There is now strong evidence that the plant hormone abscisic acid (ABA) plays an important role in the regulation of stomatal behaviour and gas exchange of droughted plants. This regulation involves both long-distance transport and modulation of ABA concentration at the guard cells, as well as differential responses of the guard cells to a given dose of the hormone. We will describe how a plant can use the ABA signalling mechanism and other chemical signals to adjust the amount of water that it loses through its stomata in response to changes in both the rhizospheric and the aerial environment. The following components of the signalling process can play an important part in regulation: (a) ABA sequestration in the root; (b) ABA synthesis versus catabolism in the root; (c) the efficiency of ABA transfer across the root and into the xylem; (d) the exchange of ABA between the xylem lumen and the xylem parenchyma in the shoot; (e) the amount of ABA in the leaf symplastic reservoir and the efficiency of ABA sequestration and release from this compartment as regulated by factors such as root and leaf-sourced changes in pH; (f) cleavage of ABA from ABA conjugates in the leaf apoplast; (g) transfer of ABA from the leaf into the phloem; (h) the sensitivity of the guard cells to the [ABA] that finally reaches them; and lastly (i) the possible interaction between nitrate stress and the ABA signal.
It has long been known that in order to conserve the water, nutrients and carbohydrates required for survival, plants respond to stresses such as soil drying by reducing leaf expansion and closing stomatal pores. In addition, root growth rates may be maintained in order that the plant may continue to access water. Traditional explanations for drought-induced regulation of gas exchange and leaf growth have emphasized the importance of the decline in shoot water status, which commonly accompanies severe soil drying (e.g. Kramer 1969). It is now accepted, however, that many of the plant’s responses to soil drying can occur in the absence of changes in shoot water status, via chemical signals. In one very clever series of experiments, pressure has been applied to roots to counteract the increasing soil suction that occurs as soil dries. This generated shoot water relations in droughted plants that were comparable to those of well-watered plants. Despite this, rates of gas exchange and leaf growth were still restricted compared to rates shown by well-watered plants (e.g. Gollan, Schurr & Schulze 1992).
Another apparent demonstration of root-sourced chemical signalling lies in the work of Gowing, Davies & Jones (1990). Here, the roots of young apple trees were split between two containers, the soil in one of which was allowed to dry, while the other container was irrigated as normal. This treatment restricted the rate of leaf growth, which could be restored to control rates by removing the roots in contact with the drying soil, i.e. by removing the source of the chemical signal. Blackman & Davies (1985) used the same technique to demonstrate that chemical signals sent from drying soil could close the stomata of wheat leaves.
It is important to note, however, that several groups have used the root pressure vessel technique to restore the shoot water relations of a range of species growing in dry soil, and have demonstrated that stomata re-open (e.g. Saliendra, Sperry & Comstock 1995; Fuchs & Livingstone 1996; Comstock & Mencuccini 1998). These results indicate that drought-induced limitations in stomatal opening in these species are mostly hydraulic, suggesting that the relative importance of hydraulic versus chemical signalling may differ between species. It might be surprising if this were not the case, but we should exercise care in using the root pressurization technique to differentiate between chemical and hydraulic signalling mechanisms, as pressurization of plant parts is known to change the pH of different compartments of the shoot and root (Hartung, Radin & Hendrix 1988). We will see below the possible significance of such a change on the modification of the signalling process, which raises the possibility that root pressurization can minimize both hydraulic and chemical limitation of stomatal behaviour simultaneously.
It has long been apparent that the major plant growth regulating hormone ABA (see Addicott & Carns 1983) strongly promotes stomatal closure (Jones & Mansfield 1970), and can often inhibit shoot growth. The synthesis of ABA is stimulated by the dehydration of plant cells (Wright 1977) including root cells. Leaf cells also synthesize ABA (Cutler & Krochko 1999) and leaf dehydration caused by severe soil water shortages massively increases bulk leaf [ABA], which often correlates well with stomatal closure. If we are to argue that ABA acts as a long-distance chemical signal which can provide information on water availability in the soil, it is important to be able to differentiate between such shoot-water deficit-induced increases in leaf ABA, and those that arise due to ABA import from the root. Zhang & Tardieu (1996) showed that all tissues of maize roots, regardless of their age and position, could synthesize ABA when dried. Much research has now shown that the amount of ABA in the xylem sap can increase substantially as a function of reduced soil water availability, and that this increased delivery to shoots can increase ABA concentrations in different compartments of the leaf (e.g. Loveys 1984; Zhang & Davies 1989). In order to ascribe an important stress detection function to the roots we need to show that the delivery of ABA by the xylem is sufficient to bring about the observed degree of stomatal closure seen in response to soil drying. This is only sometimes the case, but for reasons that will be explained below we can still argue that ABA plays an important regulating role in the droughted plant in response to perturbations at the root, even when xylem ABA concentrations are not increased at all.
Increases in the ABA concentration in the leaf in response to perturbations around the root are now well documented (see Davies & Zhang 1991). The xylem vessels give up their contents (including their ABA) to the leaf apoplast (Weyers & Hillman 1979), thereby increasing the concentration of the hormone in this compartment. The ABA is carried with the transpiration stream inside the leaf around and/or through the mesophyll cells (see below) so that it reaches the target cells (the stomatal guard cells) in the epidermis, which contain ABA receptors with external (and possibly internal) loci (see below) in their plasma membranes (pm). Once bound, the hormone then induces an internal signal transduction cascade usually involving increases in both externally and internally sourced cytoplasmic calcium, which eventually reduces guard cell osmotic potential via loss of K+ and Cl– to cause stomatal closure (for reviews see McAinsh, Brownlee & Hetherington 1997; Assmann & Shimazaki 1999). Although the ABA receptor has yet to be unambiguously identified in plants (Assmann & Shimazaki 1999) there is much evidence in the literature for an apoplastic locus of ABA perception, indicating that the apoplastic fraction of ABA will be the major determinant of stomatal aperture (Hartung 1983; Hornberg & Weiler 1984). Anderson, Ward & Schroeder (1994) demonstrated a failure of direct microinjection of ABA into the guard cells of Commelina communis L. to cause closure (but see Allan et al. 1994 and Schwartz et al. 1994). In addition, external ABA has been found to stimulate inward Ca2+ conductance and to induce elements of the IP3 signalling cascade inside guard cells that leads to the release of Ca2+ from internal stores (see MacRobbie 1997). Zhang & Outlaw (2001a) found that stressing Vicia faba L. roots could change the ABA concentration at the guard cell apoplast (2·3 µm) in the absence of bulk apoplastic changes in [ABA] (185 nm), and that the apoplastic guard cell ABA fraction correlated with changes in stomatal aperture more effectively than the guard cell symplastic fraction. Even more recently, ABA supplied via the xylem to non-stressed V. faba was seen to accumulate at the guard cell apoplast without a concomitant increase in symplastic guard cell ABA, and this correlated with stomatal closure in the absence of any other stress-induced signals (Zhang & Outlaw 2001b). These data show that apoplastically-facing guard cell ABA receptors seem to be important in the responses to very subtle ‘stress’ signals experienced by plants. Such signalling mechanisms may not be important when shoot water stress develops. Harris et al. (1988) found that when severe leaf water deficit develops, most ABA accumulated in the guard cell symplast, and it may be that ABA receptors with an internal locus (Allan et al. 1994; Schwartz et al. 1994) only operate when shoot water relations are also affected.
