Various data indicate that nitric oxide (NO) is an endogenous signal in plants that mediates responses to several stimuli. Experimental evidence in support of such signalling roles for NO has been obtained via the application of NO, usually in the form of NO donors, via the measurement of endogenous NO, and through the manipulation of endogenous NO content by chemical and genetic means. Stomatal closure, initiated by abscisic acid (ABA), is effected through a complex symphony of intracellular signalling in which NO appears to be one component. Exogenous NO induces stomatal closure, ABA triggers NO generation, removal of NO by scavengers inhibits stomatal closure in response to ABA, and ABA-induced stomatal closure is reduced in mutants that are impaired in NO generation. The data indicate that ABA-induced guard cell NO generation requires both nitric oxide synthase-like activity and, in Arabidopsis, the NIA1 isoform of nitrate reductase (NR). NO stimulates mitogen-activated protein kinase (MAPK) activity and cGMP production. Both these NO-stimulated events are required for ABA-induced stomatal closure. ABA also stimulates the generation of H2O2 in guard cells, and pharmacological and genetic data demonstrate that NO accumulation in these cells is dependent on such production. Recent data have extended this model to maize mesophyll cells where the induction of antioxidant defences by water stress and ABA required the generation of H2O2 and NO and the activation of a MAPK. Published data suggest that drought and salinity induce NO generation which activates cellular processes that afford some protection against the oxidative stress associated with these conditions. Exogenous NO can also protect cells against oxidative stress. Thus, the data suggest an emerging model of stress responses in which ABA has several ameliorative functions. These include the rapid induction of stomatal closure to reduce transpirational water loss and the activation of antioxidant defences to combat oxidative stress. These are two processes that both involve NO as a key signalling intermediate.
In recent years, nitric oxide (NO) has emerged as an important endogenous plant signalling molecule that mediates many developmental and physiological processes (Lamattina et al., 2003; Neill et al., 2003; Delledonne, 2005; Lamotte et al., 2005). Here, the focus is on the role of NO during stomatal closure and in particular on its role as an signalling intermediate involved in the guard cell movements induced by the stress hormone and endogenous anti-transpirant, abscisic acid (ABA). Recent data indicating that NO activates antioxidant defences during oxidative stress and that in these cases there may be a similar link between NO and ABA are also discussed.
Water stress or, strictly speaking, water deficit stress, which is often referred to as drought stress, is a major abiotic condition that impacts dramatically on plant and crop growth and yield. Water stress occurs in plants growing in drying soil as the water lost from the leaves exceeds that taken up by the roots and results in cellular dehydration, damage, and ultimately death. Cellular dehydration can also occur during the exposure of plants to other abiotic stresses that restrict water supply, such as, for example, either during cold and salt stress or during the anaerobic conditions resulting from root flooding (Fig. 1). Several defence responses are activated by water stress. One of the most important of these is stomatal closure induced by the redistribution and synthesis of ABA. ABA also modulates the expression of gene networks that control other ameliorative responses. These include the maintenance of root water uptake, the synthesis of osmoprotective proteins such as dehydrins, and various metabolic changes (Fig. 1) (Zhu, 2002; Chaves et al., 2003; Seki et al., 2007). Oxidative stress is a common feature of several abiotic stresses including water stress (Bailey-Serres and Mittler, 2006). During oxidative stress the redox balance of cells is disturbed by increases in the rate of generation of reactive oxygen species (ROS) such as the superoxide anion (O2·−) and hydrogen peroxide (H2O2) above that of their removal by antioxidant enzymes or by reaction with antioxidant molecules. Cell functions are altered during oxidative stress not only because of oxidative damage per se, but also because ROS themselves are centrally important signalling molecules. Thus, excessive quantities of ROS result in aberrant cell signalling (Bailey-Serres and Mittler, 2006). The activation of cellular antioxidant systems is a common feature of oxidative stress (Fig. 1), and there is increasing evidence which indicates that NO is a critical factor in such responses.
Until recently there appeared to be two well-defined plant enzymes capable of NO biosynthesis in plants, nitric oxide synthase (NOS) and nitrate reductase (NR) (Crawford, 2006). Additionally there may be other sources that appear to involve neither of these enzymes (Arnaud et al., 2006).
