivity for K+ over Na+ may be upregulated.

Genes involved in signal transduction pathways that regulate the rate of cell division and elongation in roots might be affected. The activity of processes that synthesize and deliver hormonal signals to the root tip could change, and ‘root signals’ that travel in the transpiration stream, and cause stomatal closure and reductions in leaf expansion rates, might be induced.

Genes found in roots: results of expression studies  Only one gene expression study has revealed the induction of relevant ion transporters in roots by salt treatment. A microarray was designed specifically to represent the major transporter genes from Arabidopsis that could be associated with cation excess or deficiency, and roots were examined over 96 h following salt stress (80 mm NaCl), Ca2+ deficiency and K+ deficiency (Maathuis et al., 2003). Salt stress induced the expression or repression of a number of transporters that could be involved in control of cation uptake across the plasma membrane, or in the storage of cations in vacuoles. Some particularly interesting genes included proton pumps (V-type H+-ATPases, AHA2) Na+/H+ antiporters (NHX1, NHX3) and cation/H+ antiporters (CHX10, CHX15). This paper is a valuable reference for future studies on transporters. However, there was no change in transcript level of SOS1. A possible reason is that the plants were grown on medium already containing 2 mm NaCl, and with the K+ concentration in the medium being less than 2 mm many transporters controlling the uptake and efflux of Na+ may have already been induced before the treatment started.

In the other studies that analyse roots separately, few K+ or Na+ transporters were found. The exception was the expression of a putative Na+/H+ exchanger, CHX17, but this occurred under cold and mannitol treatment, as well as under salt treatment (Kreps et al., 2002).

The reason for the absence of changed expression levels of transporters in microarray studies could be simply that the sequences were absent from the chips. Alternatively, it could be that their regulation was post-translational, that the genes were constitutively expressed, or that they had already been induced before NaCl treatment. They could be fully induced by the basal nutrient solutions commonly used. Hoagland and Arnon's solution contains 6.5 mm K+ and ≈ 0.2 mm Na+. Murashige and Skoog's medium contains 20 mm K+ and ≈ 0.8 mm Na+. The K+ and Na+ concentration might be high enough to induce all transporters regulating K+ and Na+ uptake and intracellular transport.

Genes in roots that are often upregulated by salt stress are involved in antioxidant responses, such as ascorbate peroxidase and glutathione reductase (Kawasaki et al., 2001; Kreps et al., 2002; Ueda et al., 2004). These are usually also expressed under other stresses, to different degrees.

Many transcription factors are altered by stress. An extensive and thorough study was reported by Chen et al. (2002), who examined the mRNA levels of 402 transcription factors in various plant tissues after different abiotic and biotic stresses, in highly replicated experiments with plants at different stages of development. About 15% were highly expressed in roots, and some were root-specific (Chen et al., 2002). However, few of these were induced by abiotic stress alone, and none was salt-specific.

In most studies the genes that are upregulated or downregulated have little relevance to salinity, and are expressed under unrelated stresses such as low temperature. The most likely explanation is the nature of the treatments, which (whether salt, cold, or dehydration) are sudden and severe, and plants are harvested within 1–2 h of treatment. The earlier time points are taken on the basis that the earliest response is the most important, but it is likely that most of these are transitory and are involved in repairing membranes, reinstating the volume of cell organelles, and metabolic adjustments to the sudden increases in metabolites caused by the sudden withdrawal of water.

Genes that peak over the first 2 h after salt stress are undoubtedly involved in repair of damaged membranes, particularly in plants that are exposed to concentrations of 150 mm NaCl or higher. Roots in solution will plasmolyse. The turgor of young root cells is in the range 0.5–0.7 MPa (Pritchard et al., 1991; Frensch & Hsiao, 1994). The osmotic pressure of a solution of 150 mm NaCl is ≈ 0.7 MPa, so cells will plasmolyse if plants are transferred in one step to 150 mm NaCl. If cell turgor is only 0.4 MPa, which is likely for older cells in roots (Pritchard et al., 1991), then a one-step transfer to a solution of 100 mm NaCl (0.5 MPa) will plasmolyse cells. With 200 mm NaCl, the protoplast would shrink to half its original volume to achieve equilibrium (Munns, 2002). It will regain its original volume with time by accumulation of solutes, by the process known as osmotic adjustment.

This process of recovery from osmotic shock possibly explains the findings that many genes expressed within a few hours of osmotic shock are in common with those expressed under other shock treatments. For example, in a study with Arabidopsis that compared responses to the relatively moderate shock treatments of 100 mm NaCl, 200 mm mannitol and 4°C, at 3 h, the number of shared responses was reduced tenfold over time, between 3 and 27 h after treatment (Kreps et al., 2002). That is, there was a progression with time towards stimulus-specific responses.