Modification of the xylem aba signal
Although we can accept that xylem ABA will modify leaf functioning, it is clearly not just a simple matter of the [ABA] in the xylem being that perceived by the target cells, given that the ABA concentration in apoplastic microsites around the guard cells is that to which stomata respond. In fact, there is often great variation in the apparent sensitivity of leaf conductance to a given concentration of ABA in the xylem stream (e.g. Correia & Pereira 1995). The reasons for these findings begin to become clear in the work of Trejo and colleagues (Trejo, Davies & Ruiz 1993; Trejo, Clephan & Davies 1995). Their data confirmed early work by Incoll & Jewer (1987a, b) that stomata in isolated epidermis were often sensitive to concentrations of ABA as low as 10−9 M, much lower than that found in the xylem sap even of well-watered plants. If the stomata were directly exposed to the concentrations of ABA normally seen in well-watered xylem sap (usually between 1·0 and 50 nm, e.g. Schurr, Gollan & Schulze 1992) they would be permanently closed. So the cells of the leaf must ‘filter out’ some of the ABA arriving in the transpiration stream before it reaches the stomata in the epidermis. By inhibiting the catabolic degradation of ABA in the mesophyll, Trejo et al. (1993) found that the amount of ABA reaching the epidermis was increased, and that the concentration of ABA supplied in the xylem exerted an increased effect on stomatal aperture. In other words the mesophyll normally removes (by metabolism and sequestration) much of the ABA passing through it in the transpiration stream before it reaches the guard cells. This would also explain why the same authors noted that there was a linear relationship between the [ABA] in the epidermis and the stomatal response, but a very poor relationship between bulk leaf ABA and stomatal aperture. Daeter & Hartung (1995) demonstrated that the epidermal symplast is an even better sink for apoplastic ABA than the mesophyll, because it has a catabolic rate that is five-fold greater. So once the ABA has reached the epidermis, there is still potential for ABA to be removed from its site of action (the guard cell apoplast). The importance of the contributions of ABA sequestration and catabolism in the symplast is made clear by the work of Kefu, Munns and King (1991). These authors calculated that the amount of ABA carried in the transpiration stream of cotton plants each day is nine times in excess of the amount of ABA actually detected in the leaves at the end of the day.
In the past it has been argued (e.g. Jackson 1993) that as the transpirational flux increases, for example with increasing VPD or temperature, the delivery of ABA to the leaf will increase, and that stomata will respond to the total flux of ABA into the leaf rather than to the concentration in the xylem stream. However, the fact that the leaf has a large reservoir (the mesophyll cells) in which to sequester and/or catabolize ABA entering via the xylem could explain why several lines of evidence have failed to support this theory. For example, Trejo et al. (1995) found that increases in VPD and temperature (19·9–36·1 °C) increased transpirational water loss from Phaseolus acutifolius L. leaves, but did not affect stomatal aperture. This was despite an increased delivery of ABA, as evidenced by greater than two-fold increases in bulk leaf [ABA]. Presumably the leaf symplastic reservoir for ABA sequestration is large, and only when the reservoir is full and/or catabolic enzyme activity is saturated, will increases in flux increase the [ABA] seen at the guard cells (see below). Some related work was carried out by Wilkinson, Clephan & Davies (2001). This showed that the [ABA] penetrating to the abaxial epidermis of detached C. communis leaves supplied with a concentration as high as 10−6 M via the transpiration stream, was kept uniformly low as the temperature was increased from 11 to 28 °C. As temperature, transpiration flux and ABA flux into the leaf increased, the rest of the leaf acted as a reservoir storing the extra ABA entering the leaf tissues and keeping it away from the epidermis (Fig. 1). The change in ABA flux with temperature (except between 7 and 11 °C) failed to change the epidermal [ABA].
Movements of ABA between the xylem and the leaf symplastic reservoir provide a powerful point at which the ABA stress signal response can potentially be modulated. Xylem sap/apoplastic sap pH changes are one way in which accessibility to the mesophyll (and epidermal) reservoir can be controlled. Evidence that xylem and apoplastic sap pH can increase in response to soil drying has been reviewed by Wilkinson (1999). It is important to note that sap pH changes can occur even when the shoot water status of plants in drying soil is maintained using the root pressure vessel (Schurr et al. 1992) or under partial root drying (PRD, Wilkinson & Davies, unpublished; Stoll, Loveys & Dry 2000). Schurr et al. (1992) found that correlations between xylem sap ABA from droughted sunflowers and reductions in stomatal conductance were poor, but when sap pH or sap [nitrate] or sap [calcium] were also taken into account the correlation was improved (Gollan et al. 1992). The authors proposed that increasing xylem sap pH reduced the ability of the leaf symplastic compartment to sequester ABA, so that during drought more of the ABA that enters the leaf will ultimately reach the guard cells in the epidermis. This proposal was based on very detailed information available in the literature, demonstrating that the extent of ABA uptake by all leaf cell types is reduced when the pH outside the cell is increased (e.g. Heilmann, Hartung & Gimmler 1980; Kaiser & Hartung 1981). Models were generated by this group, based on measurements of ABA uptake in response to pH in isolated cells and tissues. These predicted that the drought-induced pH increases detected in vivo would be enough to cause stomatal closure in an intact leaf, by preferentially concentrating the ABA that already exists in the leaf into the apoplastic as opposed to the symplastic compartment (Slovik & Hartung 1992a, b). It was not even necessary to incorporate an increase in xylem-sourced [ABA] into the model for stomatal closure to occur. These predictions have since been substantiated to some extent by Wilkinson & Davies (1997) and Wilkinson et al. (1998) in C. communis and tomato. We were able to conclude that as well as potentially magnifying an effect of xylem ABA on stomatal closure during drought, a drought-induced increase in xylem sap pH alone could constitute a root-sourced signal to the leaf, inducing stomatal closure. In essence, well-watered xylem sap has a pH around 6·0 and sap from droughted plants has a more alkaline pH of around 7·0 (see Wilkinson 1999). Artificial xylem sap buffered to pH 7·0 supplied to whole detached C. communis or tomato leaves via the xylem stream reduced stomatal aperture in comparison to controls supplied with pH 6·0 buffer. The reduction of transpirational water loss by high pH in this system was found to require absolutely the presence of a very low ‘well-watered concentration’ of ABA (c. 10−8 M) in the xylem stream, a concentration that does not affect stomatal aperture at control pH. The transpiration rate of the ABA-deficient tomato mutant flacca could not be reduced by increasing xylem sap pH up to 7·75 (highest tested), but the inclusion of a ‘well-watered’ concentration of ABA in the artificial sap buffer restored the pH response seen in the wild-type. It was concluded that the increased pH caused stomatal closure by allowing more of the ABA that enters the leaf to penetrate to the stomatal guard cells as a result of a pH-induced reduction in sequestration into the symplastic reservoir as described above.
We now know that soil drying-induced modulation of ABA sequestration into the leaf symplast via xylem-sourced pH changes is only one example out of many whereby the ‘basic’ ABA signal can be influenced by the environment around the plant. Further examples are given below.
