Shoot and root growth are differentially sensitive to water stress. Interest in the involvement of hormones in regulating these responses has focused on abscisic acid (ABA) because it accumulates in shoot and root tissues under water-limited conditions, and because it usually inhibits growth when applied to well-watered plants. However, the effects of ABA can differ in stressed and non-stressed plants, and it is therefore advantageous to manipulate endogenous ABA levels under water-stressed conditions. Studies utilizing ABA-deficient mutants and inhibitors of ABA synthesis to decrease endogenous ABA levels, and experimental strategies to circumvent variation in plant water status with ABA deficiency, are changing the view of the role of ABA from the traditional idea that the hormone is generally involved in growth inhibition. In particular, studies of several species indicate that an important role of endogenous ABA is to limit ethylene production, and that as a result of this interaction ABA may often function to maintain rather than inhibit shoot and root growth. Despite early speculation that interaction between these hormones may influence many of the effects of water deficit, this topic has received little attention until recently.
Shoot growth is very sensitive to water-limited conditions. Much evidence over the past 20 years indicates that the inhibition of growth is metabolically regulated, rather than being a direct consequence of altered plant water status arising from soil drying. For example, studies have shown that stem and leaf growth can be severely inhibited at low water potentials despite complete maintenance of turgor in the growing regions as a result of osmotic adjustment (e.g. Michelena & Boyer 1982). Furthermore, shoot growth can be so sensitive to soil drying that substantial inhibition can occur before the development of decreased water potentials in the aerial plant parts (Saab & Sharp 1989; Gowing, Davies & Jones 1990), and such observations have led to much interest in the involvement of non-hydraulic regulatory signals from the roots (Davies & Zhang 1991; Davies et al. 2000).
Root growth is usually less inhibited than shoot growth, or even promoted, in plants growing in drying soil, which is of obvious benefit to maintain an adequate plant water supply (Sharp & Davies 1989). An important feature of the root system response is the ability of some roots to continue elongation at water potentials that are low enough to completely inhibit shoot growth. For example, this occurs in nodal (adventitious) roots of maize, which must penetrate through the often dry surface soil (Westgate & Boyer 1985), and in primary roots of a range of species, which helps seedling establishment under dry conditions by ensuring a supply of water before shoot emergence (Sharp, Silk & Hsiao 1988; Spollen et al. 1993; van der Weele et al. 2000). Figure 1 shows for several important agronomic species that the primary root maintains substantial elongation rates at water potentials as low as −1·6 MPa, whereas shoot growth is inhibited completely at about −0·8 MPa.
The mechanisms that determine the different sensitivities of root and shoot growth to water stress are not well understood. Although hormones are likely to play important regulatory roles, and despite considerable early attention to this topic (e.g. Itai & Ben-Zioni 1976; Bradford & Hsiao 1982), the involvement of these compounds has not been elucidated. Most interest in this question has concerned the role of abscisic acid (ABA). This review evaluates the evidence for the involvement of ABA in root and shoot growth responses to water stress, with a focus on advances in understanding gained from studying effects of endogenous ABA deficiency. In particular, recent studies have revealed that an important role of endogenous ABA is to limit ethylene production, and that this interaction is involved in the effects of ABA status on root and shoot growth. Despite early speculation that interaction between these hormones may influence many of the effects of water deficit (Wright 1980; Bradford & Hsiao 1982), this topic has received little attention until recently.
Approaches to study the role of aba in growth responses to water stress
Interest in the involvement of hormones in regulating growth responses to water stress has focused on ABA because (a) it accumulates to high concentrations in shoot and root tissues under water-limited conditions, often correlating with growth inhibition, and (b) it usually inhibits growth when applied to well-watered plants. Based on these findings, a commonly proposed function of increased ABA concentrations in water-stressed plants is growth inhibition (reviewed in Trewavas & Jones 1991; Munns & Sharp 1993; Munns & Cramer 1996). Two of the most compelling examples of this type of study are those by Creelman et al. (1990) in soybean and by Zhang & Davies (1990) in maize. In the former study, relationships between root and shoot elongation and the ABA contents of the respective growth zones were compared in seedlings whose growth was inhibited by transfer to vermiculite of low water potential (−0·3 MPa) or to vermiculite saturated with various concentrations of ABA. Similar to the responses for soybean shown in Fig. 1, shoot growth was more inhibited than root growth in the water-stressed seedlings. In the well-watered seedlings treated with ABA, growth was also more inhibited in the shoot than in the root at all ABA concentrations tested. The relationships of growth inhibition to ABA content suggested that variation in endogenous ABA accumulation, together with differing sensitivity to ABA, could largely explain the differential inhibition of shoot and root growth at low water potential. In the study of maize by Zhang & Davies (1990), the relationship between inhibition of leaf elongation rate and xylem sap ABA concentration was compared in plants grown in drying soil and in well-watered plants to which a range of ABA concentrations was supplied to part of the root system. The similarity between the two data sets suggested that the limitation of leaf growth resulting from soil drying could be explained entirely by the increase in endogenous ABA in the xylem sap.