Ca2+ deficiency is a common problem with salinity studies (Cramer, 2002). It is well known that Ca2+ activity (the concentration of Ca which is fully ionized and hydrated) is lowered by adding salts to the medium, and that this causes effects on Ca2+ uptake, membrane selectivity of other cations, and ultimately growth of the plant (reviewed by Cramer & Läuchli, 1986; Reid & Smith, 2000; Tester & Davenport, 2003). For example, when half-strength Hoagland's solution (which contains 2 mm Ca) is used as the basal nutrient medium to which is added 100 mm NaCl or above, the chemical activity of Ca2+ will be about one-third of the concentration, which is below the level adequate for biological requirements. Few of the gene expression studies added supplemental Ca2+. Many effects of salt treatment on membrane transporters, including those for organic solutes, could be caused by impaired root membrane function caused by low Ca2+ activity.

2. Leaves

Gene functions expected  Genes for salt tolerance expected in mature leaves would have one of three general functions: to control stomatal conductance and photosynthesis; to prevent Na+ toxicity while maintaining turgor; or to regulate leaf senescence. These would all be under control of feed-forward signals coming from roots, and perhaps feedback signals from growing regions or sinks.

Gene activity expected would change over time, as the development of Na+ toxicity takes time. Figure 6 shows that the stress-induced reduction in stomatal conductance was seen when the leaf emerged, but after some time there was a further decline, probably caused by salt toxicity, as the Na+ and Cl concentrations in the leaf increased to about 250 mm at that time (James et al., 2002). Photosynthetic rate, in contrast, was not reduced when the leaf emerged. The high photosynthetic rate per unit leaf area was caused by the changed leaf and cell morphology, so that leaves with smaller cells of different dimensions, giving a higher chloroplast density per unit leaf area, were able to compensate for the lower stomatal conductance and maintain their photosynthetic rate (on a unit leaf area basis). This results in a high transpiration efficiency for the leaf, because the water lost per unit of carbon fixed is higher: an adaptation to dry or saline soil.

Figure 6.

Effect of 150 mm NaCl on (a) stomatal conductance; (b) net CO2 assimilation rate (A) of leaf 3 of the durum cultivar Wollaroi. Open triangles, controls, closed triangles, salt. Values are averages (n = 6) ± SE (James et al., 2002).

The need for compatible solutes would increase with time in mature leaves, as the Na+ and Cl concentrations increase to levels that must be sequestered in vacuoles or they would be toxic. Activity of the vacuolar antiporter NHX1 should increase, along with the proton pumps needed to energize it. Also, the need for protectants would increase with time, as increasingly older leaves are exposed to light with higher Na+ and/or Cl concentrations and a lesser ability to fix carbon. Genes for antioxidant molecules and enzymes should be upregulated. The increased production of ROS is common to all stress conditions, particularly where the plant is exposed to high light, but its rate of photosynthesis is limited by stomatal conductance or by reduced sink demand. Enzymes such as superoxide dismutase, ascorbate peroxidase and glutathione reductase might be upregulated.

With leaf age, the potential for the onset of senescence increases, and stress can accelerate the ageing process (Munné-Bosch & Alegre, 2004). For this reason, it is important to distinguish leaves that are young (recently produced and recently reached their final leaf area) from those that are older. Younger leaves are at their peak rate of photosynthesis (Rawson et al., 1988), and possibly the peak rate of all processes involving carbon and nitrogen assimilation. The salt concentration in these leaves is probably not yet high (James et al., 2002). Younger leaves will be conserved under stress at the expense of older leaves. In older leaves the normal, regulated senescence process may have started, and stress may trigger the early onset of this process and the production of senescence associated genes, or may trigger a somewhat different process called drought-induced senescence (Munné-Bosch & Alegre, 2004). Regrettably, gene expression studies do not separate leaves of different ages.

Genes found  Few gene expression analyses are carried out specifically on leaves or shoots (Table 5), and in few of these have expected genes been upregulated. The gene for proline synthesis, P5CS (Δ-1-pyrroline-5-carboxylate synthetase), is commonly detected (Atienza et al., 2004; Ueda et al., 2004). It was upregulated 8.5-fold by 100 mm NaCl, as well as by 200 mm mannitol and, to a lesser extent, by cold in Arabidopsis (Kreps et al., 2002). Genes for enzymes with key roles in the synthesis of other osmoprotectants are upregulated, such as myo-inositol 1-phosphate synthase (Kreps et al., 2002) and betaine aldehyde dehydrogenase (Ueda et al., 2004).