Whole-plant modulation of the ABA signal
ABA may gain entry to plant roots in the following ways: it can be synthesized in the root (or released from conjugated forms), it can be taken up from soil water surrounding the root, or it can be delivered to the root from the shoots by the phloem. Once inside the root the ABA can be taken up by the symplast and stored or degraded, or it may be transferred from cell to cell towards the xylem vessels, or carried apoplastically with the transpiration stream towards the xylem. Alternatively, ABA may be lost from the root by diffusion into the soil water (rhizosphere). If root ABA is not degraded, stored in the symplast or lost into the soil water it is transferred into the xylem vessels for transport to the leaf. ABA loading into the xylem can either occur directly from the apoplast, or indirectly via specialized xylem parenchyma cells found along the length of the root xylem. There are several points along this root ABA uptake and transport pathway that can be influenced by the environment such that the strength of the root-sourced ABA signal in the xylem can be enhanced or reduced. These are described below.
Root cells constitutively synthesize a low ‘well-watered’ amount of ABA, and as described above this rate of synthesis is massively increased by root tissue dehydration, classically associated with soil drying. However the strength of this ‘basic’ ABA signal sent from roots to shoots can be modified in other less obvious ways. Firstly the soil around the roots may be influential, not only as regards the direct effects of its moisture or nutrient content (see last section). Hartung et al. (1996) suggested that when rhizospheric ABA is high (for example at low soil water contents) the outwardly directed concentration gradient for ABA between the root and the soil could be reduced. This makes it less likely that root ABA will diffuse out of the plant, potentially strengthening the xylem ABA signal further along the transpiration stream. In addition, different soil types are characterized by different pH values, and as described above ABA tends to accumulate in alkaline compartments as a result of the anion trapping effect (see below and Wilkinson 1999). Degenhardt et al. (2000) have shown that in alkaline soil substrates considerable portions of the ABA synthesized in the roots can be leached out into the soil solution, although species with an exodermis (e.g. Zea mays) were less susceptible to this type of stress. In this way alkaline soil could weaken the root-sourced ABA signal.
More usually though, rhizospheric ABA is transported into the root with the soil water. Water flow into and across the root (see Steudle 2000) is a result of the tension created by the transpiration stream (in which case water flows apoplastically), or of the decreased osmotic potential of the xylem sap in the absence of transpiration from the leaf (in which case water flows from cell to cell). Anything that increases water flux across the root should promote the transfer of ABA across this organ towards the xylem, because the uptake and radial transport of ABA occurs alongside that of water (Freundl, Steudle & Hartung 2000). However, until recently this prediction could not have been made, as it was believed that radial transport of ABA across roots was entirely trans-cellular due to the perceived impermeability of the exo- and endodermis. This would have meant that increased flux of apoplastic water into the xylem upon increased transpiration would dilute the xylem [ABA] (e.g. Else et al. 1994). However, Freundl, Steudle & Hartung (1998, 2000) found that filtration of ABA from the apoplastic flow of water as it crossed the endodermis built its local concentration up high enough to cause ‘solvent drag’ through this potential barrier into the xylem. This is has been termed the ‘apoplastic by-pass’ for ABA.
Different species transport root water to different extents within each pathway (symplastic or apoplastic), and when the symplastic route is necessarily dominant (e.g. in heavily suberized roots), root cell plasma membranes contain large concentrations of water channels or ‘aquapoprins’ which can be actively regulated. Interestingly, ABA itself can increase the flow of water into and through the root, otherwise known as its hydraulic conductivity (Hose, Steudle & Hartung 2000). This is believed to result from ABA-induced opening of these inwardly directed water channels (Tyerman et al. 1999; Netting 2000). Soil drying and salinity have been found to increase root hydraulic conductivity, and the induction of ABA biosynthesis in the root by these stresses may effectively maximize the transport of the newly synthesized ABA within the root towards the xylem (positive feedback). As described above, water and ABA flux across the root is also regulated by the osmotic potential of the xylem sap. Again, ABA itself may reduce the xylem osmotic potential and thereby increase water and ABA flux across the root, by inducing ion loading into the xylem, e.g. Glinka (1980)– potentially another positive feedback mechanism to maximize the root-sourced ABA signal.
Because root cells may degrade or store the ABA that they synthesize or take up as it flows past them, not all the ABA that enters the root reaches the xylem vessels. Conditions that reduce the ability of root cells to degrade or take up and store ABA will increase the amounts that penetrate to the xylem. The catabolic degradation of ABA takes place constitutively in most plant cell types (see Cutler & Krochko 1999 for review), but there is some evidence that soil drying reduces the rate of this process in roots. Liang, Zhang & Wong (1997) found that the half-life of 3H-ABA fed to maize roots was extended from 1·15 to 02·27 h by drying the soil around them. Thus, more of the fed ABA could potentially penetrate to the xylem.
Soil drying also leads to changes in the pH of the various compartments of the root. That of the apoplast increases, and root dehydration causes the cell cytoplasm to acidify (Daeter, Slovik & Hartung 1993). This has multiple effects on the root ABA signal, all based on the fact that ABA always becomes concentrated in the most alkaline compartments, and all leading to an increase in the amount of ABA that finally reaches the xylem, by at least two or three-fold (Slovik, Daeter & Hartung 1995). Firstly, these pH changes inhibit ABA loss from the root to the rhizosphere. Secondly, all shoot-sourced ABA arriving at the root in the phloem is compartmentalized in the apoplast. Both Liang et al. (1997) and Neales & Mcleod (1991) have reported that root dehydration gives rise to increases in phloem-sourced ABA in the xylem sap of 25–30%, not a negligible amount. Finally, cytoplasmic acidification causes ABA synthesized in the cytoplasm or previously stored there to be released to the apoplast (see below). The fact that more root ABA becomes compartmentalized in the apoplast as a result of drought-induced pH changes means that it is less likely to be catabolized or simply stored, and a greater percentage finally reaches the xylem vessels.
To summarize, a soil drying ABA signal classically associated simply with the induction of synthesis in the root can be intensified by the simultaneous effects of dry soil to reduce ABA catabolism, to prevent rhizosphere- and phloem-sourced ABA from entering the symplast, to induce more effective efflux of the synthesized ABA to the apoplast, and potentially to increase water and therefore ABA flux towards the xylem.
Once ABA has reached the xylem vessel its transfer into this conduit can also be modulated by the environment. Xylem vessels have the ability to modify the pH of solutions passing through them (Clarkson, Williams & Hanson 1984). Fromard et al. (1995) showed that H+-pumping ATPases are much more concentrated in the plasma membranes of xylem parenchyma cells than in any other plant cell type. They showed that these pumps were involved in the control of vascular sap pH, and that this control fluctuated with season. Xylem pH values were close to neutrality in winter (when ATPase activity was low), whereas they were acidic (pH 5·5) at the beginning of spring. Hartung & Radin (1989) reported that plant water stress could reduce the activity of these H+-ATPases associated with the root xylem, and suggested that this may be one of the causes of the increased alkalinity often observed in this compartment under stress (see below for other hypotheses). Increases in xylem pH enhance ABA loading to the root xylem by increasing the capacity of the xylem sap to trap anions.
The xylem stream
As described above, soil drying can increase the pH of the xylem stream. Recently, work by Hartung’s group determined that not only will this pH change maximize the initial amount of ABA loaded into the xylem at the root, but also that ABA loading into the xylem lumen from stores found to exist in the stem parenchyma occurred more effectively (Hartung, Sauter & Hose 2002). However, this only occurred when the initial xylem [ABA] loaded in at the root was low, i.e. it will be an early response to soil drying. Nevertheless, the possibility exists that the ABA concentration arriving at the leaf in the xylem stream may be greater than that loaded in at the root, and this effect may be intensified in drying soil.