An apparent plant NOS (AtNOS1) was first identified in Arabidopsis (Guo et al., 2003). The enzyme had homology to a snail protein potentially involved in NO synthesis, and had some of the biochemical characteristics expected of a NOS. It increased NO synthesis when overexpressed in Escherichia coli and, as it had NOS activity when assayed with a commercial kit in which L-arginine is converted to L-citrulline, it appeared to use this amino acid as a substrate for NO generation. Importantly, Guo et al. (2003) identified a T-DNA insertion mutant, Atnos, that showed reduced root and guard cell NO synthesis in response to ABA. Moreover, young Atnos1 seedlings have a chlorotic phenotype that can be rescued by exogenous NO. Several reports have confirmed that Atnos1 does indeed show reduced NO accumulation and is impaired in its ability to generate NO in response to various stimuli (Zeidler et al., 2004; Bright et al., 2006; M Zhao et al., 2007; Zottini et al., 2007). Recently, however, the nature and function of AtNOS1 have been questioned and, although it is does appear to be required for normal NO synthesis to occur, it now appears that it is not in fact a NOS at all (Crawford et al., 2006; Guo, 2006; Zemojtel et al., 2006). In attempting to reproduce earlier findings, Zemojtel et al. (2006) failed to detect any NOS activity in purified recombinant AtNOS1 protein or in similar recombinant proteins encoded by orthologous genes from rice and maize. Thus, since NO synthesis and responses that require NO are undoubtedly impaired in the Atnos1 mutant, it has been re-named Arabidopsis thaliana nitric oxide-associated 1 (Atnoa1). In reality, therefore, no cloned NOS-like enzyme has been found in plants. However, there are many reports in which inhibitors of mammalian NOS such as NG-nitro-L-arginine methyl ester (L-NAME) have been shown, in correlation with an inhibition of NO production, to inhibit various processes in plants, and there is some biochemical evidence for the existence of plant enzymes that use L-arginine to generate NO (Neill et al., 2003; Lamotte et al., 2005; Crawford, 2006; Jasid et al., 2006). Thus, the implication is that NOS-like enzymes do exist in plants. However, the biochemistry of these enzymes and how they utilize L-arginine remain to be clarified.
The other characterized enzymatic source of NO is NR. The primary function of NR in plants is to assimilate nitrogen (N) by converting nitrate to nitrite. However, as shown originally in soybean (Dean and Harper, 1988) and also in vitro and in vivo in Arabidopsis and other species (Neill et al., 2003; Bright et al., 2006; Crawford, 2006), in an NAD(P)H-dependent reaction NR can also convert nitrite to NO. In Arabidopsis, NR is encoded by two genes, NIA1 and NIA2. Modolo et al. (2006) recently reported that the lack of NR activity in the nia1nia2 double mutant leads to reduced levels of L-arginine and that exogenous L-arginine can restore NO generation in this mutant. This suggests that aberrant N metabolism due to a lack of NR might somehow affect NO production via an arginine-dependent process. However, as is detailed below, in Arabidopsis guard cells it would appear that the NIA1 isoform of NR preferentially accounts for NO synthesis during ABA-induced stomatal closure and does so in a background where N metabolism is unlikely to be disturbed.
A plasma membrane-bound, root-specific enzyme, nitrite:NO oxidoreductase (Ni-NOR), may function as a further source of NO. This enzyme was identified biochemically as a result of its NO-generating activity. Unlike NR, it does not use NAD(P)H as a cofactor, but uses cytochrome c as an electron donor in vitro, and its pH optimum is more acidic than that of NR. However, neither its physiological role nor its genetic identity is yet known (Stohr and Stremlau, 2006).
Recent work has also suggested that, in addition to NR-mediated nitrite-dependent NO production, electron transport processes in mitochondria and chloroplasts can also generate NO from nitrite (Gupta et al., 2005; Modolo et al., 2005; Planchet et al., 2005; Jasid et al., 2006). Additionally, apoplastic conversion of nitrite to NO at low pH has been demonstrated in the barley aleurone layer (Bethke et al., 2004).
NO- and ABA-induced stomatal closure
Exogenous NO, supplied in the form of NO donors, was first shown to induce stomatal closure and reduce transpiration in species such as Vicia faba, Salpichroa organifolia, and Tradescantia spp by Garcia-Mata and Lamattina (2001). Such findings were confirmed by work in our laboratory demonstrating that NO donors induced stomatal closure in epidermal peels or leaves of pea (Neill et al., 2002), Arabidopsis thaliana (Desikan et al., 2002), and various other species (S Neill et al., unpublished results). Many different NO donors induce stomatal closure in a dose- and time-dependent manner, and their effects can be reversed by simultaneous co-incubation with the NO scavengers 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (PTIO) or 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (cPTIO) (Bright, 2006). Many such plant studies have used the NO donor sodium nitroprusside (SNP) that also generates cyanide. However, neither cyanide itself nor light-inactivated SNP induce stomatal closure (Bright, 2006), and cPTIO does not react directly with cyanide (Bethke et al., 2006). Moreover, the reaction product of cPTIO and NO, cPTI, which has some biological activity per se (Planchet and Kaiser, 2006b), does not inhibit ABA-induced guard cell movements (I Wilson et al., unpublished results). Thus, the effects mediated by NO donors such as SNP must indeed be due to the release and biological activity of NO.
The requirement for endogenous NO during ABA-induced stomatal closure has been demonstrated by both pharmacological and genetic approaches that either remove NO or prevent its generation and correlate stomatal responses with estimates of its accumulation. It has been shown that pea and Arabidopsis guard cells generate NO in response to ABA and that their treatment with PTIO both removes NO and inhibits ABA-induced stomatal closure (Fig. 2). Similar guard cell responses have been shown in V. faba (Garcia-Mata and Lamattina, 2002). The treatment of Arabidopsis leaves with either L-NAME or tungstate, compounds that potentially inhibit NO generation from L-arginine or nitrite, inhibits guard cell NO production and stomatal closure (Bright et al., 2006). Moreover, in the Atnos1 and nia1nia2 mutants, both ABA-induced stomatal closure and ABA-induced NO generation are impaired (Desikan et al., 2002).