Similarly to the results of these studies, primary root growth of well-watered maize seedlings is inhibited progressively when increasing concentrations of ABA are applied, and comparison of the relationship of growth inhibition with growth zone ABA content to the values obtained in seedlings grown at low water potential would suggest that endogenous ABA accumulation could explain the inhibition of root growth during water stress (Fig. 2).
However, interpretation of these and other similar studies relies on the assumption that when ABA is applied to well-watered plants and increases to internal levels similar to those of water-stressed plants, the effects on growth are the same as those resulting from endogenous ABA accumulation during water stress. Increasing evidence indicates that this is not necessarily the case. Decreases in plant water status have been shown to change the compartmentation (Hartung, Radin & Hendrix 1988; Bacon, Wilkinson & Davies 1998), apparent sensitivity (Tardieu & Davies 1992; Dodd & Davies 1996) and response (Sharp et al. 1994) to ABA in various organs.
To avoid these concerns, it is of obvious advantage to manipulate endogenous ABA levels under water-stressed conditions. However, despite the availability of ABA-deficient mutants (and recently, transgenics) in several species, and of several inhibitors of ABA biosynthesis, very few studies have used this approach to study the role of ABA in growth responses to water stress. A particular difficulty for the study of ABA in growth regulation is its effect on stomatal behaviour and, thereby, on plant water balance. Hence, ABA-deficient mutants are typically wilty due to high stomatal conductance even under well-watered conditions (Quarrie 1987), and this problem is exacerbated under soil drying conditions. The decrease in water status of ABA-deficient relative to control plants could inhibit growth independently of direct growth-regulatory properties of ABA. Jones, Sharp & Higgs (1987) subjected several ABA-deficient tomato mutants to soil drying, and reported greater decreases in leaf and root biomass than in the wild type. However, leaf water potentials also decreased more in the mutants, so interpretation of the role of ABA deficiency in the greater growth inhibition of the mutants was not possible. Reduced root biomass compared to the wild type of ABA-deficient and ABA-insensitive mutants of Arabidopsis in drying soil has also been reported (Vartanian, Marcotte & Giraudat 1994). However, plant water relations were not measured nor were comparative data for well-watered controls included, so the cause of the smaller root system development in the mutants could not be assessed.
Accordingly, studies of the effects on growth of endogenous ABA deficiency require experimental strategies that avoid variation in water status between plants with different levels of ABA. A few such studies have now been published, and the results are changing the view of the role of ABA from the traditional idea that the hormone is generally involved in growth inhibition.
Aba accumulation maintains maize primary root growth at low water potentials
When maize seedlings are grown in vermiculite at a water potential of −1·6 MPa, at which primary root but not shoot elongation continues (Fig. 1), the ABA content of the root growth zone increases about five-fold (Fig. 2). Three approaches have been used to study the effect on root growth of reducing the accumulation of ABA: (i) the inhibitor fluridone, which blocks carotenoid synthesis (at the conversion of phytoene to phytofluene) and, thereby, inhibits ABA synthesis although at an early step of the pathway; (ii) the vp5 mutant, which has a defect at the same step as that blocked by fluridone; (iii) the vp14 mutant, which has a defect in the synthesis of xanthoxin (Tan et al. 1997). Xanthoxin is converted to ABA via ABA-aldehyde, and its synthesis is considered a key regulatory step in water stress-induced ABA production (Qin & Zeevaart 1999). Initial studies used the fluridone and vp5 mutant approaches (Saab et al. 1990; Saab, Sharp & Pritchard 1992; Sharp et al. 1994). Recently, similar studies of vp14 were undertaken to strengthen the conclusion that the results obtained with fluridone and vp5 were due to ABA deficiency and not to other effects of their action (I.-J. Cho, B.-C. Tan, D.R. McCarty & R.E. Sharp, unpublished results). To circumvent variation in plant water status with ABA deficiency, seedlings were grown under conditions of near-zero transpiration (minimal shoot development, darkness and near-saturation humidity).