It is described above how ABA entering the leaf via the xylem does not all reach the guard cells because of sequestration into the symplastic ‘reservoir’, and that soil drying-induced changes in xylem and therefore in apoplastic sap pH modulate the rate of this process. However, several additional stress-induced phenomena occur in the leaf to intensify the root-sourced ABA signal that finally reaches the epidermis. One, or a combination of the following, could explain why the extent of stomatal closure observed in droughted plants is often greater than can be attributed alone to the concentration of ABA found in the xylem (e.g. Correia & Pereira 1995).
A full symplastic reservoir. Under severe drought, shoot water potentials drop and ABA biosynthesis occurs in the cells of the leaf as well as in those of the root. This means that the leaf symplast contains a great deal more ABA than that of a turgid leaf. This rapidly alters the ABA concentration at the guard cells not only as a result of the penetration of this newly synthesized ABA, but also because ABA entering the leaf at the xylem cannot enter the symplastic reservoir and therefore ends up at the epidermis. Filling of the symplastic reservoir therefore increases the sensitivity of the stomata to the ABA coming from the xylem. For example, when xylem ABA concentrations return to well-watered values after a period of water deficit, stomatal re-opening is often sluggish (e.g. Trejo et al. 1995), and this may be because the symplast is still full of ABA. Thus, even the low well-watered concentrations of ABA carried by the xylem may be capable of maintaining stomata in the closed state until normal rates of symplastic sequestration are restored. In related work, Heckenberger, Schurr & Schulze (1996) found that when ABA influx to sunflower leaves was large, either due to long-term feeding of low concentrations or short-term feeding of high concentrations via the xylem, stomatal opening after the cessation of feeding took longer to occur. ABA flux into the leaf may therefore be an important determinant of stomatal aperture under severe but not incipient drought stress, when the capacity of the symplastic reservoir to remove ABA from the apoplast is minimized by a reduction of the inwardly-directed pm concentration gradient for ABA.
Loss of ABA from the symplast. The symplastic ABA reservoir in the leaf may also lose its ability to hold on to the ABA that it contains. The occurrence of ABA release from root cells has already been described above. There is plenty of evidence in the literature that leaf cells are also capable of releasing ABA to the surrounding medium, and Hartung and co-workers have suggested that drought-induced increases in xylem sap pH cause leaf cells to release ABA to the apoplast. This could cause stomatal closure in the absence of any increase in bulk leaf ABA concentration, as has often been described (references in Wilkinson & Davies 1997). However, Wilkinson & Davies (1997) could find no evidence for ABA efflux from tissue isolated from turgid leaves in response to an increase in pHext. On the other hand, when leaves become dehydrated, the cell cytoplasm is acidified. Hartung, Kaiser & Burschka (1983) showed that wilting induced the efflux of 14C-ABA from the mesophyll of preloaded Valerianella locusta leaf tissue so that it was able to penetrate to the lower epidermis. Hartung & Radin (1989) showed that cotton leaf dehydration reduced the activity of outwardly rectifying pm H+ATPases, which would acidify the symplast. This mechanism was similar to, but separate from that involved in the alkalinization of the xylem sap of plants in drying soil referred to above. Kaiser & Hartung (1981) demonstrated that the efflux of ABA from the cytoplasm of isolated mesophyll cells of Papaver somniferum increased with pHext in the presence of KNO2, which acidified the cytoplasm. However, they were unable to detect an effect of pHext, from 5 to 8, on ABA release in the absence of this salt. ABA is usually present in its dissociated form (ABA–) in the normally high pH of the cytoplasm. This form cannot cross the plasma membrane and it is therefore trapped in the symplast. ABAH formed upon cytoplasmic acidification is lipophilic, and can diffuse out of the symplast such that ABA accumulates in the apoplast if this is alkaline enough.
Thus, episodes of soil (or air) drying that are severe enough to induce reductions in shoot water potential may increase the penetration of ABA to the guard cells in one of three ways. Firstly, the up-regulation of ABA biosynthesis in leaves means that bulk leaf ABA concentrations increase, but it also means that ABA arriving at the leaf via the xylem is more easily able to penetrate to the epidermis as the symplastic reservoir will be full. Finally, leaf dehydration-induced cytoplasmic acidification induces more effective release of symplastic ABA to the apoplast, increasing its likelihood of penetrating to the guard cells in the epidermis. These leaf-sourced hydraulically induced chemical changes may occur either in addition to or instead of the root-sourced increases in xylem sap pH, which increase the penetration of ABA to the guard cells by reducing its symplastic uptake, even in leaves that are still fully turgid.
Evidence for chemical signals originating in leaves that may control ABA penetration to guard cells. Recently it has been found that the leaf may be able to affect the penetration of ABA to the guard cells in the absence of leaf water deficits and/or root-sourced changes in xylem pH. In other words, chemical changes originating in leaves that are not obviously hydraulically induced may influence the ABA signal. This was first indicated by the finding that changes in VPD (and mild soil drying episodes) reduced leaf growth and stomatal aperture in some species without affecting water potential, apparently by sensitising growth receptors and guard cells to xylem ABA (Tadieu & Davies 1992, 1993). This is an entirely different response to that described above, in which large soil water deficits (or increases in VPD) induce obvious hydraulic signals that reduce leaf water status prior to stomatal closure. The authors coined the term ‘isohydric’ to describe the resultant stability of leaf water status that they described in the field, in the face of fluctuating VPD and soil water content. They proposed that mild stresses induce small hydraulic changes, possibly undetectable in bulk shoot water measurements, that can directly sensitize guard cells to ABA so that stomata close and maintain leaf water status. We have proposed here and elsewhere that soil drying-induced pH changes can often be responsible for that heightened sensitivity of leaf physiological responses to xylem ABA sometimes observed, such as those described by Tardieu & Davies (1992, 1993). It is also possible that leaf-sourced changes in pH can occur in response to variable VPDs in the absence of VPD-induce leaf water deficits, possibly alongside the VPD-induced hydraulic influence proposed by Tardieu & Davies (1992, 1993). Some evidence that this does indeed occur is provided below; however, it is by no means a universal phenomenon. For example, in sunflower ABA and plant water status do not interact at all even when leaf water deficits are clearly detectable (Tardieu, Lafarge & Simonneau 1996) and plant water status is variable and unregulated (anisohydric).
Recent work at Lancaster University has shown that the aerial environment can change leaf sap pH in both Forsythia×intermedia cv Lynwood and Hydrangea macrophylla cv Bluewave. Increases in sap pH correlated with reductions in stomatal conductance in the absence of measurable changes in leaf relative water content in F.intermedia. Aerial conditions that increase transpiration (high VPD, PPFD and leaf surface temperature) increased the pH of sap expressed under pressure from the upper parts of F. intermedia and H. macrophylla shoots, and the stomata were more closed (see Fig. 2A and B for data from H. macrophylla, see Davies, Wilkinson & Loveys 2002 for data from F. intermedia). The same conditions that caused stomatal closure in the afternoon did not do so in the morning in F. intermedia (and vice versa in H. macrophylla), suggesting that the sensitivity of guard cells to pH (and ABA) change over the course of the day.
