AOBPreview originally published online on February 6, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Annals of Botany 91: 503-527, 2003
© 2003 Annals of Botany Company
REVIEW |
Na+ Tolerance and Na+ Transport in Higher Plants
1 Department of Plant Sciences, University of Cambridge, Downing St, Cambridge CB2 3EA, UK
* For correspondence. Fax + 44 1223 333953, e-mail mat10{at}cam.ac.uk
Received: 1 July 2002; Returned for revision: 16 September 2002; Accepted: 17 December 2002 Published electronically: 6 February 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT THE PROBLEMS OF SALINITY...MECHANISMS OF SALINITY TOLERANCECONCLUSIONSLITERATURE CITED |
|---|
Tolerance to high soil [Na+] involves processes in many different parts of the plant, and is manifested in a wide range of specializations at disparate levels of organization, such as gross morphology, membrane transport, biochemistry and gene transcription. Multiple adaptations to high [Na+] operate concurrently within a particular plant, and mechanisms of tolerance show large taxonomic variation. These mechanisms can occur in all cells within the plant, or can occur in specific cell types, reflecting adaptations at two major levels of organization: those that confer tolerance to individual cells, and those that contribute to tolerance not of cells per se, but of the whole plant. Salt-tolerant cells can contribute to salt tolerance of plants; but we suggest that equally important in a wide range of conditions are processes involving the management of Na+ movements within the plant. These require specific cell types in specific locations within the plant catalysing transport in a coordinated manner. For further understanding of whole plant tolerance, we require more knowledge of cell-specific transport processes and the consequences of manipulation of transporters and signalling elements in specific cell types.
Key words: Review, sodium, salinity, ion transport, non-selective cation channels, long-distance transport.
|
THE PROBLEMS OF SALINITY TOLERANCE |
|---|
|
TOPABSTRACTTHE PROBLEMS OF SALINITY...MECHANISMS OF SALINITY TOLERANCECONCLUSIONSLITERATURE CITED |
|---|
Background
High concentrations of salts in soils account for large decreases in the yield of a wide variety of crops all over the world. The problem is huge; almost 1000 million ha of land is affected by soil salinity (Szabolcs, 1994), 7 % of all land area. Of the 1·5 billion ha that is cultivated, about 5 % (77 million ha) is affected by salt (Munns et al., 1999). Critically, the problem of salinization is increasing, often due to bad agricultural practices. Irrigated land is particularly at risk with approx. one-third being significantly affected by salinity. Despite its relatively small area, irrigated land is estimated to produce one-third of the world’s food (Munns, 2002), so salinization of this resource is particularly critical. However, dryland salinity is also an important, and increasing, problem, at least in some areas of the world, such as in southern Australia. The recently released Environment Australia-commissioned National Land and Water Resource Audit predicts that, on a business as usual basis, 17 million ha of Australia’s agricultural land will be significantly affected by salinity by 2050, this comprising 25 % of Australia’s wheat belt (see http://audit.ea.gov.au).
The significance of soil salinity for agricultural yields is enormous. The agricultural problem of salinity tolerance is probably best tackled by either altering farming practices to prevent soil salinization occurring in the first place, or by implementing schemes to try to remediate salinized soils (such as by planting perennials to lower water tables). However, such approaches are also well complemented by programmes to increase the salt tolerance of plants, by either traditional breeding or genetic manipulation technologies. In this way yields can be increased on affected soils whilst they are being remediated, and plants can also maintain increased growth when they encounter saline sub-soils, thus enabling them to form part of the remediation process itself.
Concentrations of various salts can be elevated in soil. For example, irrigation can lead to high concentrations of calcium and magnesium carbonates, whereas in areas with a geologically recent marine history or where there has been prolonged deposition of wind-borne marine salts, and where rainfall is too low to leach salts, soils can be high in NaCl. In addition, saline soils often have toxic concentrations of the weak acid, boric acid, whose mechanism of both transport and toxicity remain unknown.
All salts can affect plant growth, but not all inhibit growth. In addition, salts do not act alone in the soil, but interact in their effects on plants; some of these interactions are simple (e.g. interactions between Na+ and Ca2+), whereas some are complex (e.g. carbonates, and their effects via increased soil pH).
Among the most common effects of soil salinity is growth inhibition by Na+ and Cl–. For some plants, especially woody perennials (such as citrus and grapevines), Na+ is retained in the woody roots and stems and it is the Cl– that accumulates in the shoot and is most damaging to the plant (often by inhibiting photosynthesis; Flowers, 1988). Cl– transport has been reviewed recently by White and Broadley (2001). However, for many plants (such as graminaceous crops), Na+ is the primary cause of ion-specific damage. In this review, we concentrate on the effects of Na+ on plants, and use the word ‘salt’ interchangeably with Na+.
Symptoms of damage
Saline solutions impose both ionic and osmotic stresses on plants. These stresses can be distinguished at several levels. In salt-sensitive plants, shoot and to a lesser extent root growth is permanently reduced within hours of salt stress and this effect does not appear to depend on Na+ concentrations in the growing tissues, but rather is a response to the osmolarity of the external solution (Munns et al., 2000a; Munns, 2002). Na+-specific damage is associated with the accumulation of Na+ in leaf tissues and results in necrosis of older leaves, starting at the tips and margins and working back through the leaf. Growth and yield reductions occur as a result of the shortening of the lifetime of individual leaves, thus reducing net productivity and crop yield (Munns, 1993, 2002). The timescale over which Na+-specific damage is manifested depends on the rate of accumulation of Na+ in leaves, and on the effectiveness of Na+ compartmentation within leaf tissues and cells. These Na+-specific effects are superimposed on the osmotic effects of NaCl and, importantly, show greater variation within species than osmotic effects (Munns, 2002).
At the molecular level, signalling mechanisms activated by salt stress include both drought-induced and Na+-specific pathways (see ‘Signalling pathways’ below). This review focuses on the Na+-specific effects of salinity, and particularly on the transport processes that give rise to these effects. However, it is important to bear in mind the osmotic component of salt stress, particularly when evaluating methods for screening for genotypic variation in salt sensitivity. One potentially useful method to separate ionic from osmotic components of NaCl damage is to use LiCl, a salt that is toxic at approximately one-tenth the concentration of NaCl and appears to share some signalling and transport pathways with Na+ (although it may exert its toxicity via different mechanisms to Na+). However, it is unlikely that all components of osmotic and ionic stresses are completely distinct. For instance, some tonoplast Na+/H+ antiporters are activated by osmotic stress, but impact on Na+-specific toxicity (see ‘Pumping Na+ into the vacuole’ below).
Some effects of high soil Na+ are also the result of deficiency of other nutrients (Silberbush and Ben-Asher, 2001), or of interactions with other environmental factors, such as drought, which exacerbate the problems of Na+ toxicity. Specifically, other nutrient deficiencies can occur because elevated Na+ inhibits the uptake of other nutrients by (a) disrupting the uptake of nutrients directly by interfering with transporters in the root plasma membrane, such as K+-selective ion channels; and (b) inhibiting root growth, by the osmotic effects of Na+ and because of the detrimental effects of Na+ on soil structure (Wild, 1988, Chapter 27). Thus, the uptake of water, growth-limiting nutrients (such as P, Fe or Zn) and the growth of soil microorganisms, such as mycorrhizal fungi, can be inhibited.
However, these effects are probably not as widespread a basis for primary Na+ damage as the direct effects of Na+ toxicity, and they will not be discussed further.