The results obtained with the three approaches were very similar. At high water potential, root elongation rates (and ABA contents) were minimally affected. At low water potential, in contrast, reduced ABA accumulation was associated with more severe inhibition of root elongation than in wild-type or untreated seedlings (Fig. 2). The three methods produced varying degrees of root tip ABA deficiency, but yielded a common relationship of inhibition of root elongation rate to root tip ABA content. In all cases, root elongation rate fully recovered when the ABA content of the elongation zone was restored to normal levels with exogenous ABA. Similar results were obtained using fluridone-treated and vp5 mutant seedlings grown at a water potential of −0·3 MPa (Sharp et al. 1994). These studies demonstrated that accumulation of ABA is required for the maintenance of maize primary root elongation at low water potentials.
Additional experiments at a water potential of −0·3 MPa showed that when the root tip ABA content of fluridone-treated seedlings was increased to twice (or more) the level in untreated seedlings, recovery of root elongation was decreased. Also, promotion of root elongation was not observed with any applied ABA concentration in untreated seedlings; again, slight inhibition of elongation resulted when the root tip ABA content was increased to twice the naturally occurring level (Sharp et al. 1994). Thus, the normal increase in ABA content in roots at low water potential is optimal for growth maintenance (also see Fig. 7).
It is important to note that the conclusion that the accumulation of ABA in water-stressed roots helps to maintain growth cannot be inferred by applying ABA to well-watered seedlings to simulate the increase in content under water stress (Sharp et al. 1994). Figure 2 shows that in well-watered seedlings treated with ABA, root growth was substantially inhibited at the root tip ABA content that occurs in water-stressed roots. These results illustrate that the maintenance of root elongation at low water potential by ABA is not solely a function of the increase in ABA content, but also requires the change in environmental conditions that modifies the growth response to ABA.
ABA maintains root growth at low water potentials by restricting ethylene production
Extending the study of ABA-deficient maize seedlings, recent work has shown that an important function of endogenous ABA in the maintenance of primary root growth at low water potential is to prevent excess ethylene production. First, it was observed that the roots of ABA-deficient seedlings at low water potential not only are shorter but also exhibit pronounced radial swelling primarily beyond the apical 2 mm (Fig. 3). Exogenous ethylene inhibits elongation and causes a similar pattern of swelling in maize primary roots at high water potential (Moss, Hall & Jackson 1988; Whalen & Feldman 1988). Second, it was shown that under water stress, ethylene production of ABA-deficient seedlings was substantially greater than in control plants (Figs 4 & 5), and this effect was completely prevented when the root ABA content was restored with exogenous ABA, together with restoration of root elongation (Spollen et al. 2000). Among the three methods used to reduce ABA accumulation (fluridone, vp5, vp14), the magnitude of the increase in ethylene evolution correlated both with the degree of ABA deficiency and the inhibition of root elongation rate (Fig. 5). It is important to note that in all cases the root tip ABA content of the ABA-deficient seedlings at low water potential remained higher than in well-watered seedlings (Fig. 2), indicating that an increased concentration of ABA is necessary to prevent excess ethylene production under water stress. Third, it was shown that root elongation could be largely restored by each of three inhibitors of ethylene synthesis or action (Spollen et al. 2000). Significantly, when ethylene evolution was restored to normal levels [using the inhibitor of ethylene synthesis aminooxyacetic acid (AOA)], root growth was completely restored (Fig. 4), suggesting that ABA deficiency did not also increase the sensitivity of growth to ethylene.
Importantly, since none of the inhibitors of ethylene synthesis or action substantially increased root elongation when ABA-deficiency was not imposed (results with AOA are shown in Fig. 4), ethylene does not appear to be an important cause of the inhibition of elongation in water-stressed roots that accumulate normal levels of ABA (Figs 1 & 2). In other words, the accumulation of ABA in water-stressed roots is sufficient to prevent excess ethylene production, in agreement with the above-mentioned finding that roots at low water potential exhibit an optimum ABA content for growth maintenance. The possible involvement of ethylene in the inhibition of growth during water stress is a long-standing question (El-Beltagy & Hall 1974), but there is no previous information in relation to root growth. This question is considered further below in relation to the inhibition of shoot growth in water-stressed plants.