In F. intermedia, high VPD/PPFD, increased pH and reduced stomatal conductance correlated with increases in bulk leaf (but not xylem sap) ABA, and we suggest that ABA removal by the phloem may have been inhibited (see below). In H. macrophylla the same set of conditions had no effect on bulk leaf [ABA], and it can only be hypothesized that the increased pH increased the penetration of xylem-sourced ABA to the guard cells and/or induced its release from the symplast as a result of the types of processes described above. That the climatically induced changes in sap pH originated from changes within the leaf apoplast rather than from the incoming xylem sap itself were indicated by effects of soil drying on the response. Unusually (but see below), soil drying acidified the xylem sap of F. intermedia plants. In plants in drying soil the correlation between apoplastic sap pH and both the changes in aerial conditions and changes in stomatal conductance were poor (Davies et al. 2002). This was hypothesized to result from the combination of the sap from the two different sources (root and leaf, see also Mühling & Lauchli 2001), which would partially override the climate-induced pH change. There is plenty of evidence in the literature that such gradients in pH exist between the leaf apoplast and the xylem, with pH often being higher in the apoplast (Hoffmann & Kosegarten 1995; Mühling & Lauchli 2000). Regulation of gas exchange will be the result of a processing by the guard cells of environmental information received by both the roots and the shoot. Changes in sap pH may be a unifying influence on guard cells through which both classes of information may act.
It is as yet unknown which aerial factor caused the changes in pH and stomatal aperture described above. High VPD may have reduced leaf cell pm H+-ATPase activity by causing slight (localised) changes in water relations (see above, Hartnung & Radin 1989). Alternatively (or additionally) at high PPFD, increased removal of CO2 from the apoplast for photosynthesis may alkalize this compartment. Stahlberg, Van Volkenburgh & Cleland (2001) found that photosynthetic removal of CO2 from Coleus leaves induced an increase in apoplastic pH that could be propagated over a distance of 2 cm. A high PPFD- and/or temperature-induced increase in nitrate uptake by leaf cells could have the same effect (see below).
The multiple effects of pH (in the root, stem and leaf) on the ABA signalling process raise an interesting possibility that species that exhibit a very low stomatal sensitivity to ABA may do so because the xylem/apoplastic pH has not become more alkaline with soil drying/climatic changes. Schurr & Schulze (1995, 1996) have shown that there are marked differences between species in the effects of soil drying on hormone and nutrient fluxes to shoots. In our experience, this is also true for changes in pH (soil drying increases sap pH in tomato, C. communis and barley; decreases pH in F. intermedia, and has no effect on sap pH from H. macrophylla and Cotinus coggyria cv Royal Purple). These results suggest that the differences in signalling responses to drought are especially significant between woody and herbaceous species (see below).
Release of ABA from inactive transport forms within the leaf. Another way in which the leaf may modify the ABA signal has been described by Dietz et al. (2000). ABA conjugates, which are themselves unable to affect stomatal aperture or plant growth, have been detected in the xylem sap of stressed plants (Netting, Willows & Milborrow 1992). ABA-GE (ABA-glucose ester) seems to be the most common of these. Glycosylation of ABA is thought to be a mechanism to allow for the export of ABA from cells independently of the prevailing transmembrane proton gradient. In addition, its release from root stelar cells and liberation to the xylem seems to be greatest under conditions of inhibited ABA metabolism (Sauter & Hartung 2000). ABA-GE transport may therefore be important as a root-sourced stress signal during soil flooding when ABA synthesis becomes inhibited. ABA-GE in barley xylem sap has also been shown to increase under salinity, whereas its concentration in the apoplast of primary leaves was unchanged (Dietz et al. 2000). This is because the barley leaf apoplastic sap contained the enzyme β-glucosidase, which rapidly cleaved ABA from the inactive ABA-GE pool arriving at the leaf apoplast via the xylem. The data support the hypothesis that ABA-GE is a long-distance transport form of ABA. Anything that affects the activity of β-glucosidase in the leaf will therefore affect the [ABA] in the leaf apoplast, presumably thereby affecting stomatal behaviour. There is also some evidence that ABA can be released from its conjugated form in the root before transport in the xylem (Sauter & Hartung 2000).
Removal of ABA from leaves by the phloem. Finally, leaves may alter the ABA signal by influencing the rate at which ABA is removed from them via the phloem. If less ABA exits the leaf then presumably there is an increased likelihood that the remaining ABA will penetrate to the guard cells. Jia & Zhang (1997) detected a large reduction of ABA transport out of detached maize leaves in the phloem when buffers at a ‘droughted’ pH of 7·4 as opposed to a ‘well watered’ pH of 5·5 were injected into the xylem. Wilkinson & Davies (1997) found that detached C. communis leaves taking up 3 µm ABA from the external medium had a greater bulk leaf [ABA] 3 h later when this was fed at a ‘droughted pH’ of 7 as opposed to pH 6. These data suggest that an alkaline apoplastic pH prevents ABA movement out of the leaf via the phloem. This could either result from the decreased pH gradient between the apoplast and the phloem, trapping the ABA in the leaf apoplast (if apoplastic ABA transfer is important), or from a decrease in the amount of ABA present in the sieve element companion cell (if symplastic phloem transfer is predominant). Phloem sieve tubes are living cells with an internal pH of 7·5 (Baier & Hartung 1991), so that the phloem’s efficiency as an ABA reservoir (and export conduit) will depend on the same types of processes described above. Support for an effect of apoplastic pH on loading of ABA into the phloem comes from work by Peuke, Jeschke & Hartung (1994), who showed that ABA flux in the phloem from the leaves of ammonium-treated plants is extremely large. As will be explained below, ammonium nutrition reduces the apoplastic pH of leaves.
Other work has shown that the ABA concentration actually increases in the phloem in response to stress (Wolf, Jeschke & Hartung 1990). This presumably occurs only under more severe and/or prolonged stress in which [ABA] will be high in all tissues due to both root and shoot biosynthesis.
The sensitivity of guard cells to a perceived concentration of aba
As described above, plants control the amounts of ABA that reach the guard cells in the leaf epidermis via processes that originate in the root, the stem or the leaf. However, they also control the sensitivity with which guard cells respond to the ABA concentration that finally reaches them. Changes in guard cell sensitivity to ABA occur over the day/night cycles and as a plant matures, often in response to changes in electrical and osmotic membrane potentials generated by guard cells themselves (for reviews see Assmann & Shimazaki 1999; Blatt 2000).
Effects of apoplastic composition on guard cell sensitivity to ABA
Calcium ions are known to enter guard cells and promote stomatal closure and inhibit opening when applied to isolated epidermis, and apoplastic calcium sensitizes guard cells to ABA (De Silva, Hetherington & Mansfield 1985). This is because ABA-induced stomatal closure involves increases in cytosolic-free Ca 2+ in guard cells as part of the signal transduction chain leading to loss of turgor (McAinsh, Brownlee & Hetherington 1990). The addition of 8 mm calcium to the transpiration stream of detached leaves initiates rapid reductions in stomatal conductance (Ruiz, Atkinson & Mansfield 1993). Variations in calcium and ABA concentration in the xylem with soil drying (Schurr et al. 1992) may interact to limit stomatal opening. To add further complexity, guard cell sensitivity to external calcium can also change under some circumstances (see below).
Potassium is centrally involved in the up-regulation of turgor-driven stomatal opening, and the effectiveness of ABA at closing stomata may be reduced when its application to the medium bathing epidermal strips exceeds a certain concentration (Wilson, Ogunkanmi & Mansfield 1978; see below). The significant reductions in xylem sap [K+] observed at high bulk soil density and in dry soil (Mulholland et al. 1999; Roberts & Snowman 2000) may have maximized a potential reduction by ABA of stomatal conductance. It is also interesting to note that ABA in the root inhibits potassium loading into the xylem by xylem parenchyma cells (Roberts & Snowman 2000), indicating that ABA in the root could potentially sensitize stomata to ABA in the leaf by reducing K+ availability as a guard cell osmoticum.