The main bases for Na+ toxicity
In shoots, high concentrations of Na+ can cause a range of osmotic and metabolic problems for plants. Leaves are more vulnerable than roots to Na+, simply because Na+ (and Cl–) accumulate to higher concentrations in shoots than in roots. Roots tend to maintain fairly constant levels of NaCl over time, and can regulate NaCl levels by export to the soil or to the shoot. Na+ is transported to shoots in the rapidly moving transpiration stream in the xylem, but can only be returned to roots via the phloem. There is limited evidence of extensive recirculation of shoot Na+ to roots, suggesting that Na+ transport is largely unidirectional and results in progressive accumulation of Na+ as leaves age.
Metabolic toxicity of Na+ is largely a result of its ability to compete with K+ for binding sites essential for cellular function. More than 50 enzymes are activated by K+, and Na+ cannot substitute in this role (Bhandal and Malik, 1988). Thus, high levels of Na+, or high Na+ : K+ ratios can disrupt various enzymatic processes in the cytoplasm. Moreover, protein synthesis requires high concentrations of K+, owing to the K+ requirement for the binding of tRNA to ribosomes (Blaha et al., 2000) and probably other aspects of ribosome function (Wyn Jones et al., 1979). The disruption of protein synthesis by elevated concentrations of Na+ appears to be an important cause of damage by Na+.
Osmotic damage (i.e. osmotically driven removal of water from cells) could occur as a result of the build up of high concentrations (possibly several hundred mM) of Na+ in the leaf apoplast, since Na+ enters leaves in the xylem stream and is left behind as water evaporates. This mechanism of Na+ toxicity was first proposed by Öertli (1968), and direct supporting evidence has been provided by x-ray microanalysis measurements of Na+ concentrations in the apoplast of rice leaves by Flowers et al. (1991). These authors calculated that there was about 600 mM Na+ in the apoplast of leaves of rice plants that were moderately salt-stressed. The universal applicability of the importance of apoplastic Na+ accumulation is nevertheless in question. By infiltrating leaves with water and rapidly washing it out by centrifugation, Muhling and Läuchli (2002) failed to measure large concentrations of Na+ in the leaf apoplast of salt-stressed maize and cotton. How much these contrasting results reflect biological or technical differences remains to be seen.
The cellular toxicity of Na+ causes another type of osmotic problem. Plants need to maintain internal water potential below that of the soil to maintain turgor and water uptake for growth. This requires an increase in osmotica, either by uptake of soil solutes or by synthesis of metabolically benign (‘compatible’) solutes. This drought component of salinity poses a dilemma for plants: the major, cheap solutes in saline soils are Na+ and Cl–, but these are toxic in the cytosol (see below). Compatible solutes are non-toxic, but are energetically much more expensive.
With high concentrations of Na+ in the leaf apoplast and/or vacuole, plant cells have difficulty maintaining low cytosolic Na+ and, perhaps as importantly, low Na+ : K+ ratios (Gorham et al., 1990; Dubcovsky et al., 1996; Maathuis and Amtmann, 1999). Cytosolic Na+ concentrations are probably in the order of 10–30 mM {measured in plants grown at various NaCl concentrations ranging from 1–200 mM using x-ray microanalysis [Jeschke and Stelter, 1976; Behl, cited in Jeschke (1984); Koyro and Stelzer, 1988] or ion-sensitive microelectrodes (Carden, 1999)}, although there have been reports of concentrations of 100 mM or even higher in saline conditions (Harvey, 1985; Hajibagheri et al., 1985, 1987; see also Cheeseman, 1988).
Methods by which to study the bases for Na+ tolerance
Physiological studies of salinity tolerance have undoubtedly benefited from advances in molecular and cellular techniques, but many experiments could be improved by improvements in the method of plant growth and manipulation. For example, measurements of responses over short times after a rapid and large change in salinity are inappropriate for the study of responses relevant to plant growth in saline soils where smaller, slower changes occur.
In addition, if Na+-specific toxicity is the subject of study, then care must be taken to distinguish those effects due to the osmotic component of salinity. For instance, plants growing in conditions of very low transpiration, such as on a sterile plate, are likely to be more sensitive to the osmotic component of salinity (as Na+ delivery to the shoot will be reduced). Likewise, inhibition of root elongation is probably not a good measure of Na+-specific toxicity, because it is not well correlated with Na+ transport to the shoot (Munns, 2002, but see the discussion of sos mutants below).
The choice of model systems to be used also requires more careful consideration. For example, results from our laboratory suggest that tobacco appears to be more sensitive to the low osmotic potential of growth solutions containing NaCl, rather than to the specific effects of Na+ (or Cl–) (Murthy and Tester, 1996). Thus, some studies in these plants using addition of NaCl may be more relevant to studies of drought tolerance rather than to salinity tolerance in particular.
We have also found that there is no relationship between Na+ accumulation and salt sensitivity in Arabidopsis thaliana. Likewise, as detailed later in this review, mutations that increased salt sensitivity of arabidopsis often did not cause increased accumulation of Na+ (see, for example, Mechanisms of salinity tolerance, Introduction). Moreover, although addition of Ca2+ reduced Na+ influx and accumulation in arabidopsis, it did not alleviate salt-induced inhibition of growth (P. A. Essah, R. Davenport and M. Tester, unpubl. res.). Although our observations are similar to reports of the Ca2+-insensitivity of some brassicas (Ashraf and Naqvi, 1992; Schmidt et al., 1993), they contrast with results for many crop species, where Na+ accumulation is correlated with toxicity, and supplemental Ca2+ both reduces Na+ accumulation and improves growth in saline conditions (Rengel, 1992; Cramer, 2002). This does not preclude the use of these model plants for the study of Na+ transport, but their use as a model for studying salinity tolerance in crop plants may be limited. In addition to highlighting the need for caution when using model plants, these considerations also recognize the diversity of plant responses to salinity, and the difficulty in making generalizations about plant responses to elevated salinity. These issues are addressed further below.
In addition to comparative physiological studies, techniques exploiting the ability to manipulate levels of expression of genes are starting to provide insights into the mechanisms of salinity tolerance. Both misexpression of targeted genes (e.g. Zhang and Blumwald, 2001) and the random alteration of genes (e.g. Zhu, 2000) have been used to good effect. One feature of note with random mutagenesis studies, though, is that many more salt-sensitive than salt-tolerant mutants appear to have been identified {Tsugane et al. (1999) and Quesada et al. (2000) being notable published exceptions for arabidopsis; and Zhang et al. (1999) for rice [The salt tolerant mutants identified by Saleki et al. (1993) and Werner and Finkelstein (1995) are not included here, because they were mutants tolerant only in germination, and this tolerance did not appear to extend to the growth of seedlings.]}. It is not known if this is related to the fact that most screens appear to use knock-out mutagenesis rather than activation mutagenesis, or to the fact that the mutants studied to date have had activity affected constitutively rather than in specific cell types.
This difference in numbers of tolerant and sensitive mutants also begs the question of whether tolerance is necessarily the opposite of sensitivity, or if salt-sensitive mutants are revealing processes distinct from those involved in tolerance. Another unusual feature of the mutagenesis studies published to date is that genes shown to affect salinity tolerance encode proteins involved in cytosolic or transport processes—we know of no transcription factors that have yet been identified from mutant screens. This is in stark contrast to screens for developmental mutants, for example. Given the potentially complex (multigenic) nature of salinity tolerance, forward genetic screens are an appropriate strategy for revealing genes involved in these processes, which makes the lack of appearance of genes controlling cellular processes at a higher level (i.e. transcription factors) the more surprising.
|
MECHANISMS OF SALINITY TOLERANCE |
|---|
|
TOPABSTRACTTHE PROBLEMS OF SALINITY...MECHANISMS OF SALINITY TOLERANCECONCLUSIONSLITERATURE CITED |
|---|
Introduction
We know that some plants tolerate high soil salinity better than others; agriculturally, yield is not reduced as much at a given salinity for some plants. This variability occurs between major groups of plants (mangroves and chenopods are examples of taxonomically distinct groups of plants dominated by salt-tolerant species), between closely related species, and between different varieties or even individuals within a varietal line. Differences between closely related plants are particularly interesting because they suggest that it may be possible to identify a small number of factors that influence salt tolerance.