Much of the aforementioned work has used the NO-sensitive fluorescent dye 4,5-diaminofluorescein diacetate (DAF-2DA) and related compounds to detect the presence of NO. However, as highlighted recently by Kaiser and colleagues (Planchet and Kaiser, 2006a, b), DAF-2DA can generate artefactual fluorescence arising from its reaction with compounds other than NO. Nonetheless, the use of appropriate control dyes such as 4-aminofluorescein diacetate (4AF-DA), that do not react with NO (Fig. 2), and the NO scavenger PTIO allows the conclusion that NO is truly being detected in guard cells. The use of other NO assays such as electron paramagnetic resonance spectroscopy or photoacoustic laser detection is difficult within the guard cell system. However, it has been possible to demonstrate ABA-enhanced NO generation in pea epidermal peels using the haemoglobin assay (Neill et al., 2002). Thus, data from various sources do demonstrate that ABA-induced stomatal closure requires NO generation and action.
Recent reports indicate that guard cells also generate NO in response to elicitor challenge (Melotto et al., 2006) and methyl jasmonate (MeJA) (Munemasa et al., 2007), treatments that both induce stomatal closure. As with the effect of ABA, in these cases the prevention of NO accumulation also inhibited the associated stomatal closure.
Clearly there are several potential mechanisms by which NO could be generated in guard cells in response to ABA and other stimuli. ABA-induced stomatal closure and guard cell NO generation are both inhibited by potential NOS inhibitors such as L-NAME (Neill et al., 2002; Guo et al., 2003; She et al., 2004; Bright et al., 2006), and both stomatal closure and ABA-induced guard cell NO generation are impaired in the Atnos1 mutant (Guo et al., 2003; Bright et al., 2006). Thus, whatever its exact function may turn out to be, the requirement for AtNOS1 (AtNOA1) is clear, as is a requirement for at least one mechanism that is inhibited by L-arginine analogues.
However, in Arabidopsis at least, there are also data indicating that ABA-induced NO synthesis in guard cells is mediated by NR. The NR substrate, nitrite, induces stomatal closure and NO production in a dose-dependent manner (Desikan et al., 2002; Bright, 2006; Bright et al., 2006). NO generation by NR is inhibited in vitro by tungstate, probably as a result of its competitive binding to the molybdenum-binding site in the enzyme. Tungstate also inhibits either ABA- or nitrite-induced NO generation and stomatal closure in vivo (Bright, 2006; Bright et al., 2006). The NR-deficient nia1nia2 mutant, which contains only 0.5% of wild-type NR activity (Wilkinson and Crawford, 1993), exhibits neither guard cell NO synthesis nor stomatal closure in response to either ABA or nitrite, indicating that one or both of the two Arabidopsis NR isoforms are required,. The observation that nitrite fails to induce NO synthesis in nia1nia2 guard cells suggests that this requirement for NR reflects its in vivo ability to produce NO from nitrite. However, as raised by Crawford (2006), the lack of NR may have several effects on plant N metabolism that could potentially limit NOS-like activity and, as mentioned previously, Modolo et al. (2006) have reported that levels of the NOS substrate, L-arginine, are substantially reduced in nia1nia2 leaves. In resolving this issue, the Arabidopsis single mutants, nia1 and nia2, have been used to study further the involvement of NR in guard cell NO synthesis during ABA-induced stomatal closure (Bright, 2006). Of the two isoforms, NIA2 accounts for 90% of the total nitrate assimilatory activity of NR in seedlings, while NIA1 accounts for the remaining 10% (Wilkinson and Crawford, 1991). The sequences of the Arabidopsis mRNAs encoding the two isoforms have ∼70% identity and give rise to proteins which are 83.5% similar at the amino acid level (Wilkinson and Crawford, 1993). Both genes are probably expressed in guard cells (Bright, 2006). Guard cell NO synthesis and stomatal responses to ABA were examined in the loss-of-function T-DNA insertion mutant, nia1::Ds (Parinov et al., 1999), which has no NIA1 activity (Wang et al., 2004), and in the loss-of-function deletion mutant, nia2-5 (Wilkinson and Crawford, 1991). Nitrite-induced stomatal closure was abolished in the nia1::Ds plants and ABA-induced stomatal closure was substantially reduced. NO accumulation in the guard cells of this mutant in response to these stimuli was effectively absent in comparison with that of the same cells of wild-type leaves. nia1::Ds stomatal closure in response to SNP was as for the wild type, indicating that these aberrant stomatal responses were not related to a general impairment in guard cell movement,. In contrast to the nia1::Ds mutant, stomata of the nia2-5 mutant behaved as those of wild type, closing in response to either ABA or nitrite, and their guard cells showed wild-type levels of NO accumulation after such stimulation. Thus, it would seem highly likely that the activity of NIA1, but not NIA2, is required for effective ABA signal transduction in guard cells during stomatal closure (Bright, 2006). These data would also suggest that the aberrant NO biology in the nia1nia2 mutant is due specifically to a lack of NIA1 as opposed to generally aberrant N metabolism. Interestingly, some years ago Yu et al. (1998) also concluded that NIA1 and NIA2 have distinct signal transduction and N assimilatory roles. NIA1 and NIA2 have similar amino acid sequences, but there are distinct regions that differ between the two proteins. Thus, key questions must concern their subcellular localization, interacting protein partners, and activation characteristics.