Protons can also directly affect stomatal aperture and/or its sensitivity to ABA. There may be an optimum apoplastic pH for stomatal opening related to the activity of guard cell pm H+-ATPases and other pm ion channels that directly control turgor. Wilkinson & Davies (1997) and Wilkinson et al. (1998) have provided evidence that the direct effect of pH on stomatal aperture in isolated epidermis can be very different to that seen when buffers of differing pH are supplied to intact leaves via the xylem. Increasing the pH of the medium on which isolated epidermal strips of C. communis were floating actually opened stomata and reduced their sensitivity to ABA (see also Schwartz et al. 1994). Feeding artificial sap of pH 7 to intact detached leaves of the ABA-deficient tomato mutant flacca increased stomatal aperture and transpirational water loss compared to when the fed sap was buffered to pH 6, but the response to pH was reversed when ABA was supplied in the fed sap. Several other studies have shown that, whereas pH per se had no effect on stomata in isolated epidermis, reduced pH did sensitize stomata to ABA (e.g. Anderson et al. 1994), as described above for C. communis. These authors proposed that because guard cells take up ABA more efficiently at more acidic pH, its contact with internally located receptors would be increased, thereby inducing more sensitive stomatal closure. However, given the evidence that guard cell symplastic ABA is not always necessary for stomatal closure, it may instead be the case that an acidic external pH directly modifies the ionic balance and thereby the turgor of the guard cell, as suggested above, to predispose it toward stomatal closure. For example, there is already evidence that under some circumstances (but see below) reduced external pH can directly reduce K+ influx and increase K+ efflux at the guard cell plasma membrane (Roelfsema & Prins 1998). The reason for the fact that in some cases (Wilkinson & Davies 1997) but not others (Anderson et al. 1994) acidic pH closes stomata in isolated epidermis in the absence of ABA could be related to the ionic composition of the bathing medium used. Thompson et al. (1997) describe how in the presence of 75–100 mm KCl stomatal aperture no longer varied with pHext as it did in the presence of 50 mm KCl, whereas the sensitivity of aperture to ABA varied with pH as usual in both media. It must be supposed, however, that the more potent effect of pH in the intact plant is on the distribution of ABA between the compartments of the leaf. This must normally override the opposing direct effect of pH to modulate guard cell sensitivity to local ABA, otherwise we would find that high xylem sap pH opens stomata, and this is often clearly not the case (e.g. Patonnier, Peltier & Marigo 1999, Wilkinson 1999).
There is some evidence in the literature that apoplastic sugars could potentially increase guard cell sensitivity to ABA. As the water potential in soil around ash seedlings decreased, there was a significant increase in the concentrations of malate and mannitol in the xylem and a concomitant decrease in maximum stomatal conductance (Patonnier et al. 1999). A transpiration bioassay using detached leaves from control plants found that mannitol had no effect on stomata, whereas malate between 0·5 and 3 mm (the range found in xylem sap from the droughted plants) prevented stomatal opening. Other structural malate analogues (citrate, aspartate or shikimate) had the same effect, whereas neutral amino acids did not. The authors suggested that the effect of malate on stomata was to specifically increase the activity of the anion efflux channel at the pm of the guard cells (GCAC1), as described by Hedrich & Marten (1993) for V. faba. As ABA also induces anion loss and therefore reduces turgor in guard cells, it may be supposed that malate and ABA act synergistically at GCAC1 to close stomata. Alternatively, because of the non-specificity of the effect in ash, Patonnier et al. (1999) suggest that the effects of the inorganic acids on stomatal aperture were directly dependent on the negative charge of the molecules. However, given that under some circumstances increasing the local apoplastic pH at the guard cell induces stomatal opening and/or reduced guard cell sensitivity to ABA as described above, it is possible instead that the negative charge of the inorganic acids exerted an indirect effect on stomata. Namely, the increased pH that they induce (see below) may cause preferential accumulation of ABA in the apoplast, rather than directly affecting guard cell sensitivity to a given ABA concentration.
High CO2 in the substomatal cavity (see below) also reduces stomatal aperture and may enhance the response of guard cells to ABA. Hedrich & Marten (1993) have suggested that high CO2 induces photosynthetically derived malate to be released from mesophyll cells adjacent to guard cells, which, as described above may directly induce loss of guard cell turgor. Mesophyll-derived apoplastic sucrose is another CO2-sensing candidate which can reduce stomatal aperture, by virtue of its osmotic effect to draw out guard cell water, but this will presumably occur only in species which load sucrose into the phloem apoplastically (Lu et al. 1997). Apoplastic malate or sucrose may signal the plant that mesophyll-derived photosynthetic products are high and stomata can therefore close to limit water loss. However, Schmidt & Schroeder (1994) have suggested that the efflux of osmotically active malate from guard cells themselves (which also contain photosynthetic apparatus) via anion channels may provide a positive feedback mechanism for ‘plentiful photosynthate’-induced stomatal closure. As described above, it may be supposed that ABA and malate or sucrose therefore act synergistically to reduce guard cell turgor, and this may be the basis of a high CO2-induced sensitization of stomata to ABA. Alternatively, the apoplastic mode of sucrose loading into the phloem could remove H+ from the apoplast and increase its pH (see Tetlow & Farrar 1993), which in an intact leaf would cause ABA to compartmentalize in the apoplast. However, Zeiger & Zhu (1998) provide evidence that in the light high CO2 detection occurs at the blue light photoreceptor in the guard cell chloroplast. It is therefore possible that CO2 and ABA could interact directly at the guard cell pm H+-ATPase, as its down-regulation comprises part of the signal response chain of both ABA and CO2 perception in guard cells. Apoplastic sugars may only be involved in the stomatal response to CO2 when its concentration is high due to high light-saturated photosynthesis, whereas the direct effect of CO2 to close stomata may occur when its concentration is high because photosynthesis is limiting at low light intensities (which include the blue light wavelengths). However, because stresses such as drought increase the concentration of sugars in the xylem as described above, they may still exert an important influence on stomatal sensitivity to ABA.
Cytokinins in the xylem have been shown to decrease stomatal sensitivity to xylem ABA in cotton (Radin, Parker & Guinn 1982), and their presence at high concentrations in the transpiration stream can act directly on stomata to increase their aperture (Incoll & Jewer 1987a, b). Blackman & Davies (1985) showed that these effects became larger as leaves aged. Fusseder et al. (1992) report data from field studies of almond trees under desert conditions that suggest an ameliorating influence of increasing xylem cytokinin concentration on ABA’s effect on stomatal aperture. More recently, Stoll et al. (2000) showed that under partial root drying there was an increase in xylem sap and leaf ABA in grape vines which correlated with reduced stomatal conductance. Concomitant increases in xylem sap pH and reductions in cytokinins were observed, both of which could have enhanced stomatal sensitivity to ABA, the latter via direct sensitization.