Differences in yield under saline conditions often simply reflect differences in plant vigour. For example, Quarrie and Mahmood (1993) found a very close correlation between yield in control conditions and that in salinized conditions (Fig. 1). Likewise, vernalization can confer a degree of salinity tolerance (Taeb et al., 1992), as can earlier development. In fact, breeding for more vigorous plants has been argued to be agronomically the most effective approach for increasing yield in saline soils (Richards, 1992). However, such an approach must have its limits, with yield increases as a result of increased vigour now reaching their limit (Conway, 1997), and a strategy whereby plants are selected for an ability to maintain yields in salinized conditions must also be employed. If extra tolerance can be conferred on top of other desirable features, then this must surely be beneficial.
|
By selecting Triticum genotypes with similar biomass and growth rates under control conditions, Schachtman and Munns (1992) demonstrated that variations in leaf growth under saline conditions were correlated with leaf concentrations of Na+, independent of plant vigour. This type of variation is also evident in Fig. 1: line 19 showed unusually high productivity in highly saline conditions, and this was correlated with unusually low Na+ concentrations in the shoot. Empirically, for a wide range of species, it is often found that plants that are more able to tolerate moderately saline environments have a greater ability to exclude Na+ from the shoot, or at least the leaf blades, and concurrently maintain high levels of K+ (e.g. Greenway, 1962; Gorham, 1990; Schachtman et al., 1992; Colmer et al., 1995; Dubcovsky et al., 1996; Munns et al., 2000b; Flowers and Hajibagheri, 2001; Zhu et al., 2001). This has been found in crop plants bred for salinity tolerance, as well as in crops that were screened for tolerance (having been selected originally for other properties), and in wild relatives of some crop plants (e.g. Nevo et al., 1992).
This correlation appears to be particularly strong in the graminaceous crop species, although a notable exception to this generalization appears to be maize (Alberico and Cramer, 1993; Cramer et al., 1994; N. Al-Mansour and M. Tester, unpubl. res.), with growth inhibition being correlated with sensitivity of leaf extension to salt-stimulated increases in abscisic acid (ABA) (Cramer and Quarrie, 2002). Another exception is rice. Although sensitivity of rice cultivars to Na+ is related to shoot Na+ accumulation, the transport characteristics of Na+ in rice have been shown by Flowers and colleagues to be quite distinct from those of other cereals, with a large component of Na+ influx into roots being due to ‘leakage’ past the endodermis (Yeo et al., 1987, 1999; Yadav et al., 1996) (see ‘Bypass flow’, below).
Dicotyledonous species vary more than monocotyledonous species in the extent to which tolerance is associated with low shoot Na+. For example, a highly salt-tolerant wild relative of tomato, Lycopersicon peruvianum, accumulates higher concentrations of Na+ than the salt-sensitive domesticated tomato, L. esculentum (Tal, 1971; Santa-Cruz et al., 1999). Importantly, the sensitivity of wild-type arabidopsis (P. A. Essah, R. Davenport and M. Tester, unpubl. res.), and that of salt-sensitive (Zhu et al., 1998) and salt-accumulating (Nublat et al., 2001) mutants does not appear to be closely related to shoot levels of Na+. For example, the salt-sensitive mutant of arabidopsis, sos1, has a lower shoot Na+ content and lower Na+ influx than wild-type plants (at least in the presence of low K+) (Ding and Zhu, 1997).
This relationship between salt tolerance and low shoot Na+ does not hold for halophytes (for present purposes considered as those plants that show little growth reduction at NaCl concentrations of 300 mM or more). Some halophytes can accumulate extravagant quantities of Na+ in the shoot (up to 50 % of dry weight) without dying. The tendency to accumulate high concentrations of NaCl in the shoot is particularly associated with dicotyledonous halophytes, and in these plants NaCl may account for almost the total osmotic potential of the shoot (that is, NaCl is employed preferentially as an osmoticum; Flowers and Yeo, 1986; Glenn et al., 1999). Monocotyledonous halophytes tend instead to take up less Na+ and maintain higher K+ in the shoot than dicotyledonous halophytes, and osmotic balance is achieved partially by sugar synthesis. Interestingly, there seems to be large variation even within halophytic taxa with respect to strategies. For example, chenopods vary in the extent of Na+ uptake and K+ : Na+ ratios in the shoot (Reimann and Breckle, 1993), and mangroves may employ different strategies depending on whether or not they have salt glands (Atkinson et al., 1967).
Although dicotyledonous halophytes are often considered to ‘accumulate’ Na+ in the shoot, it is unlikely that halophytic species have higher rates of Na+ transport at high salinities than salt-sensitive plants. In fact, halophytes are much more effective at controlling uptake of Na+ than salt-sensitive plants (see ‘Pathways for initial entry into the root’). It is probably only the fact that they live longer than glycophytes at higher salinities that leads to the high concentrations of shoot Na+ often observed. Studies of halophyte ion transport at low salinities tend to obscure this, because many halophytes show a growth stimulation upon addition of NaCl to the growth medium due to accumulation of NaCl at moderate salinities (the growth optimum varies for different species, and may be as high as 400 mM NaCl or more).
Although NaCl accumulation in these halophytes are often higher than those in glycophytes at low salinities, uptake is not proportional to external salinity, and is curtailed at higher salinities (e.g. Glenn et al., 1999; Munns et al., 1999). By comparing external and xylem Na+ concentrations, Munns et al. (1999) estimated that at 200 mM external Na+, about 97 % of all Na+ presented to the root surface must be excluded—whether in a glycophyte or halophyte. Moreover, where it has been examined, it has been found that Na+ concentrations in halophytes are lower in growing tissues than in mature ones, implying a need, in common with glycophytes, to protect metabolically active cells from Na+ (Flowers and Yeo, 1986). This issue is discussed further below, but here we can say that evidence from studies of halophytes indicates that the ability to regulate Na+ uptake and transport to the shoot is critical in all plants. This seems to hold true even for halophytes able to secrete Na+ from the shoot. For instance, in two mangrove species, Na+ excretion via salt glands does not keep pace with Na+ uptake over a range of salinities, reducing plant growth as salinity increases (Ball, 1988).
Although halophytes probably do not accumulate salt any faster than glycophytes, they nevertheless end up with shoot concentrations of Na+ that are lethal to glycophytes. Thus, control of Na+ transport must operate at several levels: Na+ entry and supply to the shoot must be controlled, and the Na+ that reaches the shoot (and the interior of root cells) must be partitioned to avoid deleterious osmotic and cytotoxic effects. While there is evidence that some halophytes may have slightly modified forms of some cytosolic enzymes (Flowers and Dalmond, 1992), it is generally considered that high shoot concentrations of Na+ in halophytes imply compartmentation into the vacuole. The broad differences between halophytic monocotyledons and dicotyledons have been attributed to differences in water content and hence vacuolar volume. Dicotyledons may maintain higher shoot Na+ and Na+ : K+ ratios because they can store most of the Na+ in the vacuole and require relatively little K+ for cytosolic metabolism, whereas monocotyledons may have less Na+ storage capacity, and require more K+ and compatible osmotica for cytosolic compartments (Flowers and Yeo, 1988; Glenn et al., 1999).