NO signal transduction in guard cells
Given that exogenous and endogenous NO induces stomatal closure, there must be some mechanism(s) by which NO is perceived in guard cells and the signal subsequently transduced into the alterations in membrane transport that effect ion and water movement across the tonoplast and plasma membrane. The small size, ready diffusibility, and reactive nature of NO perhaps make the ligand-specific, proteinaceous receptor interaction paradigm unlikely. On the other hand, the reactive nature of NO does make it likely that it reacts, directly or indirectly, with either several or perhaps many proteins. Thus, there may be a number of different ‘NO perceptors’ that exist in different cell types and during different physiological states. On the basis of mammalian work and emerging studies in plant systems, it is probable that both the transient activation of soluble guanylyl cyclase (sGC), which generates the second messenger cGMP, and the reversible modulation of protein activity via S-nitrosylation-induced conformational changes (Delledonne, 2005; Wang et al., 2006) are major components of NO signalling.
The evidence that cGMP is an NO signalling intermediate has been obtained in several systems (Neill et al., 2003; Delledonne, 2005). Both salt and osmotic stress, two conditions which would both be expected to induce ABA synthesis, induced a rapid increase in the cGMP content of Arabidopsis seedlings (Donaldson et al., 2004). Using a sensitive radioimmunoassay technique, the cGMP content of pea epidermis and Arabidopsis guard cell fragments has been similarly measured as being in the pmol g−1 range, and transient increases in cGMP levels following either ABA or SNP treatment have been observed that could be prevented by co-incubation of the treated tissues with PTIO (J Harrison et al., unpublished reults). Further evidence that cGMP can mediate the effects of ABA in stomatal guard cells has resulted from pharmacological work using the cell-permeable cGMP analogue 8-bromo cGMP (8BrcGMP) and inhibitors of NO-sensitive sGC such as 1H-[1,2,4] oxadiazolo[4,3-a]quinoxalin-1-one (ODQ). The ABA- or NO-induced closure of pea (Neill et al., 2002) and Arabidopsis (J Harrison et al., unpublished results) stomata was inhibited by ODQ. This inhibition could be prevented by co-incubation of the ABA-/NO-stimulated, ODQ-treated guard cells with 8BrcGMP. However, 8BrcGMP alone did not induce stomatal closure. Thus, it would appear that although an elevated level of cGMP is required for effective ABA-induced stomatal closure, additional signalling pathways stimulated by ABA must operate in concert for such an increase to mediate its effects.
In non-plant systems, cADP ribose (cADPR), an agent that mobilizes Ca2+ from internal stores, is a downstream messenger of NO. Nicotinamide, a potential inhibitor of cADPR synthesis, blocks ABA- and NO-induced stomatal closure (Neill et al., 2002). Garcia-Mata et al. (2003) have also shown that NO-induced intracellular Ca2+ release and the regulation of guard cell plasma membrane K+ and Cl− channels are mediated by a cGMP- and cADPR-dependent pathway.
cGMP may also signal by binding to and directly activating cyclic nucleotide-gated ion channels (CNGCs) or by similarly activating cGMP-dependent protein kinases. To date, no cGMP-activated plant protein kinases have been identified, and the potential role of CNGCs in guard cell NO signalling awaits clarification.
Although cGMP has been unequivocally identified in various plant tissues (Neill et al., 2003) and pharmacological data indicate a role for it during stomatal closure, very little is known of the mechanisms by which cGMP might be turned over in plant cells. The one plant guanylyl cyclase gene (AtGC1) cloned from Arabidopsis encodes a protein which shows many domain differences from mammalian sGC, lacks a haem-binding motif, and is insensitive to NO (Ludidi and Gehring, 2003). The biological roles of AtGC1 remain to be described and it may be that plants contain other enzymes capable of cGMP production. Interestingly and indicating that there may well be novel plant guanylyl cyclases awaiting discovery, a recent report has indicated that the Arabidopsis brassinosteroid receptor, AtBRI1, contains a domain with guanylyl cyclase activity (Kwezi et al., 2007). If cGMP is an intracellular plant signal, then mechanisms for its rapid degradation will exist. Plant cells do indeed possess cGMP hydrolysis activity and the Arabidopsis genome contains several genes for potential phosphodiesterases including that encoding a putative cGMP phosphodiesterase (GenBank accession no. NM_118011) (Maathuis, 2006).