The discovery that increased levels of endogenous ABA in roots at low water potential are required to limit ethylene production confirms ideas first suggested by Wright (1980) and developed further by Bradford & Hsiao (1982) but which had not been tested. These suggestions were based on the finding that pre-treatment with exogenous ABA prevented the increase in ethylene production caused by wilting of excised wheat leaves. It should be noted that although several other studies reported that ABA treatments of well-watered plants inhibited ethylene production (e.g. Gertman & Fuchs 1972; Yoshi & Imaseki 1981; Tan & Thimann 1989), there are also many reports of ABA-stimulated ethylene production (Riov et al. 1990 and references therein). Interpretation of these results is complicated by uncertainty that effects of applied ABA in well-watered plants are predictive of the role of endogenous ABA accumulation in water-stressed plants, as discussed above. The approach of using chemical and genetic means to manipulate endogenous ABA levels in water-stressed plants avoided these concerns.
Aba accumulation can inhibit and promote maize seedling shoot growth at low water potentials
As discussed above, ABA is generally regarded as an inhibitor of shoot growth (Trewavas & Jones 1991; Davies 1995; Munns & Cramer 1996). Initial studies of the effect of decreasing endogenous ABA levels in maize seedlings grown at low water potential (under conditions of near-zero transpiration) were consistent with this expectation (Saab et al. 1990, 1992). Seedlings were grown in vermiculite at a water potential of −0·3 MPa, at which slow rates of shoot growth continue (Fig. 1), and fluridone and the vp5 mutant were again used to decrease endogenous ABA accumulation. Studies were confined to the first 50 h after transplanting to low water potential, and showed that ABA-deficiency resulted in substantial promotion of shoot elongation. Similar results of experiments with fluridone are shown in Fig. 6a and Table 1 (first 40 h after transplanting). These experiments indicated an involvement of ABA in the inhibition of shoot growth of water-stressed maize seedlings, in contrast to its role in root growth maintenance.
Table 1. Effects of treatments with ABA, fluridone (FLU) and silver thiosulphate (STS) on shoot elongation rate of maize seedlings during early and later periods after transplanting to vermiculite at a water potential of −0·3 MPa. Elongation rates were calculated from the shoot length increases shown in Fig. 6. (Modified from Feng 1996)
Shoot elongation rate (mm h−1)
However, experiments of a longer duration revealed a greater complexity in the relationship of shoot growth to ABA status at low water potential (Feng 1996). After 50 h from transplanting to a water potential of −0·3 MPa, the effect of ABA-deficiency reversed so that shoot growth became inhibited relative to the control (Fig. 6a). Table 1 shows that shoot elongation rate between 72 and 135 h was inhibited by over 70% in fluridone-treated compared to control seedlings. Both the fluridone-induced promotion of shoot growth early after transplanting and the inhibition during the later period could be prevented with exogenous ABA. Moreover, when the same ABA treatment was applied without treatment with fluridone, shoot growth was further inhibited in the early phase and further promoted during the later phase (Fig. 6a), such that the shoot elongation rate was more than two-fold greater than in the control seedlings after 70 h from transplanting (Table 1). These results indicate that the normal accumulation of ABA in water-stressed maize seedlings functions to inhibit shoot growth only at an early stage of development, and subsequently helps to maintain growth, as in roots. In contrast to the roots, however, in which ABA levels were optimal for growth maintenance, endogenous ABA accumulation was apparently insufficient for maximal shoot growth during the second phase of the experiment.
In the experiment shown in Fig. 6a, supplemental ABA was supplied both during germination and after transplanting, and the resulting growth promotion after 70 h was largely attributable to growth of the mesocotyl. Since it is also the mesocotyl (and coleoptile) that is promoted by ABA deficiency during the first 40 h (Saab et al. 1990, 1992), the change in response of shoot growth to ABA from inhibition to promotion was not due to differential sensitivity among organs but occurred within the same organ. The shift to growth promotion was not restricted to the mesocotyl, however. In other experiments in which ABA was supplied only via the low-water-potential vermiculite into which the seedlings were transplanted, overall shoot elongation was similarly promoted after 50 h but the response was specific to leaf growth (Fig. 7). This difference probably reflects the different developmental stage of the shoots relative to the timing of ABA uptake. Importantly, there was no effect of the supplemental ABA treatment on root growth (Fig. 7), demonstrating that the promotion of shoot growth was not attributable indirectly to root growth inhibition (see shoot and root dry weight data in Fig. 7, legend), and in agreement with the above-mentioned conclusion that the endogenous ABA content of water-stressed roots is optimal for growth maintenance.