In this review, we shall restrict ourselves to considerations of plant adaptations to moderate levels of salinity, which are those most likely to be of agricultural relevance. However, we will argue, where appropriate, that halophytes do not necessarily differ in kind from salt-sensitive species, but in general are more effective in employment of common mechanisms to avoid salt damage. Mechanisms to minimize damage from high salinity vary among plants, and several mechanisms must operate in a coordinated fashion to manage Na+ (Fig. 2). For example, plants can: (1) minimize initial entry; (2) maximize efflux; (3) minimize loading to the xylem or maximize retrieval before reaching the shoot; (4) maximize recirculation out of the shoot in the phloem; (5) maximize intracellular compartmentation or allocation to particular parts of the shoot (e.g. pith cells or old leaves); or even (6) secrete salt onto the surface of the leaf.
|
In this review, we shall separate adaptations involving processes common to all cells within the plant (‘cellular adaptations’) from those involving processes specific to particular cell types within the plant (‘whole plant adaptations’). We have organized this review in such a way because we consider that there are two quite distinct mechanisms of Na+ tolerance: (1) the tolerance of single cells to high salinity, involving, for example, intracellular compartmentation and damage repair; and (2) tolerance at a higher level than that of the single cell, involving, for example, control of long-distance transport and the site of Na+ accumulation within the plant. A review by Neumann (1997) also discusses these issues, and this division is also employed by Adams et al. (1992) when comparing responses of Mesembryanthemum tissue culture and intact plants. Both levels of process are clearly important, and for tolerance of a whole plant, the relative importance of these two mechanisms probably varies with species and conditions.
The need for multiple responses to salinity within a plant is indicated by the high rate of failure of programmes to regenerate stably salt-tolerant plants from salt-tolerant cells selected in tissue culture systems (Dracup, 1991; Winicov, 1998). Thus, salt-tolerant cells rarely regenerate to produce fertile salt-tolerant plants with heritable salt tolerance, consistent with cell-specific processes usually being more important than general cellular processes. The reverse finding is also pertinent, namely that calli derived from naturally occurring salt-tolerant plants do not reflect the tolerance of the plants from which they were derived (Smith and McComb, 1981).
Recent genetic manipulations showing increased tolerance resulting from constitutive overexpression of specific genes (e.g. Apse et al., 1999; Sugino et al., 1999; Gaxiola et al., 2001; Piao et al., 2001; Rus et al., 2001a; Zhang and Blumwald, 2001; Zhang et al., 2001; Pandey et al., 2002) raises questions about the general validity of this thesis. Although it is not known whether such manipulations do, in fact, generate salt-tolerant cells per se, they are altering expression in all cell types within a plant, indicating that manipulation of several processes at the cellular level can increase salinity tolerance of plants. It is interesting that such manipulations appear not to have been selected for when salt-tolerant cells were selected in tissue culture systems.
Cellular adaptations to high Na+
1. Intracellular compartmentation—with synthesis of ‘osmoprotectants’
The most direct way to maintain low cytoplasmic Na+ is to sequester it in vacuoles within each plant cell. To maintain equal osmotic potentials in vacuole and cytoplasm, solutes whose accumulation is not detrimental to cellular biochemistry must accumulate in the cytoplasm. Intracellular compartmentation can also be associated with succulence, to provide a larger volume of vacuole in which the Na+ can be stored.
Pumping Na+ into the vacuole.
In the process of intracellular compartmentation, Na+ enters leaf cells and is then pumped into the vacuole before concentrations increase in the cytoplasm. The pumping of Na+ into the vacuole is catalysed by a vacuolar Na+/H+ antiporter, the difference in H+ being initially established by H+-pumping ATPase and pyrophosphatase proteins (Blumwald et al., 2000). Na+/H+ antiporter activity can increase upon addition of Na+ (e.g. in barley roots: Gabarino and DuPont, 1989; in tomato roots: Wilson and Shannon, 1995; in sunflower roots: Ballesteros et al., 1997), and this induction was found to be much greater in the salt-tolerant species, Plantago maritima, than in the salt-sensitive species, P. media (Staal et al., 1991). Consistently, salinity did not induce tonoplast Na+/H+ antiport activity in salt-sensitive rice (Fukuda et al., 1998). This induction is reflected in increased transcript levels of some members of the arabidopsis AtNHX gene family which encode vacuolar Na+/H+ antiporters (Yokoi et al., 2002). Salinization also induces activity of both vacuolar primary H+ pumps, although this appears to occur in both Na+-tolerant and Na+-sensitive species (reviewed by Hasegawa et al., 2000; Maeshima, 2000).
The central importance of vacuolar sequestration has recently been underlined by experiments in which constitutive overexpression of vacuolar transporters has greatly increased salinity tolerance of a range of species. Overexpression of an arabidopsis vacuolar Na+/H+ antiporter (NHX1) increased salinity tolerance of arabidopsis (Apse et al., 1999), tomato (Zhang and Blumwald, 2001) and Brassica napus (Zhang et al., 2001), and overexpression of the native vacuolar H+-translocating pyrophosphatase gene (AVP1) increased salinity tolerance of arabidopsis (Gaxiola et al., 2001). The apparent K+-transporting activity of the arabidopsis Na+/H+ antiporter (Zhang and Blumwald, 2001; Venema et al., 2002) does not appear to cause perturbations in cytosolic K+ homeostasis, at least under the experimental conditions tested to date. Likewise, overexpression does not appear to have interfered with the possible roles for the tonoplast Na+/H+ antiporter in control of pH of the cytoplasm (Viehveger et al., 2002) or vacuole (Fukada-Tanaka et al., 2000).
The mechanisms underlying the enhancement of salinity tolerance by these transporters are not yet clear. Although overexpression of NHX1 in arabidopsis caused a small increase in shoot Na+ accumulation (Apse et al., 1999), further experiments at lower salinities, where control plants are still alive, are required to confirm this for tomato and B. napus. While increased uptake of Na+ to shoot vacuoles could facilitate osmotic adjustment (as for halophytes), enhanced activity of Na+ compartmentation mechanisms would also confer greater tolerance by reducing Na+ accumulation in the cytosol, without any change in cellular Na+ concentrations. This is possible because Na+ transport across the tonoplast is bidirectional and dynamic: a ‘leak’ of Na+ from the vacuole to the cytosol (discussed below) would require compensatory pump activity to maintain constant levels of cytosolic Na+. A higher level of tonoplast active Na+ transport may allow transgenic plants to exclude Na+ more effectively from the cytosol. It has also been suggested that overexpression of NHX1 could cause ‘spillover’ of the protein from the tonoplast to the plasma membrane pathway, resulting in enhanced cellular Na+ exclusion (Frommer et al., 1999). In the case of plants overexpressing the vacuolar pyrophosphatase gene AVP1, there was both increased accumulation of Na+ and enhanced tolerance compared with wild-type plants (Gaxiola et al., 2001), suggesting a significant role for vacuolar sequestration. Given the higher levels of Na+ measured in shoots of transgenic arabidopsis plants, these experiments suggest that it is not increased exclusion from the shoot per se that is important, at least in arabidopsis, but the ability to compartmentalize effectively the Na+ that reaches the shoot.