NO and its related species can oxidize, nitrate, or nitrosylate proteins (Mur et al., 2006; Wang et al., 2006). Peroxynitrite, formed by the reaction of NO with superoxide, can oxidize proteins on cysteine, methionine, or tryptophan residues or nitrate tyrosine residues to form nitrotyrosine. These post-translational modifications may well turn out to have roles in intracellular signalling and the subsequent physiological effects. For example, recent work indicates a role for protein tyrosine nitration in plant defence responses (Saito et al., 2006).
S-nitrosylation is the reversible covalent attachment of NO to the thiol group of cysteine residues forming an S-nitrosothiol (SNO) and may well be an ancient highly conserved cell signalling mechanism (Wang et al., 2006). Nitrosylation can occur either directly through the interaction of NO and other NO-related species with the cysteine group or indirectly by trans-nitrosylation where the NO is derived from S-nitrosglutathione (GSNO) or other S-nitrosylated proteins. Some recent studies indicate that this redox-based mechanism plays a pivotal role in plant biology and will, therefore, also be important with regard to NO signalling in guard cells.
S-nitrosylated proteins can be detected using the ‘biotin switch assay’ which is based on specifically biotin labelling any S-nitrosylated cysteines within proteins. This assay is useful for demonstrating the existence of potentially S-nitrosylated proteins but, like many assays, may be prone to identifying false positives. However, S-nitrosylation can be demonstrated unequivocally using mass spectrometry (Lindermayr et al., 2006). The biotin switch assay has been used to show the presence of S-nitrosylated proteins in plants and to demonstrate that specific proteins can be S-nitrosylated (Lindermayr et al., 2005, 2006; Wang et al., 2006; Belenghi et al., 2007). A large number of potentially S-nitrosylated proteins have been identified which include stress-related proteins, redox-related proteins, signalling proteins, cytoskeletal proteins, and proteins involved in photosynthesis and metabolism (Lindemayr et al., 2005). Moreover, conserved protein S-nitrosylation and GSNO-binding motifs are present in plant proteins (Wang et al., 2006), and the effects of S-nitrosylation on protein activity and plant physiology are now being addressed (Lindermayr et al., 2006; Belenghi et al., 2007). It would seem likely that all cells contain nitrosylated proteins and that the spectrum and levels of these, the ‘nitrosylome’, will alter during NO accumulation in, for example, guard cells after their challenge with ABA. In fact some evidence that this is the case already exists, and data published by Sokolovski and Blatt (2004) suggest that NO regulates outward rectifying K+ channels in V. faba guard cells by S-nitrosylation.
Alterations in the calcium concentrations of specific cellular compartments and reversible protein phosphorylation are two processes central to cell signalling. Thus, it is not surprising that NO is likely to affect these processes in guard and other cells. ABA and NO activate mitogen-activated protein kinases (MAPKs) (Desikan et al., 2004a; Pagnussat et al., 2004; A Zhang et al., 2007), and MAPK activation has been suggested as a convergence point for guard cell H2O2 and NO signalling during ABA-induced stomatal closure (Desikan et al., 2004a). NO also activates other protein kinases such as the osmotic stress-activated kinase, NtOSAK, in tobacco (Lamotte et al., 2006). NO affects calcium mobilization (Lamotte et al., 2006), and Sokolovski et al. (2005) demonstrated a requirement for both protein phosphorylation and altered calcium flux in order for normal NO responses to occur in guard cells. Further evidence that NO signalling requires protein dephosphorylation results from the observation that stomata of the Arabidopsis ABA-insensitive phosphatase mutant, abi1, are insensitive to NO (Desikan et al., 2002).
NO and interactions with hydrogen peroxide
It is commonly observed that NO and ROS such as superoxide and H2O2 are generated in response to similar stimuli and with similar kinetics. NO and ROS can interact in various ways. For example, they can interact chemically as in the formation of compounds such as peroxynitrite. NO may also affect the activities of enzymes that alter ROS levels, and vice versa. Thus, they could both impact either negatively or positively on the same or related signalling pathways and thereby lead to additive and possibly synergistic responses. Stomatal closure in response to ABA is one such example where this occurs. ABA induces the production of H2O2. In Arabidopsis guard cells this is likely to arise from superoxide generated by the isoforms of NAD(P)H oxidase encoded by the genes AtrbohD and AtrbohF (Kwak et al., 2003; Desikan et al., 2004b; She et al., 2004). Consequently, the simultaneous presence of both superoxide and NO in the same location may lead to the generation of peroxynitrite anions (Neill et al., 2003). Peroxynitrite may well have unique signalling properties, but as yet there is no evidence that it affects guard cell movements, and the co-incubation of ABA-treated leaves with a peroxynitrite scavenger does not inhibit the induction of stomatal closure (Bright, 2006).