Effects of the aerial microclimate around the leaf on guard cell sensitivity to ABA
We have shown recently that chilling temperatures applied to epidermal strips reduce the sensitivity of C. communis stomata to ABA in the bathing medium (Fig. 3). This is despite the fact that the same conditions induce stomatal closure in the intact leaves of this chill-tolerant species to counteract chill-induced reductions in water uptake at the root (Wilkinson et al. 2001). Given much evidence in the literature that chilling temperatures can eventually induce ABA biosynthesis, it seems counterintuitive that stomata become less responsive to this hormone. This finding may explain why some research has shown that low temperatures can actually open stomata in vivo in chill-sensitive species. Although we found that stomata remained open when C. communis epidermis was isolated from the rest of the leaf and floated on pre-chilled media, the closing response to low temperatures could be restored (Wilkinson et al. 2001) by supplying the epidermal strips with the calcium concentration estimated to exist in the leaf apoplast (0·01–0·05 mm, De Silva, Honor & Mansfield 1996). Stomata in strips floating on the same calcium concentration remained open at room temperature. Thus, chill-induced stomatal closure occurs via a sensitization of guard cells to calcium in the apoplast, even though stomatal sensitivity to ABA is reduced (Wilkinson et al. 2001). In the context of this review, it might be more pertinent to state that increasing temperature increases the sensitivity of guard cells to ABA (Fig. 3), presumably by increasing ABA binding to its receptors in the guard cell pm, and/or by increasing the activity of the ion transporters involved in the ABA signal response chain.
Although early responses to chilling temperatures directly reduce guard cell sensitivity to ABA, there is evidence in the literature that a previous period of chilling stress followed by recovery can sensitize guard cells to close in response to a second stress such as drought (Wilson 1976) –‘chill-hardening’. Allan et al. (1994) found that guard cells in epidermis isolated from cold-acclimated C. communis leaves did not require cytoplasmic increases in calcium for ABA to cause stomatal closure. Our work (Wilkinson et al. 2001) demonstrates that chilling temperatures may replenish internal stores of calcium in guard cells. In effect, cells are ‘primed’ by cold pre-treatment for closure in response to a subsequent exposure to ABA, as internal calcium is centrally involved in the ABA signal transduction chain.
There is much evidence in the literature that increasing VPD closes stomata and increases their sensitivity to ABA, although again the underlying cellular mechanisms for these responses remain under debate. We have proposed some explanations for this response in the sections above (see also Zhang & Outlaw 2001a), based on ABA redistribution. However, Assmann, Snyder & Lee (2000) found that increasing VPD closed stomata in both ABA-deficient and ABA-insensitive mutants of Arabidopsis thaliana, showing that, at least in some species, it is necessary to invoke the more traditional view that dry air has a more direct effect on guard cell and/or epidermal cell turgor. Like responses to soil drying, responses to VPD may differ between species, and could be hydraulic or chemical or a combination of both. Saliendra et al. (1995) found that in western water birch,restoration of shoot water potential using a root pressure vessel fully reversed stomatal closure in response to both a soil water deficit and increased VPD. By contrast, Fuchs & Livingstone (1996) found that stomatal responses to VPD in Douglas fir and alder seedlings were only partially reversed by restoration of shoot water relations.
It is well established that indirect responses of stomata to light involve photosynthetic effects on intercellular CO2 concentrations (see above). Stomata are also directly responsive to light (for reviews see Assmann & Shimazaki 1999 and Zeiger & Zhu 1998): photosynthetically active radiation (400–700 nm) induces guard cell photosynthesis and opens stomata via the production of osmotically active solutes such as malate. In addition, blue light (425–475 nm) induces stomatal opening by transposing a signal cascade in guard cells that includes a stimulation of H+ efflux to the apoplast via pm H+-ATPases. This provides the driving force for uptake of K+ and Cl– and therefore increases guard cell turgor (depending on the prevailing membrane potential, see Roelfsema & Prins 1998). Stomata may be more sensitive to ABA under conditions of low light because: (a) the intercellular [CO2] will be higher (less photosynthetic removal); (b) reduced guard cell photosynthesis means lower guard cell [malate]; and/or (c) the blue-light-dependent H+-ATPase in the guard cell pm, which is directly down-regulated as part of the ABA signal response chain, will already be less active when light is limiting. However, work by Clephan in our laboratory could only detect evidence for a sensitization of C. communis stomata to ABA at low light intensities when leaves were intact, and not in isolated epidermis (personal communication). This suggests that only indirect effects of light (e.g. on the mesophyll-determined [CO2]) affect the response of stomata to ABA. The same series of experiments revealed that more ABA reached the epidermis of intact leaves at lower light intensities (despite decreased ABA fluxes into them). It was suggested that light-induced alkalinization of the mesophyll chloroplast stroma and therefore ABA retention in chloroplasts (Kaiser & Hartung 1981) may have inhibited ABA penetration to the epidermis in saturating light, such that stomata were less sensitive to the [ABA] supplied.
These responses occur over very low light intensities, and will be very different to those suggested to result from stressfully high PPFD and/or VPD described earlier, which correlate with apoplastic alkalinization and stomatal closure.
Nitrate stress and aba signals may interact as soil dries
Plants growing in dry soil frequently show a sensitive reduction in the delivery of nitrate to shoots and this signal can regulate leaf processes independently from a variation in plant water relations (Shaner & Boyer 1976). When nitrate supply is limiting (regardless of soil water availability), stomata close, leaves grow more slowly, root growth is maintained, and is often characterized by greater lateral root proliferation. All of these are also symptoms of water stress. It may be that the soil drying response that is often characterized is at least in part a response to a limitation in N supply.
One way in which limiting nitrate could reduce leaf growth would simply be the result of a decrease in the supply of N-derived amino acids and structural proteins. However, there is much evidence that responses to N-deprivation are governed instead by fast root–shoot (or shoot–root) chemical signals. Krauss (1978) found marked increases in ABA in the xylem sap of rooted sprouting potato tubers deprived of N, and increases in shoot and leaf ABA in N-deprived plants have since been observed in several species (see Radin et al. 1982; Clarkson & Touraine 1994). However Peuke et al. (1994) found that ABA in low nitrate-grown Ricinus communis was only increased in the xylem and not in the leaves. Similar data were reported by Brewitz, Larsson & Larsson (1995) for barley. Some studies detected no long-term effect of reduced nitrate supply on tissue ABA concentrations at all (Chapin et al. 1988, barley and tomato). When Peuke et al. (1994) grew R. communis on ammonium as a substitute N source, rather than on low N, they were able to demonstrate large increases in leaf and xylem ABA, and also large increases in phloem ABA transport, but the stomata in the leaves did not close in response to this increase in ABA flux around the plant.
To explain these rather variable responses, it has been suggested that rather than inducing increases in [ABA] per se, low nitrate availability may modulate responses in the shoot via a sensitization of stomata/leaf growth receptors to ABA. Radin et al. (1982) have shown that both N and P deficiency can enhance stomatal sensitivity to an ABA signal. Schurr et al. (1992) found that the concentration of nitrate in the xylem sap of droughted sunflower plants regulated stomatal sensitivity to ABA. It is possible that ABA and nitrate can interact to influence stomata via nitrate-induced changes in the pH relations of the leaf (Raven & Smith 1976).