These results have several implications. First, it is remarkable that the overexpression of a single native gene product can produce such striking increases in salt tolerance. In arabidopsis, NHX1 and a number of related NHX-type genes are ubiquitously transcribed, and show up-regulation by drought and/or NaCl treatments (Yokoi et al., 2002). That a further constitutive increase in expression can cause an increase in salt tolerance begs the question of why breeding programmes have failed to select for such a radical increase in tolerance. It is possible that such a manipulation comes with a significant cost in yield, and that this has yet to be revealed in the glasshouse-based trials that have been undertaken to date.
Another implication of these results is that up-regulation of a component of tonoplast transport appears to have caused pleiotropic up-regulation of other genes (or of the activity of the gene products) to enable appropriate levels of Na+ accumulation to occur. For example, for overexpression of H+-translocating pyrophosphatase activity to cause greater accumulation of Na+ in the vacuoles (Gaxiola et al., 2001), increased vacuolar Na+/H+ antiporter activity must have occurred, as well as increased synthesis of osmoprotectants (see below). Elucidating the mechanism(s) for such up-regulation is an area of much interest—we do not even know at present if increased transcription or translation of the necessary genes is occurring, or whether all up-regulation is occurring at the post-translational level. In this context, it is noteworthy that constitutive overexpression of a protein kinase also increased both tolerance and Na+ accumulation in arabidopsis (Piao et al., 2001).
Another interesting output from these experiments is that plants are accumulating high levels of Na+ whilst retaining the ability to maintain high rates of growth. Thus, extraordinarily, such plants could prove able to reduce significantly the equilibrium concentrations of Na+ in saline soils. For example, given a dry matter production of 10 tonnes per hectare and a tissue Na+ concentration of 5 %, 0·5 tonnes per hectare of Na+ could be removed in above-ground biomass. This is equivalent to 10 % of the amount of Na+ per hectare in 1 m depth of soil containing 10 % water (on a volume basis) at a concentration of just over 200 mM Na+. If these plants can sustain high rates of growth whilst accumulating such high levels of Na+ in field situations, this does provide possibilities for a degree of phytoremediation of salinized lands. However, removal and disposal of the above-ground biomass would be expensive relative to the value of much salinized land, so this is probably a practical option only for high-value lands (and in cases where soil salt levels are not continuing to increase owing to aerial inputs, irrigation or rising saline water tables).
Na+ leaking back out of the vacuole.
In parallel with the pumping into vacuoles of Na+ is a leak back into the cytoplasm, probably through ion channels. Cation-selective channels, such as the so-called ‘SV’ and ‘FV’ channels, are particularly abundant in the vacuolar membrane of plant cells. Based on their selectivity, these channels are classified as non-selective cation channels (Demidchik et al., 2002), and are generally thought to be highly permeable to Na+ as well as other cations, such as K+. If open in intact cells, this could readily cause problems for plant cells, compromising their ability to sequester Na+ in the vacuole. However, because of the large number of interacting controls on the activity of these channels, their activity in vivo is still uncertain (Demidchik et al., 2002). Nevertheless, given their relatively high permeability to cytotoxic cations (such as Na+ and Ca2+), it must be presumed that their activity is low.
This presumption is supported by the observation that there are no major differences in the properties of vacuolar channels in salt-tolerant and salt-sensitive species of Plantago (Maathuis and Prins, 1990). Likewise, there were no particular features of the vacuolar membrane from leaves of the extreme halophyte, Suaeda maritima (Maathuis et al., 1992). One interesting observation, however, is that plants of Plantago that had been grown at high NaCl concentration had much reduced channel activity in their vacuolar membrane compared with those grown at low NaCl concentration (Maathuis and Prins, 1990), implying that the energetic cost of recycling leaked Na+ back to the vacuole may be significant.
To summarize, we know very little about the control of Na+ efflux from the vacuole, and thus know little about the possibilities for manipulation of the control of this pathway to increase the ability of cells to sequester Na+.
Synthesis of osmoprotectants.
As Na+ accumulates in the vacuole, osmotic potential in the cytoplasm must be balanced with that in the vacuole. This is done by the synthesis and accumulation in the cytoplasm of organic solutes that do not inhibit biochemical reactions. In fact, many of them appear to protect such reactions from increased levels of inorganic ions (e.g. Shomer-Ilan et al., 1991). Such ‘osmoprotectants’ (or ‘compatible solutes’) are often highly soluble, neutral or zwitterionic compounds and include secondary metabolites such as quaternary ammonium compounds (e.g. glycinebetaine) and polyols (e.g. mannitol), as well as core metabolites such as proline and sucrose (Hu et al., 2000).
Compatible solutes have been extensively researched and many transgenic plants have been made that over-accumulate these solutes (by stimulating synthesis, inhibiting breakdown or increasing transport to growing tissues). This work has been reviewed elsewhere (e.g. McNeil et al., 1999; Chen and Murata, 2002; Garg et al., 2002) and will not be discussed in detail here. With the notable exception of proline, transgenic manipulations do not usually lead to osmotically significant over-accumulation of metabolites (Chen and Murata, 2002), suggesting that the main mode of action of most manipulations made to date is not to balance osmotic potential in the cytosol, thus enabling increased vacuolar salt accumulation. Instead, increased osmoprotectant accumulation may be protecting plants against damage, such as helping maintain protein structure or increasing scavenging of reactive oxygen species. Nevertheless, account of the metabolic costs of such manipulations must still be made (Huang et al., 2000), and to minimize effects on the diversion of photosynthate to such solutes, their synthesis should be inducible.
Separating the effects of these compounds on salinity tolerance rather than drought tolerance has been difficult, especially when it is possible that the favourite model plant in which over-production has been generated (tobacco) may be more sensitive to the osmotic component of NaCl addition than to the ions themselves (Murthy and Tester, 1996). Nevertheless, in plants such as maize, there is a clear correlation between osmoprotectant production and salt tolerance (Saneoka et al., 1995). It seems highly likely that these compounds are effective both in maintaining a negative osmotic potential in the cytoplasm and in directly protecting protein and ribosome structures from deleterious effects of elevated cytoplasmic Na+.
Although osmoprotectants alone could go some way to enabling a plant to tolerate salinity, significant amounts of Na+ would still need to be compartmentalized. It is therefore likely that over-production of osmoprotectants would need to be accompanied by a pleiotropic up-regulation of Na+ pumps (such as must occur for compatible solute synthesis in transgenic plants with elevated vacuolar Na+/H+ antiporter activity; see above).
2. Tolerance of high cytoplasmic Na+
Cheeseman (1988) questioned whether Na+ is necessarily toxic in the cytoplasm, a point supported by biochemical evidence showing that cytoplasmic reactions can appear to be tolerant of quite high concentrations of Na+ (in the order of 100 mM) in the presence of compatible solutes, and even metabolic intermediates such as phosphoenolpyruvate (Shomer-Ilan et al., 1991). There is evidence for the tolerance, in vitro, of cytoplasmic enzymes to high Na+ in some salt-tolerant plants (e.g. of ribosomal activity; Flowers and Dalmond, 1992). However, because salt sensitivity of both ribosomes and enzymes can depend very much on relatively modest changes in the conditions used during the in vitro enzyme assays (e.g. Gibson et al., 1984; Shomer-Ilan et al., 1991; Solomon et al., 1994), the relevance of any assay to the situation in the intact plant is uncertain. Attempts to identify biologically relevant tolerance of enzymes to high Na+ would require complementation of in vitro assays with in vivo assays that also include concurrent monitoring of substrate levels and Na+ concentrations.