Lum et al. (2002) observed that exogenous H2O2 induced NO generation in the guard and other cells of Phaseolus aureus leaves. In a chemiluminescence-based assay, the H2O2 induced a substantial increase in an apparent NOS-like activity. This increase was reduced by using a potential NOS inhibitor. Interestingly, the H2O2-induced NO generation was inhibited by the calcium channel blocker verapamil. Thus, it is possible that Ca2+ ions may mediate this effect of H2O2. She et al. (2004) and He et al. (2005) also reported that H2O2 induced NO synthesis in V. faba guard cells and that again this accumulation could be reduced by a NOS inhibitor. Moreover, they found that both NO generation and stomatal closure stimulated by darkness or UV-B were dependent on guard cell synthesis of H2O2. Similarly, in Arabidopsis, H2O2 induces guard cells to synthesize NO which is required if the induction of stomatal closure is to follow (Bright et al., 2006). ABA-induced NO synthesis is dependent on prior H2O2 generation in both V. faba (Dong et al., 2005) and Arabidopsis (Bright et al., 2006). Removal of the H2O2 with antioxidants or inhibition of its synthesis by inhibiting NAD(P)H oxidase activity prevents NO generation and stomatal closure. Similarly, removal of the NO using PTIO prevents either H2O2- or ABA-induced stomatal closure. Guard cells of the AtrbohD/F double mutant (Kwak et al., 2003) also fail to make NO in response to ABA, as do those of the NR-deficient nia1nia2 mutant. However, H2O2 stimulation of guard cell NO accumulation in the Atnos1 (Atnoa1) mutant was as in the wild type (Bright et al., 2006). Thus, the requirement of AtNOS1 for NO synthesis must be epistatically upstream of H2O2 and its signalling effects. Regardless, the current evidence would suggest that H2O2 induces guard cell NO synthesis through the activity of NR. More detailed analysis has demonstrated that only the NIA1 isoform is required, as H2O2-stimulated NO generation is as wild type in the nia2-5 mutant (Bright, 2006). Critically, it has been suggested that the presence of oxidants could increase the sensitivity of the DAF-2DA used in these studies. Thus, the implication is that increased concentrations of H2O2 might result in an increase in the fluorescence of this dye without any actual increase in the level of NO (Planchet and Kaiser, 2006a, b). While this is difficult to discount completely in vivo, it would appear that H2O2 does not increase the sensitivity of DAF-2DA in vitro (Bright et al., 2006). Furthermore, DAF-2DA fluorescence does not increase in the guard cells of the nia1nia2 mutant following ABA or H2O2 challenge. The fact that this mutant generates increased H2O2 levels in guard cells in response to ABA strongly suggests that H2O2 does indeed induce NO production and that it this that is measured by the increased DAF-2DA fluorescence.
It also appears that the ABA–H2O2–NO cascade is not restricted to guard cells. A Zhang et al. (2007) have now demonstrated that it also operates in the mesophyll cells of Zea mays leaves and that the effects of ABA and H2O2 on the induction of antioxidant defences depended on NO generation and could be reduced using cPTIO or L-NAME.
While it is well established that H2O2 induces NO synthesis and accumulation, there has also been some suggestion that NO may modulate H2O2 levels. Both She et al. (2004) and He et al. (2005) have reported that this was the case in V. faba guard cells. However, Dong et al. (2005), also working with V. faba guard cells, did not observe this phenomenon. NO does not appear to induce H2O2 generation in Arabidopsis guard cells (Bright et al., 2006) or in maize mesophyll cells (A Zhang et al., 2007). Thus, there is argument as to whether or not this occurs, but feedback mechanisms and autocatalysis are part and parcel of signalling pathways, and it may be that such discrepancies simply reflect differences in the physiological states of the tissues examined.
Stomatal closure during water stress
Although Magalhaes et al. (2000) reported decreased NO emission following wilting, other studies have shown that water stress causes enhanced NO synthesis and accumulation in guard cells (Leshem and Haramaty, 1996; Garcia-Mata and Lamattina, 2002; Gould et al., 2003). Recently such findings have been confirmed by A Zhang et al. (2007) who showed that polyethylene glycol-induced water stress stimulated NO production in guard cells as a result of increased ABA synthesis. However, the effects of dehydration and turgor loss on the level and emission of NO in and from plants, respectively, certainly requires clarification. ABA synthesis is undoubtedly induced by turgor loss. Thus, the question arises as to whether or not NO plays a direct role in mediating drought-induced stomatal closure. Recent work in our laboratory (D Ribeiro et al., unpublished results) indicates that NO is not required for rapid water stress-induced stomatal closure. Rapid dehydration of detached Arabidopsis leaves caused equally rapid stomatal closure. This was not inhibited by pre-incubation of the leaves in PTIO and occurred in both the nia1 and nia1nia2 mutants in the same way as in leaves from wild-type plants. The stomata of the abi1 mutant do not close at all during this treatment, indicating that the stomatal closure response is active and ABA dependent,. Thus, in contrast to the requirement for NO synthesis during ABA-induced stomatal closure in well-hydrated leaves, stomatal closure during rapid wilting may result from ABA signalling that bypasses the need for NO synthesis and perception. It may be different during stomatal closure induced in intact plants by a more gradual reduction in water potential, but it is worth noting that neither the Atnos1 nor the nia1 mutants are obviously wilty. It is also possible that the effects of NO differ depending on its concentration and the physiological state of the guard cell. For example, Sakihama et al. (2003) reported that NO donors actually caused stomatal opening when used at high concentrations. Similar results have been obtained in that, unlike low (0.1–50 μM) concentrations of SNP, high (0.5–1 mM) concentrations failed to cause stomatal closure (Bright, 2006). Another interesting occurrence that has yet to be explained concerns the interaction of ABA, ethylene, and NO signalling in guard cells. ABA, ethylene, and NO all induce stomatal closure when applied alone, but stomatal closure is reduced when either ABA or NO is combined with ethylene (Desikan et al., 2006), a situation likely to arise during the wilting of plants. Whichever, NO does appear to have ameliorative benefits in warding off the effects of drought. Garcia-Mata and Lamattina (2001) reported that prolonged SNP treatment provided plants with such beneficial long-term effects. Wheat plants watered with SNP and then exposed to water stress had higher relative water contents than those not receiving the SNP pre-treatment. Presumably the long-term exposure to NO in some way enables plants to retain more water during subsequent dry periods.