As was the case for root growth (Fig. 2), shoot growth of seedlings grown at high water potential was inhibited progressively with increasing concentrations of applied ABA (Feng 1996). Thus, the promotive effect of supplemental ABA on shoot growth at low water potential appeared to be specific to the water-stressed condition.
Effects of ABA on shoot growth are mimicked by inhibiting ethylene action
The change with time after transplanting to low water potential in the response of shoot growth to supplemental ABA, from inhibition to promotion, was closely mimicked by treatment with silver thiosulphate (STS) to inhibit ethylene action (Fig. 6b, Table 1). Thus, treatment with STS inhibited shoot growth early after transplanting and then markedly promoted shoot growth relative to the control. Similar results were obtained using AOA to inhibit ethylene synthesis (Feng 1996). Consistently, preliminary experiments showed that shoot growth could be increased during the early phase, and inhibited during the later phase, by applying either ethylene or the ethylene precursor 1-aminocyclopropane -1-carboxylate (ACC), simulating the effects of ABA deficiency.
Taken together, the results in Fig. 6 and Table 1 suggest that, as in roots, an important role of ABA in shoot growth of water-stressed maize seedlings is to restrict ethylene production and/or sensitivity. ABA accumulation thereby played either a growth-inhibitory (early phase) or growth-maintaining (later phase) role in the response of shoot growth to water stress, due to a change in the effect of ethylene from promotive to inhibitory.
The substantial promotion of shoot growth at later stages of development by treatment with STS indicates that the normal accumulation of ABA in the shoot was suboptimal for growth because it was insufficient to fully prevent ethylene-induced growth inhibition. This contrasts with the apparent sufficiency of ABA accumulation in the roots, and this difference accounted at least partly for the differential ability of the roots and shoots to grow at low water potential.
It should be noted that the extent to which ethylene caused the inhibition of shoot growth in water-stressed seedlings at later stages of development, and therefore the extent to which growth could be promoted by supplemental ABA, was influenced by nutritional status. It was subsequently discovered that the seedlings used for the experiments illustrated in Figs 6 and 7 and Table 1 were grown with suboptimal Ca2+ availability. When seedlings were grown with a different culture protocol involving supplemental Ca2+ (because the properties of the vermiculite had changed; Spollen et al. 2000), responses to STS and ABA after 50 h from transplanting were less pronounced (M.A. Else & R.E. Sharp, unpublished results). Such an interaction with nutrient status is not unexpected, because nutrient availability can markedly alter plant ethylene relations (e.g. He, Morgan & Drew 1992).
Generality of the promotive effect of aba on shoot growth
As discussed above, most of the research underlying the view that ABA is an inhibitor of shoot growth during water stress has involved applications of ABA to non-stressed plants or correlations of growth inhibition to increased endogenous ABA levels during stress. Indeed, the early phase of the maize seedling studies described above (Fig. 6; Saab et al. 1990, 1992) remains the only demonstration of enhanced shoot growth in response to endogenous ABA deficiency at low water potentials. The apparent change in response to ethylene during shoot development in those experiments, from promotive to inhibitory, is consistent with reports that ethylene stimulates mesocotyl growth in some species (Suge 1971; Cornforth & Stevens 1973), whereas it is usually inhibitory to shoot growth of terrestrial plants at later stages of development (Abeles, Morgan & Saltveit 1992; Lee & Reid 1997; Hussain et al. 1999). The following sections summarize recent studies which suggest that restriction of ethylene production may be a common function of ABA and therefore that endogenous ABA may often act to maintain rather than inhibit shoot growth.