However, in halophytic bacteria, where the structure of some enzymes is altered to provide unequivocal tolerance to high Na+ (by, for example, having more acidic residues on the surface of the enzymes), enzymes become dependent on high concentrations of Na+ for activity (Dym et al., 1995). Thus, engineering of salt-tolerant enzymes in plants would not only require alteration of multiple proteins, but could reduce metabolic efficiency in non-saline conditions. Thus, a more desirable strategy would be to alter concentrations of protective substances that help maintain protein structure in the face of elevated concentrations of Na+ (see ‘Synthesis of compatible solutes’, and below).
3. Damage response and repair
Increased synthesis of a wide variety of proteins occurs in response to salt stress. Many of these, such as the osmotins and dehydrins, have properties similar to chaperones, and appear to be involved in the maintenance of protein structure in the face of elevated salts and other conditions in which protein–water interactions are disrupted (Ingram and Bartels, 1996; Campbell and Close, 1997). They may also act to stabilize membrane structure. These proteins are often highly hydrophilic and structured as random coils. Many of these stress-response proteins are probably elicited by the sudden osmotic shock of the experimental treatment. Others are probably involved mainly in plant responses to the osmotic component of high salinity. Detection of stress response proteins specific to Na+ damage requires suitable osmotic controls and could, perhaps, be distinguished, at least in the shoot, by a relatively late appearance in the time course of Na+ treatment (days to weeks after NaCl introduction) (Munns, 2002). Constitutive overexpression of a late embryogenesis abundant protein (LEA) from barley conferred tolerance to salt stress in transgenic rice (Xu et al., 1996), although similar effects upon addition of mannitol suggest that the LEA was protecting against the osmotic effects of NaCl, rather than the ion-specific component. Constitutive overexpression in tobacco of a heat shock protein from the halotolerant cyanobacterium, Aphanothece halophytica, increased tolerance to high salinity and reduced shoot Na+ concentrations, whilst having no measurable effect on unstressed plants (at least in the growth conditions employed in the experiments) (Sugino et al., 1999). However, the lack of osmotic controls makes it difficult to assess whether the effects were ion-specific.
The osmoprotectant glycinebetaine has also been suggested to act as a chaperone, as well as potentially reducing lipid peroxidation and protecting mitochondrial electron transport reactions (Chen and Murata, 2002).
Synthesis of polyamines (such as putrescine and spermine) increases greatly in response to stresses, including salinity (Galston and Sawhney, 1990; Mansour, 2000), and the increase in putrescine levels (especially) appears to be greater in salt-tolerant than salt-sensitive lines of tomato (Santa-Cruz et al., 1999). The function of these compounds is presumed to be protective, with a role in scavenging free radicals being a favoured hypothesis (Mansour, 2000). However, Lefevre et al. (2001) found an ion-specific increase in putrescine and tyramine concentrations which was higher in the roots of a salt-tolerant cultivar of rice (Pokkali) than in a salt-sensitive cultivar. However, the opposite trend was found in the shoots, perhaps reflecting different functions of these polyamines in roots and shoots.
Minimization of the generation of reactive oxygen species (ROS) as a result of inhibition of photosynthesis, and/or maximization of their removal (‘scavenging’) is likely to be an important component of plant responses to high salinity, among other stresses (Zhu, 2001). Interestingly, activities of enzymes involved in reactive oxygen scavenging were higher in a salt-tolerant species of tomato than in its salt-sensitive cultivated relative (Shalata and Tal, 1998). This suggests that ROS scavenging activity was part of the active salt tolerance mechanisms of this species, rather than a secondary response to damage arising from salt stress (such as mediated by the osmotic component of salt stress).
Reactive oxygen species are also scavenged by osmoprotectants, such as proline and mannitol, and the role of these and other ROS scavenging systems is reviewed elsewhere (Xiong et al., 2002). Although, intuitively, it might be more sensible to minimize damage rather than maximize its repair, it is notable that one of the few Na+-tolerant mutants identified to date appears to be tolerant owing to mutation of a negative regulator of enzymes involved in scavenging reactive oxygen species (Tsugane et al., 1999). Although elevated levels of ROS may be a result primarily of osmotic stress causing reduced stomatal conductance (unfortunately not directly tested in the paper), that the mutation also provided some protection against low concentrations of LiCl suggests there was a substantial ion-specific component to the effect.
The identification of a salt-sensitive mutant through a screen for UV-sensitivity suggests that components of salinity tolerance may also be related to processes of DNA repair (Albinsky et al., 1999).
4. Genomic-scale observations of alterations in expression of genes
In the current scramble to employ microarrays to study changes in gene transcription, a wealth of data is emerging on effects of salinity on gene expression (Bohnert et al., 2001; Kawasaki et al., 2001; Seki et al., 2001; Chen et al., 2002; Ozturk et al., 2002). Experiments need to be done in a physiologically realistic manner (e.g. to minimize the measurement of shock responses rather than adaptive responses to NaCl), and time courses will also be invaluable in enabling separation of primary damage limitation from secondary damage repair and acclimation. The relevance to salt-tolerance of studying responses in a salt-sensitive plant also needs to be considered. From the experiments published to date, it appears as if salinity affects the level of transcription of approx. 8 % of all genes. About 70 % of genes whose transcription is altered by salinity are distinct from those altered in response to drought stress, and distinct genes are activated depending on the tissue, developmental stage, and the rate and extent of onset of the stress (Bohnert, 2001). Consistent with this observation are data from differential display-PCR techniques suggesting that most salt-induced changes in gene expression are ABA-independent (Wei et al., 2000). Similar differential subtraction techniques to compare expression in wild-type and sos mutants of arabidopsis in control and stressed conditions identified 84 salt-regulated genes, six of which are controlled by the SOS signalling pathway (which is described in more detail below) (Gong et al., 2001).
Such data sets can be reconciled, at least in part, with older data reporting changes both in levels of protein expression in response to salinity and biochemical measurements of changes in enzyme activities. Interestingly, correlations made to date of responses in different varieties displaying different tolerances to salinity suggest that tolerance is most associated with the early responses to salinization, which include transcription of genes involved in signal transduction (Bohnert, 2001; Bohnert et al., 2001; Kawasaki et al., 2001). Such conclusions need to be reconciled with physiological experiments suggesting that physiological responses immediately after Na+ shock are of little relevance to long-term responses to Na+ (Munns, 1993, 2002). However, the treatments used in these microarray studies were extreme and could potentially lead to death of even tolerant plants over several days. For example, the description by Kawasaki et al. (2001) of irreversible wilting and death upon addition of 150 mM NaCl is not reassuring. As such, lack of differences after longer time intervals may reflect more the general effects of severe stress—more experiments need to be done using much lower concentrations of NaCl. The apparent lack of osmotic controls in many experiments also makes it difficult to separate Na+-specific effects from osmotic effects. Major bioinformatic analyses are now required to identify signalling pathways from such data and to predict likely biochemical consequences of these changes in transcription. Confirmation of such predictions with proteomic and biochemical studies directed by information gained from such transcriptomic studies will be necessary. Despite the likelihood that alternative splicing of mRNA may not be used as widely in plants as it is in animals (Arabidopsis Genome Initiative, 2000), the recent observation of the possible importance of pre-mRNA splicing in Na+ tolerance also needs to be considered (Forment et al., 2002).
5. Signalling pathways
The effects of salt stress include both salt-specific and osmotic components, the latter including effects mediated by abscisic acid. These pathways, from the receptor perceiving salt stress to alteration of protein activity and gene transcription via signalling intermediates and phosphoprotein cascades have been reviewed extensively elsewhere (e.g. Leung and Giraudat, 1998; Hasegawa et al., 2000; Schroeder et al., 2001; Xiong et al., 2002; Zhu, 2002). We shall add only a few comments to these more detailed accounts.