Stomatal movements during light/dark transitions
The inhibition of stomatal closure by cPTIO when bean leaves are transferred into the dark coupled to increased fluorescence of dark-incubated guard cells stained with DAF-2DA indicates that NO is involved in light/dark-regulated stomatal movements (She et al., 2004). As also reported for pea leaves (Desikan et al., 2004b) and indicating the involvement of H2O2, this dark-induced stomatal closure could also be inhibited by the addition of ascorbate or exogenous catalase. Adding weight to the concept that NO is involved in dark-induced stomatal closure, stomata of the Arabidopsis nia1 mutant fail to close after transfer from the light to the dark, and their guard cells show markedly reduced DAF-2DA fluorescence compared with those of the wild type which close after such treatment. Re-opening of Arabidopsis stomata on transfer from the dark to the light can be prevented by the addition of ABA or the NO donor SNP (Bright, 2006; I Wilson et al., unpublished results). The addition of the NO scavenger PTIO effectively prevents this ABA inhibition of re-opening. It is generally considered that the dark-induced cessation of photosynthesis allows the pH gradient between the chloroplast and cytosol to dissipate such that the cellular distribution of ABA changes, with more accumulating in the cytosol. At this point the redistributed ABA may induce stomatal closure directly in the guard cells or indirectly after its movement from the mesophyll to epidermal cells. These data suggest that stomatal closure on transfer from the light to dark involves a linear signalling process from H2O2 and NO and that their re-opening, either on transfer to the light or on return to conditions of hydration, may also be partly dependent on the metabolic removal of H2O2 and NO. The degree to which stomata close in the dark depends on numerous factors including the stomatal density and leaf age, the position of the plant in its circadian rhythm, the light intensity and quality of the preceding photoperiod, the ambient CO2 concentration, the relative humidity and soil water availability, and whether the plant operates a C3, C4, or crassulacean acid metabolism (Caird et al., 2007). Thus, the involvement of NO with respect to these variables remains to be fully determined.
Stomatal responses to phytopathogenic bacteria
Gaining entry to subepidermal tissues is one of the first steps in the bacterial infection of plants, and it is often assumed that openings such as stomata serve as passive ports of entry for such pathogens. However, active stomatal closure has recently been described as part of the innate immunity of plants to bacteria (Melotto et al., 2006). For example, in what has been termed a pathogen-associated molecular pattern (PAMP)-induced basal defence mechanism, Arabidopsis guard cells are able to perceive Pseudomonas syringae cell surface molecules such as lipopolysaccharide (LPS) and flagellin which trigger NO synthesis and result in stomatal closure. Inhibitors of NO synthesis prevent this stomatal closure. In contrast to those of the wild type, stomata of either the ABA-insensitive stomata mutant, ost1-2, or the ABA-deficient mutant, aba3-1, failed to close in response to LPS and flagellin. Escherichia coli also induced stomatal closure that was prevented by inhibitors of NO synthesis. Thus, it would appear that the ABA to NO signalling mechanism is required for stomatal closure during bacterial challenge. Interestingly, virulent pathogenic bacteria have a mechanism to overcome this barrier and cause stomata to re-open. Using mutants of P. syringae which were unable to elicit stomatal re-opening, Melotto et al. (2006) identified the phytotoxin coronatine (COR) as the bacterial virulence effector responsible for overcoming this barrier to infection. COR effectively prevented ABA-induced stomatal closure in wild-type Arabidopsis, but not in the COR-insensitive mutant, coi1-20. COR did not prevent ABA-induced NO synthesis in wild-type plants. Thus, it is likely that COR acts downstream of ABA and NO to prevent stomatal closure (Melotto et al., 2006).