There is now much evidence available for effects of N-containing compounds on plant pH balance. Mengel, Planker & Hoffmann (1994) found that young sunflower plants growing on a sole NO3– (or bicarbonate) medium had chlorotic pale leaves, but those on ammonium nitrate did not. On NO3– (or bicarbonate), the pH of the leaf apoplast was increased in comparison to that of plants on ammonium, and this reduced leaf chlorophyll content by inhibiting iron reduction in the apoplast (this process has an acidic pH optimum). Tagliavini et al. (1994) reported similar findings in field-grown Kiwi fruit: many of the plants displayed chlorotic pale leaves, and these were found to contain xylem sap with a higher pH than those with healthy green leaves. The plants were growing on calcareous soils, which contain high levels of HCO3– and NO3–. When the leaves were sprayed with citric acid or H2SO4, to counteract the (presumably) nitrate-derived alkalinization, re-greening occurred. More recently, Kosegarten, Hoffmann & Mengel (1999) found that nitrate nutrition (and nitrate plus HCO3–) gave rise to a high apoplastic pH in immature but not mature sunflower leaves, especially at distinct interveinal sites. This did not occur in sole NH4+ or NH4/NO3 nutrition. Mühling & Lauchli (2001) also found that nitrate nutrition gave rise to more alkaline leaf apoplastic sap than ammonium nutrition, in both Phaseolus vulgaris and sunflower, but not in Vicia faba or Zea mays. Mengel et al. (1994) hypothesize that NO3– may alkalize apoplastic sap because it is co-transported into the leaf cytoplasm with a proton (Ullrich 1992), and this presumably occurs to a greater extent in young leaves because these are the site of greatest N reduction. HCO3– may simply neutralize H+ pumped into the xylem sap by ATPases. Soil-sourced NH4+ is reduced in roots, and the neutral amino acids which are produced are transported in the xylem to the leaf instead of NO3– (see below). Amino acids are already protonated at apoplastic pH values, so uptake into the mesophyll will remove fewer protons from the apoplast than NO3–/H+ co-transport, presumably giving rise to a more acidic apoplast. These results could help to explain the findings of Peuke et al. (1994) described above, that even though ammonium-grown R. communis plants contain more root-sourced ABA than nitrate-grown plants, they have open stomata and poor water use efficiencies. Ammonium nutrition may have given rise to a more acidic leaf apoplast than nitrate nutrition, as it did in sunflower, which as described above can reduce the accessibility of ABA to guard cells by increasing its symplastic sequestration and increasing the efficiency of its transfer into the phloem and out of the leaf. Similarly, Allen & Raven (1984) found that R. communis with NH4+ as the N source transpired more water per g dry matter synthesized than when NO3– was the N source. Raven, Griffiths & Allen (1984) went a step further and showed that ammonium-grown plants actually had increased stomatal apertures.
So far, the data described above would tend to suggest that a reduction in nitrate availability, by whatever means (soil nutrient deficiency, soil moisture deficiency, soil compaction or soil flooding), would tend to decrease the leaf apoplastic pH, as nitrate nutrition alkalizes this compartment. This is counterintuitive because, as discussed above, acidic apoplastic sap pH reduces the accessibility of ABA to stomata, and would tend to oppose stomatal closure (as in ammonium nutrition). However, all of these stresses are characterized by stomatal closure and increased sensitivity of stomata to ABA. Some of them are characterized by increases in xylem sap pH. How can we explain these discrepancies?
The answer lies in the fact that N availability influences the species of N-containing molecule that is carried in the xylem. Different N-containing molecules have different abilities to influence the xylem and/or apoplastic pH. When nitrate is plentiful, nitrate is transported in the xylem to the leaf, where it is reduced in the mesophyll. When nitrate is limiting, either due to low soil availability or to soil drought/flooding/compaction, then nitrate reduction is switched from the shoot to the root (see Lips 1997). Root nitrate reductase activity produces hydroxyl ions, which are then converted to COO– (often as malate). This is carried in the xylem to the leaf. Malate and related compounds alkalize apoplastic sap to an even greater extent than nitrate can. It is well known that organic anion salts such as malate have higher pH values than mineral anion salts such as nitrates (Wallace, Wood & Soufi 1976). Kirkby & Armstrong (1980) showed that under nitrate nutrition, organic acids were present in only trace amounts in the xylem sap of castor oil plants, which had a pH of 5·6. Under NO3– stress the pH of the sap rose to 7·3, and considerable amounts of organic acid anions were present. Patonnier et al. (1999) describe drought-induced increases in xylem sap pH and in the organic acid fraction of the xylem (that presumably result from a drought-induced switch of nitrate reduction to the root) that could reduce stomatal aperture in ash. Neutral amino acids (the product of ammonium nutrition that is transported in the xylem) fed into the xylem of detached leaves had no affect on stomatal aperture. These authors suggested that the organic acids had a direct effect on guard cells (see above), but it is just as likely that their effect to increase xylem sap pH allowed ABA more access to the guard cells by reducing symplastic sequestration.
Rather than having a direct affect on xylem and therefore leaf apoplastic pH, organic acids could increase apoplastic sap pH by virtue of their di-anionic form: upon transfer of these to the symplast, the uptake of two protons and therefore their removal from the apoplast will be required. Whether they directly increase xylem sap pH, or increase apoplastic sap pH by removing protons during symplastic sequestration, the result will be the same: greater penetration of ABA to its site of action at the guard cells. This could explain why much research shows increased stomatal sensitivity to ABA in response to N-deficiency, whereas N-deficiency does not always lead to increases in leaf [ABA]. It could also be one explanation for the effect of soil drying to increase xylem sap pH.
To summarize, the literature shows that xylem/apoplastic sap seems to be most alkaline under N deficiency when nitrate reduction in the root produces inorganic anions and loads them to the xylem. It is of intermediate alkalinity when nitrate is carried to leaves in the xylem stream, and most acidic when ammonium is reduced in roots and the neutral amino acids produced are loaded to the xylem stream.
An explanation for differences between species in the effects of rhizospheric stresses on xylem sap pH could lie in the inherent species response to low nitrate availability. Most crop species exhibit very active nitrate uptake and xylem loading, carry out nitrate reduction in the leaf and have fast growth rates. However, some slow-growing trees carry out N assimilation in their roots (see Lips 1997). Perhaps fast-growing crop species will exhibit xylem sap pH changes as N deprivation switches nitrate reductase activity to the root where, e.g. malate will be transferred to the xylem. However, in species that reduce N in the root as a matter of course, N deficiency or soil drying will not change the position of N reduction, such that xylem sap concentrations of organic acids show little change as stress develops, and xylem pH remains stable. Our finding that xylem sap pH in woody species is unresponsive to soil drying supports this hypothesis.
Early work on the chemical control of plant growth and functioning under stress led to the seductively simple hypothesis that ABA produced in the roots of plants undergoing rhizospheric stresses moved to the shoot to provide a measure of resource availability. Shoot responses could then be modified as a function of the intensity of this signal. We show here how a range of influences interact to modify the ‘root signal’ and much of this modification occurs via a modulation by the environment of the pH relations of the different compartments of the root and the leaf. This provides some understanding of how rhizospheric and aerial influences may interact to influence leaf growth and functioning. We are now moving towards a more complete understanding of the whole-plant signalling process and an understanding of why the intensity of chemical signalling and the response to these signals may vary between environments and plant species. To aid in our thinking we now have available a range of new technology to assess pH and local accumulation of ABA in the apoplast adjacent to the guard cells. This technology will allow us to test some of the hypotheses proposed in this review.
The authors would like to thank the members of MAFF Horticulture Link project HL0132LHN/HNS 97 for financial support towards the work with ornamental species. We are grateful to the American Society of Plant Biologists for permission to reprint Fig. 1 (represented as Fig. 2 at source) from: Wilkinson, Clephan & Davies (2001). This work is copyrighted by the American Society of Plant Biologists.
Received 12 June 2001; received in revised form 7 September 2001; accepted for publication 10 September 2001