The relevance to field situations of many studies of salinity signalling may well be compromised because responses are often measured over a short time period after a rapid and large change in salinity. Thus, how much the responses relate to a shock response rather than physiologically relevant responses to elevated salinity is uncertain (see also Munns, 2002).
Cytosolic calcium activity.
Although the mechanisms of signal perception remain unknown, and may involve proteins such as stretch-activated channels and ion-specific receptors, it is clear that an early response to the sudden addition of extracellular Na+ is a transient rise in cytosolic Ca2+ (Knight et al., 1997). However, the relevance of such rapid rises to plant responses in a fairly constant saline environment is uncertain, although salinity-induced changes in activity of Ca2+ transporters and components of Ca2+-related signal transduction pathways suggest that Ca2+ plays a physiologically relevant role in plant responses to salinity. For example, InsP3-activated Ca2+ channel activity is induced by hypotonic pre-treatment (Allen and Sanders, 1994) and transcription of a gene encoding a phosphatidylinositol-specific phospholipase C also occurs upon salinization, consistent with a role for inositol phospholipid-based signalling responses in salinity responses (Hirayama et al., 1995). Up-regulation by salt stress of expression of an endoplasmic reticulum (ER) Ca2+-translocating ATPase (Wimmers et al., 1992) may also indicate an ongoing role for Ca2+ in plant responses to salinity, as does the effects of SOS3, a Ca2+-binding protein, on salinity tolerance (see ‘sos mutants’).
Protein phosphorylation and dephosphorylation.
There is extensive evidence for the central importance of protein phosphorylation and dephosphorylation in plant responses to salinity stress (Xiong et al., 2002). It is also possible that membrane-bound histidine kinases could act as Na+ sensors, as might occur in Synechocystis (Mikami et al., 2002). In arabidopsis, overexpression of a GSK1/shaggy-like protein kinase caused plants to show symptoms similar to those of salt stress in unstressed conditions (anthocyanin synthesis, transcription of some NaCl stress-responsive genes), but also increased tolerance to salt stress (Piao et al., 2001). These plants also accumulated higher concentrations of Na+ in the shoot, as is observed in plants overexpressing a vacuolar Na+/H+ antiporter (see ‘Pumping Na+ into the vacuole’). It is possible that the increased kinase activity directly or indirectly increases activity of this Na+ pump. Overexpression of a Dbf2 kinase also conferred tolerance to salinity (among other stresses) in arabidopsis suspension cells (Lee et al., 1999).
There are myriad kinases and phosphatases in plants likely to be related to salinity responses and/or tolerance. These include metabolic phosphatases (e.g. HAL2; Gil-Mascarell et al., 1999), Ca2+-dependent protein kinases (CDPKs), serine/threonine protein kinases (e.g. SOS2; Halfter et al., 2000), mitogen-activated protein kinase (MAPK) cascades (Kovtun et al., 2000; Tena et al., 2001), two component histidine kinases (Urao et al., 1999), Ca2+/calmodulin-activated serine/threonine-specific protein phosphatases (Pardo et al., 1998; Kudla et al., 1999), and so on. Determining how these interact to direct appropriate responses to salt stress will be a major undertaking requiring an integration of genomic and biochemical techniques with a framework of network analyses, such as speculated upon by Trewavas and Malho (1997) and Knight and Knight (2001).
The sos mutants.
The sos (salt overly sensitive) mutants are a set of recessive mutants with increased sensitivity to NaCl. They were identified by an inability to maintain root growth in elevated NaCl, and the defect appears to be NaCl-specific rather than osmotic because the mutants are specifically sensitive to NaCl and LiCl but not to mannitol. sos mutants have provided significant insights into components of signal transduction pathways that are likely to be involved in controlling plant responses to salinity. At least one of the proteins (SOS1) involved in the SOS signalling pathway shows putative cell-specific expression, indicating that these signalling pathways are involved in the regulation of whole plant Na+ transport. We will discuss the proposed SOS signal transduction mechanisms here, but will refer to cell-specific processes affecting plant distribution of Na+ in later sections.
SOS3 is a myristoylated calcium-binding protein that is thought to respond to salt-induced cytosolic Ca2+ elevations (Liu and Zhu, 1997; Ishitani et al., 2000). SOS3 interacts directly with SOS2, a serine/threonine protein kinase (Halfter et al., 2000). One of the targets of this signalling pathway is SOS1, an Na+/H+ antiporter localized to the plasma membrane (Shi et al., 2002). SOS2 appears both to increase SOS1 transcription and to increase the activity of the SOS1 protein through a direct protein–protein interaction (Shi et al., 2000; Qiu et al., 2002; Quintero et al., 2002). In addition to its role as an Na+ transporter, SOS1 has a large, putatively cytosolic domain which has been suggested to comprise an Na+ sensor, raising the possibility of feedback by SOS1 on the SOS signalling pathway (Zhu, 2002). Promoter-GUS fusions indicated localization of SOS1 transcripts in the epidermal and stelar cells of roots, suggesting a role in either removal of Na+ from or loading of Na+ into the xylem (Shi et al., 2002) or phloem.
sos1, sos2 and sos3 mutants demonstrate different phenotypes with respect to Na+ accumulation, indicating that SOS2 and SOS3 regulate the expression and/or activity of other Na+-related proteins in addition to SOS1 (Zhu et al., 1998). sos2 and sos3 mutants accumulate more Na+ than the wild type, whereas sos1 mutants accumulate less, whilst still showing strong symptoms of Na+ toxicity. The sos1, sos 2 and sos 3 mutants were initially identified in a screen for K+ transport mutants, and all show defects in K+ nutrition, being unable to grow normally in media with less than 20 mM K+. The interaction between Na+ toxicity and K+ deficiency in sos mutants has not yet been established, although the lack of K+ transport activity by SOS1 when expressed in yeast suggests that the effects of the sos1 mutation on K+ transport are indirect (Quintero et al., 2002). sos4 encodes a pyridoxal kinase that is important for the synthesis of pyridoxal-5-phosphate. It is speculated that this may be involved in regulation of SOS1 activity because SOS1 contains a putative pyridoxal-5-phosphate binding motif in the C-terminal cytoplasmic tail (Zhu, 2002).
It is curious that despite evidence for the existence of other components of the sos signalling pathway, only four sos mutants have been identified, despite the size of the screen (confirmed by the high frequency of identification of alleles of most of the four sos mutants) (Zhu et al., 1998). Moreover, a screen for knockout mutations complementing sos3 Na+ sensitivity identified multiple mutations of HKT1 and no other genes (Rus, 2001b). This may reflect the use of root growth assays over relatively short periods (7 d) to identify the sos mutants, which may not detect differences in Na+ transport that take longer to manifest and primarily affect shoot growth. For instance, sos2 mutants showed almost no root growth in 50 mM NaCl, yet showed 100 % survival in 100 mM NaCl, whereas sos3 mutants showed less than 50 % survival in 100 mM NaCl yet were considered less sensitive to NaCl than sos2 mutants on the basis of root growth inhibition (Zhu et al., 1998). sos1, sos2 and sos3 mutations all affect K+ nutrition, and sensitivity to inhibition of root growth by NaCl was correlated with tissue [K+] but not [Na+] (Zhu et al., 1998). A screen for survival, shoot size or leaf [Na+] at high salinity might pick out other mutants of relevance to Na+ toxicity and Na+ transport.
Related to the SOS pathway, the observation by Pandey et al. (2002) that increased expression of a protozoan calcium-binding protein in tobacco increased tolerance to salinity is also of interest, although how much of this tolerance is due ‘simply’ to increased vigour of the transgenic in control conditions is uncertain.