NO as an antioxidant and protectant
Exogenous NO, applied via the roots in the form of SNP, has been shown to provide some protection against the oxidative damage associated with water stress (Garcia-Mata and Lamattina, 2001). Several other studies have also shown that exogenous NO ameliorates the oxidative stress induced by a range of abiotic conditions including those resulting from the presence of heavy metal ions, salinity, high temperatures, the presence of H2O2, dehydration, UV irradiation, and the presence of paraquat (Cheng et al., 2002; Hung et al., 2002; Uchida et al., 2002; Kopyra and Gwóźdź, 2003; Zhang et al., 2003, 2006; Hsu and Kao, 2004; Hung and Kao, 2005; Laspina et al., 2005; Shi et al., 2005; Wang and Yang, 2005; Yu et al., 2005; Tian et al., 2007). Some of the antioxidant effects of NO may be due to its direct interaction with ROS such as superoxide to form peroxynitrite that might then be scavenged by other cellular processes. In other cases NO may enhance the antioxidant capacity of cells by increasing the activities of antioxidant enzymes such as superoxide dismutase, which converts superoxide to H2O2, and catalase and ascorbate peroxidase, which both remove H2O2 (Cheng et al., 2002; Hung et al., 2002; Uchida et al., 2002; Kopyra and Gwóźdź, 2003; F Zhang et al., 2007).
Whether or not endogenous NO has an antioxidant function is debatable. Guo and Crawford (2005) detected greater levels of oxidized proteins and lipids in the Atnos1 mutant. Atnos1 has enhanced sensitivity to the oxidative stress induced by either methyl viologen or salinity that can be ameliorated by exogenous NO (M Zhao et al., 2007; MG Zhao et al., 2007). Moreover, in calli of stress-tolerant and stress-sensitive reed ecotypes, heat and salt stress induce endogenous NO generation in the tolerant, but not in the sensitive ecotype. NO enhanced stress tolerance in both ecotypes, but more so in the sensitive ecotype, and depletion of endogenous NO by cPTIO reduced survival in the tolerant ecotype. The enhanced antioxidant enzyme activity and stress tolerance associated with the ability of the tolerant ecotype to make NO was abolished by this cPTIO treatment (Zhao et al., 2004; Song et al., 2006). Similarly, in Populus callus and Stylosanthes guianensis, stress tolerance has been associated with endogenous NO synthesis (Zhou et al., 2005; F Zhang et al., 2007).
Given that various stresses induce ABA and, as a consequence NO, and given that NO appears to enhance antioxidant enzyme activity, one might expect a signalling link between ABA, NO, and oxidative stress tolerance. In S. guianensis, ABA enhances the activity of antioxidant enzymes, and this is dependent on NO production (Zhou et al., 2005). Recent work by A Zhang et al. (2007) has clearly demonstrated this link. Their data show that the enhanced MAPK activity and antioxidant gene expression and enzyme activity induced in maize leaves by ABA and H2O2 are in fact dependent on endogenous NO generation. Thus, the data indicate a key ‘ABA–H2O2–NO–MAPK–antioxidant survival cycle’ and suggest that during water stress ABA has several ameliorative functions that involve NO as a key signalling intermediate and which include the rapid induction of stomatal closure to reduce transpirational water loss and the activation of antioxidant defences to combat oxidative damage (Fig. 1).
Conclusions and perspectives
The data available to date show that NO is a key factor involved in stomatal closure in response to ABA and probably in response to other stimuli such as the light/dark transition and exposure to phytopathogens. They suggest that in some cases, such as in response to ABA, NO generation is dependent on the production of H2O2. They indicate that endogenous NO is a key factor in the tolerance of cells to the oxidative stress induced by a range of abiotic conditions including water stress, and that this probably involves the enhanced expression of genes encoding antioxidant enzymes. They suggest that during water stress ABA has several ameliorative functions that involve NO as a key signalling intermediate and which include the rapid induction of stomatal closure and the activation of antioxidant defences.
Of course, there are numerous unanswered questions and important areas for further research. The mechanisms by which NO is generated are still largely unresolved, and elucidation of how it is made by different plant cells in different situations is clearly a research priority. How ABA perception by the recently characterized ABA receptors, two of which mediate stomatal movements in response to ABA (Shen et al., 2006; Liu et al., 2007) and one which apparently does not (Razem et al., 2006), is transduced into elevated NO production is not yet known. Even though NO is a short-lived molecule that reacts with oxygen, it is an extremely active one. Thus, the mechanisms by which either NO or molecules participating in NO signalling, such as GSNO, are removed are also important. The reaction of NO with haemoglobins, for example, may be one such mechanism that is potentially subject to regulation (Perazzolli et al., 2006), as could be the metabolism of GSNO by GSNO reductase (Wang et al., 2006). Another question to be resolved relates to the mechanism(s) by which NO is perceived in cells. Transient elevation of the second messenger cGMP is likely, as is the direct regulation of the activity of enzymes, transcription factors, and ion channels by the reversible S-nitrosylation of cysteine residues located therein. The functions of many other components of the ABA–H2O2–NO–cellular response signal transduction chain also require clarification. These include the protein kinases OST1 (Mustilli et al., 2002) and OXI1 (Rentel et al., 2004) and many other as yet unknown proteins and signalling molecules. There is much to do and much to discover in this area of plant biology.
Work in the authors’ laboratory was supported by the BBSRC (UK), the Leverhulme Trust (UK), and the CNPq (Brazil).