Reassessment of the role of ABA in shoot growth of tomato and Arabidopsis under well-watered conditions
Paradoxically to the long-standing view that ABA is generally inhibitory to shoot growth, it has been observed for over 30 years that ABA-deficient mutants are often shorter and have smaller leaves than the corresponding wild types, and that leaf and stem growth can be substantially restored by applying ABA (Imber & Tal 1970; Bradford 1983; Quarrie 1987). As already mentioned, in addition to reduced growth, ABA-deficient mutants are typically wilty even when the soil is well supplied with water. This results from high stomatal conductance and has also been shown to involve decreased root hydraulic conductance. These effects can also be prevented by applying ABA (Imber & Tal 1970; Tal & Nevo 1973; Koornneef et al. 1982; Bradford 1983). Accordingly, the inhibited shoot growth of ABA-deficient mutants of tomato and Arabidopsis has been attributed to shoot water deficits (Bradford 1983; Neill, McGaw & Horgan 1986; Nagel, Konings & Lambers 1994; Léon-Kloosterziel et al. 1996). The growth-promotive effect of applied ABA in these cases has therefore been assumed to result from improvement of plant water balance. Consequently, these findings have generally not been regarded as evidence against a direct inhibitory role of ABA in leaf and stem growth, although the observations have led some authors to question this view (Jones et al. 1987; Quarrie 1987; Taylor 1987; Zeevaart & Creelman 1988).
Alternatively, endogenous ABA may be required to maintain shoot growth of well-watered plants independently of its effect on plant water balance. Consistent with the finding that ABA restricts ethylene production in water-stressed maize seedlings, it was reported that under well-watered conditions, ethylene production was greater in shoots of the flacca (flc) mutant of tomato (Tal et al. 1979) and in whole plants of the aba1 mutant of Arabidopsis (Rakitina et al. 1994) than in the corresponding wild types. In flc, it was also shown that ethylene production could be reduced to normal levels with exogenous ABA. In addition, the ABA-deficient mutants of tomato exhibit morphological symptoms characteristic of excess ethylene such as leaf epinasty and adventitious rooting (Tal 1966; Nagel et al. 1994). Despite these observations, the possibility that ethylene is a cause of reduced shoot growth in ABA-deficient mutants was not considered until recently.
To distinguish between these possibilities, it is necessary to assess the growth of ABA-deficient mutants at the same plant water status as the wild type. If the inhibition of growth normally observed is caused by adverse water relations, then under such conditions growth should be restored or even enhanced relative to wild-type plants. In contrast, if endogenous ABA is required to maintain shoot growth independently of effects on plant water balance, then the mutants should remain smaller. Jones et al. (1987) grew flc and two other ABA-deficient mutants of tomato, notabilis (not) and sitiens (sit), under mist in a greenhouse, and observed that stem height became greater than in the wild type, consistent with ABA playing an inhibitory role in stem elongation. In contrast, later-formed leaves remained smaller and total leaf biomass remained substantially decreased. However, leaf water potentials also remained considerably lower than in the wild type, so interpretation of the role of ABA in leaf growth was not possible. Taylor (1987) reported that the inhibition of shoot growth in double mutants of flc, not and sit was partially alleviated when the plants were grown at high humidity, but information on plant water relations was not included so, again, full interpretation was not possible. In other studies in which flc was grown at high humidity, effects on growth were not reported (Puri & Tal 1977; Tal et al. 1979).
To assess whether the reduced leaf and stem growth of well-watered flc and not are attributable to water deficits, in a recent study plants were grown under controlled-humidity conditions in a growth chamber, such that the leaf water potentials of the mutants were equal to or higher than those of wild-type plants throughout development (Sharp et al. 2000). Most parameters of shoot growth remained markedly impaired; for example, total leaf area in flc was 48% less than in the wild type (Fig. 8a). Root growth was also greatly reduced. Consistent with the observation of Jones et al. (1987), the stems of the mutants initially elongated more rapidly than those of the wild types; however, this effect was not sustained, and the mutants became shorter at later stages of development. Shoot growth substantially recovered when wild-type levels of ABA were restored by treatment with exogenous ABA (Fig. 8a,b), even though the experimental strategy prevented improvement in leaf water potential. The ability of applied ABA to increase growth was greatest for leaf expansion, which was restored by 75%. The ethylene evolution rate of growing leaves was doubled in flc compared with the wild type (Fig. 8c), and treatment with STS to inhibit ethylene action partially restored leaf, stem and root growth (Sharp et al. 2000).
Similar results have been obtained in preliminary studies of Arabidopsis. When grown at the same leaf water potentials as well-watered wild-type plants throughout development, total leaf area of the aba2 mutant (ABA-deficient) remained greatly inhibited, and was restored by about 50% in a double mutant of aba2 and etr1 (ethylene-resistant) (M.E. LeNoble, W.G. Spollen & R.E. Sharp, unpublished results). The leaf ethylene evolution rate of aba2 was approximately twice that of the wild type under these conditions.