Transcription factors.
The response to high salinity requires the coordinated induction of transcription of many genes, thus requiring the activity of specific sets of transcription factors and their binding to specific sequences in the promoter regions of target genes (Chen et al., 2002; Xiong et al., 2002). Such promoter regions include dehydration-responsive elements (‘DREs’) and ABA-responsive elements (‘ABREs’), which are likely to be involved in plant responses to the osmotic effect rather than an Na+-specific effect. Various proteins binding to such elements have been identified using yeast one-hybrid screening (e.g. Liu et al., 1998; Uno et al., 2000). Constitutive overexpression of some of the genes encoding these proteins can induce strong constitutive overexpression of several stress-inducible genes, along with increased tolerance to abiotic stresses (Liu et al., 1998). Although this increased tolerance was associated with reduced growth in unstressed conditions, placing such genes behind a strong stress-inducible promoter reduced this growth inhibition in unstressed conditions (Kasuga et al., 1999).
Thus, by analysing information coming from microarray studies (e.g. Chen et al., 2002), there are many possibilities for future identification of promoter elements that are specific to controlling gene expression in response to salinity stress. Identification of transcription factors binding to these elements will then enable their inducible overexpression in transgenic plants, potentially increasing salinity tolerance.
Cell-specific signalling responses.
It is also important in these studies to try to differentiate responses that are relevant to a positive response of the plant towards tolerance of salinity, as opposed to responses that are related to a negative spiral towards death. To put this another way, genes related to damage limitation are more important than those related to a plant’s attempts at damage repair. Such differences are best separated by the judicious use of time courses and different concentrations of NaCl. Another approach could be to identify responses in particular cell types, with cell type-specific responses being more likely to be involved in damage limitation, whereas constitutive responses are more likely to be related to damage repair.
In this review, signalling has been discussed in the context of cellular adaptations to salinity because, at present, very little is known about cell-specific signalling responses to salinity. Most workers look at gene expression, Ca2+ changes, etc. in whole organs, or even whole plants. The study by Kiegle et al. (2000) is the only one of which we are aware in which an attempt has been made to compare responses in different cell types, although this was limited to measurements of changes in cytosolic Ca2+ in response to the sudden addition of large amounts of Na+. Thus, although distinctive oscillations in cytosolic Ca2+ were observed in endodermal and pericycle cells, which have distinct functions in plant responses to salinity (as explained below, see ‘Control of xylem loading’ and ‘Retrieval from the xylem’), more refined experiments must be performed to test the relevance of such observations to physiologically realistic conditions.
Whole plant adaptations to high salinity
In addition to features of every cell within the plant that promote cellular survival and thus contribute to the tolerance of the whole plant to salinity stress, plants can also have a wide range of other mechanisms that involve particular activities of specific cell types. These generally involve the control of uptake of Na+ and of its distribution within the plant (Fig. 2). This includes processes such as (1) regulation of Na+ delivery to the shoot, a process comprising several steps, namely: initial entry into root epidermal and cortical cells, a balance between influx and efflux; loading to the xylem; retrieval from the xylem before reaching the shoot (and storage in, e.g. pith cells). (2) Recirculation out of the shoot in the phloem. (3) Allocation to particular parts of the shoot (e.g. old leaves). (4) Secretion onto the surface of the leaf. (5) Control of transpiration. Any or all of these mechanisms may operate in a particular plant. We shall now consider each in turn.
1. Regulation of Na+ transport to the shoot
Many of the mechanisms that enable plants to tolerate high soil salinity involve maintenance of low Na+ in shoots, as evidenced by the frequently observed inverse correlation between amounts of shoot Na+ and growth. Within crop species, it is often the case that varieties that are more tolerant of salinity have smaller amounts of salt reaching the shoot, or at least accumulating in the shoot. However, this relationship does not necessarily hold among species; for instance, barley and cotton are relatively salt-tolerant crop species, and tolerate higher shoot Na+ concentrations than, for instance, wheat.
Concomitant with the maintenance of low shoot Na+ is the maintenance of high K+. In fact, it is possible that a high K+ : Na+ ratio is more important for many species than simply maintaining a low concentration of Na+ (Gorham et al., 1990; Dubcovsky et al., 1996; Maathuis and Amtmann, 1999; Cuin et al., 2003), which makes sense given that much of the basis for Na+ toxicity is due to competition with K+ for K+-binding sites (see ‘The main bases for Na+ toxicity’). In this context, it is interesting that the sos1, sos2 and sos3 mutants are defective in K+ nutrition as well as Na+ transport. However, we will not discuss K+ specifically, except when it interferes directly with Na+.
Pathways for initial entry into the root.
It is clear that initial entry of Na+ from the soil solution into the root cortical cytoplasm is passive (Cheeseman, 1982), with Na+ influx being favoured energetically by differences in both concentration and voltage. Net accumulation in the plant is due to the balance between passive influx and active efflux. Exclusion of Na+ from the root at the site of initial entry has been suggested to be an important mechanism to minimize damage in saline soils for some plants (e.g. Schubert and Läuchli, 1990).
An important feature of the initial unidirectional entry of Na+ into roots is the very high rate of influx that has been measured. Unidirectional influx has been recorded as being in the order of 0·5–2·0 µmol g–1 f. wt min–1 with 50 mM external Na+, a value found in species as diverse as wheat (Davenport, 1998), rice (L. Wang, R. Davenport and M. Tester, unpubl. res.) and arabidopsis (P. A. Essah, R. Davenport and M. Tester, unpubl. res.). This value is somewhat higher than those published previously (e.g. Zidan et al., 1991; Elphick et al., 2001), probably because the previous measures were influenced by significant efflux (due either to long periods of uptake or long rinse periods). We have undertaken high resolution time courses in each species, and these show clear non-linearities in influx in 5–10 min. Thus, very short exposures to 22Na+ are necessary to obtain an accurate measure of influx (Fig. 3). To put this high rate of influx into perspective, it can be calculated that without compensatory fluxes to the apoplast or vacuole, the cytoplasm of a cell containing no Na+ would equal an extracellular concentration of 50 mM within 3 min of being exposed to that concentration (assuming a cell diameter of 40 µm and 10 % of the cell volume being cytoplasm). This corresponds to the range of time over which non-linearity in influx occurs, consistent with the need for very short influx times.
|
By using the very rough conversion of Pitman (1963) (that 1 µmol g–1 h–1
3 nmol m–2 s–1), an influx of 1 µmol g–1 root f. wt min–1 equates to a flux per unit area of root plasma membrane of 180 nmol m–2 s–1, or a current density of 34 mA m–2. This is of a similar magnitude to Na+ inward currents at physiologically reasonable potentials measured in patch clamped protoplasts isolated from roots of maize (Roberts and Tester, 1997), wheat (Tyerman et al., 1997; Buschmann et al., 2000), rye (White and Lemtiri-Chlieh, 1995) and arabidopsis (Demidchik and Tester, 2002). These high rates of unidirectional influx of Na+ do not result in rapid accumulation of Na+. Root Na+ concentrations tend to remain fairly constant over time in saline conditions, and shoot Na+ content tends to rise rather slowly. Thus, the net rate of uptake of Na+ is usually much lower than the unidirectional influx measured by short-term uptake of radioactive tracers or patch-clamp measurements, implying substantial efflux of Na+ across the plasma membrane. This efflux must be energetically expensive (see ‘Na+ efflux pathways’, below).
Interestingly, the few measurements of which we are aware suggest that unidirectional influx into halophyte roots is much lower than that into roots of glycophytes. Cheeseman et al. (1985) measured 22Na+ uptake into