Genes coding for proteins, such as the antioxidants peroxidase, glutathione reductase and superoxide dismutase, are often expressed (Kreps et al., 2002), as are dehydrins and LEAs (Atienza et al., 2004). These are rarely stress-specific.

Photosynthesis-related transcripts, such as those for chlorophyll-binding proteins, are downregulated as predicted (Ozturk et al., 2002; Atienza et al., 2004; Gu et al., 2004).

Most gene expression studies are carried out on plants after shock treatments, and the gene response over the first few hours is probably dominated by recovery from the shock. Leaf cells will not suffer the same trauma as roots after an osmotic shock. If the cells are not in contact with the plasmolysing solution, and are still surrounded by air, the wall will collapse around the shrinking protoplast in the phenomenon known as cytorrhesis (Munns, 2002). Although the integrity of the cell wall is affected, the plasma membrane remains attached, so there is presumably less damage to the plasma membrane than with plasmolysis. This also applies to root cells in drying soil, or when dehydrating on a laboratory bench. However, volume shocks still occur. A K+ channel protein, AKT3, was upregulated by salt stress, but also by mannitol and cold stress (Kreps et al., 2002), so it is probably related to the adjustment of K+ transport and regaining of cell volume to maintain K+ homeostasis, for the same reason that aquaporins are often upregulated as well.

3. Growing tissues: leaf growing zones and root tips

Gene functions expected  Gene activity in growing tissues would be involved with regulation of the rate of cell production, the initiation of primordia, and the rate at which cells expand. They would be under the control of hormones and regulating genes in a signal transduction pathway. Such genes would encode transcription factors, protein kinases, protein phosphatases and other signalling molecules such as calmodulin-binding proteins. Genes involved in osmotic adjustment would be important in providing nontoxic solutes for generating sufficient turgor for the expanding cells. Genes to mobilize carbon storage compounds and generate more ATP would be needed, as an increased growth rate of cells would demand higher energy production.

Genes found  No expression studies have been carried out on growing tissues such as root tips or elongating zones of leaves, even in those studies focusing on transcription factors and signal transduction, which are likely to be tissue- or cell-specific.

V. Conclusions

Application of the knowledge about candidate genes is currently hampered by our lack of understanding of their function at cell, tissue and whole-plant levels. Despite the increasing use of microarrays and RT–PCR, few expression studies have revealed transcripts that have a clear function in salt tolerance, and which can be tested by transformation experiments or by other means. One reason for this may be that the biochemical or physiological function of the gene or its product is unknown, as in the case of transcription factors and trehalose; and the function of intensively researched molecules such as proline and glycine betaine is incompletely understood. Also unknown is the tissue or cell type in which a gene might be acting. A second reason is that many experiments are not designed to expose genes that confer salinity tolerance under natural conditions. Many experiments with NaCl do not include supplemental Ca2+, and even NaCl concentrations as high as 150 or 250 mm NaCl have been given in one hit, with no additional Ca2+. It is unlikely that important genes controlling the transport of Na+ or K+ under normal conditions of soil salinity would be discovered by such experiments. Important genes are more likely to be revealed by contrasting genotypes undergoing realistic treatments.

Realistic experimentation should:

  • • avoid large osmotic shocks;
  • • avoid measurement while the cell is recovering from plasmolysis;
  • • avoid Ca2+ deficiency;
  • • measure traits associated with the gene action rather than just plant growth;
  • • measure the dry weight of plants, rather than take photographs;
  • • when comparing genotypes with different sizes, measure relative growth rates; and
  • • be replicated.

More than 700 patents have been granted that invoke salt tolerance as part of the invention, most involving specific genes (BIOS, 2005). However, the successful use of these genes to produce a salt-tolerant cultivar is very slow. Useful knowledge about genes would be increased if studies on plant transformation included measurements of biochemical or physiological processes in which the transgene is acting, and targeted particular tissues and salinity-specific responses.

The future of transgenic approaches is uncertain, first because of the reluctance of most consumers to accept transgenically modified food; and second because a single gene change may not make any difference. The redundancy in many gene families may have evolved to compensate for the natural mutation and thereby change in expression of any one important gene. However, the realization of the extent of man-made and natural subsoil salinity is still awakening, and plants with greater salt tolerance will be more in demand than ever before. Innovative application of molecular tools can provide new approaches to increasing the salt tolerance of crops. This is an exciting and important field that will flourish under close interactions between molecular biologists, geneticists and physiologists, and will benefit from timely feedback from plant breeders.


I thank many colleagues for helpful suggestions and critical comments on the manuscript: Romola Davenport and Mark Tester for advice on ion-transport mechanisms, and John Boyer, Tim Colmer, Tim Flowers, André Läuchli, John Boyer, Daniel Schachtman and John Passioura for pertinent comments.