These results demonstrate that (a) normal levels of endogenous ABA are required to maintain shoot development, particularly leaf expansion, in well-watered tomato and Arabidopsis plants independently of effects on plant water balance, and (b) the impairment of leaf and shoot growth caused by ABA deficiency is at least partly attributable to increased ethylene production.
It should be noted that these findings do not negate the necessity of avoiding differences in plant water status among genotypes in future studies of the effects of ABA manipulation on growth. Differences in water relations, when they do occur, might also contribute directly or indirectly to growth responses. It will be particularly critical (and challenging) to extend the strategy of circumventing variation in plant and also soil water status between control and ABA-deficient (or ABA-overproducing) genotypes to address the function of ABA accumulation in the responses of plant growth to soil drying.
The conclusion with well-watered tomato and Arabidopsis that endogenous ABA is required to maintain shoot growth contrasts with the many examples in the literature where ABA has caused shoot growth inhibition when applied to well-watered plants. The explanation for this difference might be simply explained by suggesting that the effects of supplemental ABA in well-watered plants are non-physiological. However, Bacon et al. (1998) published compelling evidence that the normal levels of ABA in well-watered barley can cause leaf growth inhibition. They showed that the leaf elongation rate of wild-type plants declined as the pH of artificial xylem sap fed to the leaves was increased, but that this response did not occur in the ABA-deficient mutant Az34 unless a normal well-watered concentration of ABA was supplied in the sap. (The pH-induced reduction in leaf elongation is suggested to result from increased accumulation of ABA in the apoplast.) This result is difficult to reconcile with the inhibited growth of the ABA-deficient mutants of tomato and Arabidopsis. It is possible that whole plant ABA-deficiency leads to an overriding inhibitory effect on growth of increased ethylene production. However, the above-described experiments with water-stressed maize seedlings indicate that, for root growth at least, this is not the case. In ABA-deficient seedlings in which ethylene evolution was restored to the normal water-stressed rate by treatment with AOA, root growth recovered but only to the rate of the control seedlings (Fig. 4). In other words, preventing the effect of ABA-deficiency on ethylene production did not uncover a growth-promotive effect of ABA deficiency.
Resolution of the differing conclusions on the role of ABA in leaf growth of well-watered plants from the studies of barley (Bacon et al. 1998) and of tomato and Arabidopsis will require further investigation.
Role of ABA in shoot growth response to compaction
Plants growing in compacted soil often exhibit reduced shoot growth. Evidence that endogenous ABA plays a role in shoot growth maintenance rather than inhibition in compaction-stressed barley was reported by Mulholland et al. (1996a, b). Leaf growth of plants growing in compacted soil was more inhibited in Az34 than in the wild type, and the evidence suggested that changes in leaf water relations were not the cause. Recent studies from the same group demonstrated that ethylene was a major cause of the inhibition of shoot growth in tomato plants grown with their root system divided between pots of uncompacted and compacted soil (Hussain et al. 1999, 2000). As shown in Fig. 9c, ethylene production increased greatly in wild-type plants under compaction. This response was almost fully prevented in the ethylene-deficient transgenic ACO1AS (ACC oxidase antisense; Hamilton, Lycett & Grierson 1990), as was the inhibition of leaf growth (Fig. 9a). The increase in ethylene production in the wild type occurred despite an increase in xylem sap ABA concentration (Fig. 9b). Interestingly, when supplemental ABA was supplied to the wild type, leaf ethylene production was partly reduced and leaf growth was partially restored. The same ABA treatment had a minimal effect on growth or ethylene evolution in ACO1AS, indicating that the effect of ABA on growth in the wild type was likely an indirect effect via inhibition of ethylene production. These results indicate that the normal increase in ABA in compaction-stressed plants was insufficient to fully prevent excess ethylene production in the leaves.
In addition, Hussain et al. (2000) attempted to use the ABA-deficient not mutant to demonstrate that the increase in endogenous ABA in the wild type helped to maintain leaf growth during compaction by limiting the production of ethylene. However, ethylene production was similar and inhibition of shoot growth was actually less in the mutant than in the wild type in compacted relative to non-compacted plants. Interpretation of these findings was confounded, however, because growth of the mutant was already substantially inhibited in uncompacted control plants. The inhibition in the control plants was associated with, and probably caused by, an already-increased rate of leaf ethylene evolution, consistent with the results for flc shown in Fig. 8. The problem of impaired growth of ABA-deficient control plants is discussed further below.
Does aba maintain shoot growth in water-stressed plants at later stages of development?
As described above, in well-watered tomato and Arabidopsis plants, the normal low levels of ABA are sufficient to prevent excess ethylene production. In contrast, in roots of water-stressed maize seedlings, ABA deficiency resulted in increased ethylene production although ABA levels were still substantially higher than in well-watered plants (Figs 2, 4 & 5). Similarly, in tomato plants exposed to soil compaction, leaf ethylene production increased to growth-inhibitory levels despite an increase in xylem sap ABA concentration (Fig. 9). Taken together, these results indicate that, for reasons unknown, stressed relative to non-stressed plants require an increased concentration of ABA to prevent excess ethylene production. Furthermore, in both water-stressed maize seedlings and compaction-stressed tomato plants, normal increases in endogenous ABA were insufficient to fully prevent ethylene-induced shoot growth inhibition. In contrast, ethylene was not an important cause of the inhibition of elongation in water-stressed maize roots that accumulated normal levels of ABA. Based on these findings, it seems reasonable to hypothesize that increased levels of endogenous ABA function to maintain rather than inhibit shoot growth of water-stressed plants at later stages of development, but that the accumulation of ABA may be insufficient to fully prevent ethylene-induced growth inhibition, hence contributing to the greater sensitivity of shoot relative to root growth.
Consistent with this hypothesis, there are many reports that ethylene production can be increased by plant water deficits (e.g. El-Beltagy & Hall 1974). However, Morgan et al. (1990) and Narayana, Lalonde & Saini (1991) reported that in intact cotton, bean and wheat plants subjected to slow soil drying, ethylene production rates decreased. They cautioned that many of the earlier observations might have been erroneous due to the use of excised plant parts and rapid drying. Restriction of ethylene production resulting from ABA accumulation provides a possible explanation for decreased ethylene production under water deficits. Definitive experiments to test whether increased ethylene production in drying soil, when it occurs, is a cause of the inhibition of leaf growth are lacking. In particular, despite the availability of ethylene deficient or insensitive mutants and transgenic plants of several species, there are no reports of the growth responses of these genotypes to soil drying.
To assess the role of ABA in the responses of growth to soil drying, a system in which ABA levels are adequate for normal growth under control conditions is essential. Otherwise, it may not be possible to distinguish the role of stress-induced increases in ABA concentration from the function of normal ABA levels. As described above, shoot growth is already substantially inhibited in ABA-deficient tomato and Arabidopsis mutants under non-stressed conditions independently of effects on plant water balance, at least partly due to excess ethylene production. This would confound interpretation of the effect of decreasing ABA accumulation on growth under soil drying conditions. It is also essential that root as well as shoot development is normal under non-stressed conditions. Otherwise, after withholding water, the patterns of soil drying and root stress exposure would differ between genotypes. Such an effect would probably influence the shoot growth response because of the likely involvement of non-hydraulic signals from roots in drying soil. To avoid the problem of impaired growth in control plants, a partially complemented line of the not mutant of tomato has recently been identified that exhibits normal ABA levels and shoot and root growth when well watered yet is deficient in water stress-induced ABA accumulation (E. T. Thorne, A. C. Jackson, A. Burbidge, A. J. Thompson, I. B. Taylor & R. E. Sharp, unpublished results). This genotype is being utilized to examine specifically the role of increased ABA concentrations in the response of shoot growth to soil drying.
The role of ABA in determining plant growth responses to water stress is a long-standing question. The studies of ABA-deficient genotypes described in this review have provided compelling insight into this issue. The commonality of the findings in maize, tomato and Arabidopsis suggests that restriction of ethylene production may be a widespread function of ABA, and that, as a result, endogenous ABA may often function to maintain rather than inhibit plant growth. In addition to root and shoot growth regulation, this hormonal interaction is likely to be relevant to other stress responses that are thought to involve ethylene, for example early leaf senescence and leaf, flower and fruit abscission (Morgan & Drew 1997). Moreover, because increases in ABA concentration and ethylene production occur in response to a range of environmental stress conditions, the findings may be of wide applicability.
I thank Dr Mary LeNoble for useful discussions and helpful comments on the manuscript. The work was supported by the National Science Foundation, the National Research Initiative Competitive Grants Program of the U.S. Department of Agriculture, and the University of Missouri Food for the 21st Century Program. Contribution no. 13,171 from the Missouri Agricultural Experiment Station Journal Series.
Received 12 June 2001; received in revised form 15 August 2001; accepted for publication 15 August 2001