NATURE OF THE SYSTEMIC SIGNAL
Early grafting experiments demonstrated that the infected leaf produces a systemic signal for SAR, and this signal is not species specific. The nature of the systemic signal has been a subject of controversy for many years.
Salicylic Acid
The detection of increased SA levels in systemic leaves and in the phloem led many researchers to believe that SA might be a systemic signal for SAR. The evidence for and against this hypothesis has been the subject of previous reviews. Labeling studies in TMV-infected tobacco showed that most of the SA (69%) accumulating systemically was made and exported from the inoculated leaf. Similarly, in cucumber infected with TNV, SA found in systemic leaves was both imported from the infected leaf and synthesized de novo. A more recent study suggests that signaling might occur through the conversion of SA to the volatile compound methyl salicylate, which could induce resistance not only in the uninfected parts of the same plant but also in neighboring plants.
A number of experiments argue against SA being the systemic signal. Detachment of Pseudomonas syringae-infected cucumber leaves before SA levels had increased in the petiole did not block the development of SAR. Furthermore, grafting experiments in tobacco between wild-type scions and nahG-expressing rootstocks showed that, although the rootstock was unable to accumulate SA, the SAR signal was still produced and translocated to the scion. The reciprocal grafting experiment showed that the systemic tissue must accumulate SA for the SAR signal to be perceived. p lants react to pathogen attack by activating an elaborate defense mechanism that acts both locally and systemically. In many cases, local resistance is manifested as a hypersensitive response, which is characterized by the development of lesions that restrict pathogen growth and/or spread 1 (Fig. 1). Associated with the hypersensitive response is the induction of a diverse group of defenserelated genes. The products of many of these genes play important roles in containing pathogen growth, either indirectly, by helping to reinforce the defense capabilities of host cell walls, or directly, by providing antimicrobial enzymes and secondary metabolites (Fig. 2). These products include cell wall polymers, such as lignin and suberin, as well as phenylpropanoids and phytoalexins. Several faroilies of pathogenesis-related (PR) proteins are also induced during the hypersensitive response. Some of these proteins are hydrolytic enzymes [e.g. ~-l,3-glucanases (PR-2) and chitinases (PR-3)], but the functions of other PR proteins reviews have yet to be determined. Most of the PR proteins have been shown to possess antimicrobial activity in vitro or the ability to enhance disease resistance when overexpressed in transgenic plants 2'3. Additionally, the hypersensitive response is associated with a massive increase in the generation of reactive oxygen species (the oxidative burst), which precedes and then accompanies lesion-associated host cell death. Over a period of hours to days after the primary infection, systemic acquired resistance develops throughout the plant. The systemic acquired resistance is manifested as an enhanced and long-lasting resistance to secondary challenge by the same or even unrelated pathogens. The application of molecular, genetic and biochemical techniques has led to the identification of key components of the signaling pathways leading to defense responses. Here, we focus on recent advances, and discuss the central role of salicylic acid in resistance to pathogens. Salicylic acid and disease resistanc .The signaling pathways involved in the initiation and maintenance of the hypersensitive response and systemic acquired resistance are still poorly understood. Only recently has salicylic acid emerged as a key signaling component involved in the activation of certain plant defense responses. For several years, it was known that the application of salicylic acid or aspirin to tobacco induced PR gene expression, and enhanced resistance to pathogens such as tobacco mosaic virus (TMV). However, in the early 1990s, it became apparent that salicylic acid is an endogenous compound that operates in the signaling pathway for plant defense. After TMV infection, salicylic acid accumulates to high levels at the site of infection, with a subsequent, but much smaller rise, in the uninfected systemic tissues. In tobacco, this increase paralleled the transcriptional activation of PR genes in both the inoculated and un-inoculated leaves. Strikingly, exogenously supplied salicylic acid induced the same set of nine genes that are activated systemically upon TMV infection. An increase in salicylic acid levels in the phloem of cucumber plants infected with either tobacco necrosis virus or Colletotrichum lagenarium was also shown to precede the development of systemic acquired resistance 2,~. More recently, the participation of salicylic acid in plant defense responses has been demonstrated through analysis of transgenic tobacco and Arabidopsis expressing the nahG gene, which encodes the enzyme salicylate hydroxylase from Pseudomonas putida 4'~. These plants accumulate little, if any, salicylic acid, and as a consequence show reduced or no PR gene expression, fail to establish systemic acquired resistance, and are compromised in their ability to prevent pathogen growth and spread from the primary infection site. The importance of salicylic acid in the activation of resistance was further underscored by the demonstration that Arabidopsis plants become susceptible to avirulent fungal pathogens when phenylalanine-ammonia lyase (PAL) activity is specifically inhibited ~. Since PAL catalyzes the first step in salicylic acid biosynthesis, and resistance can be restored in PAL-inhibited plants by treatment with exogenous salicylic acid, increased susceptibility is presumably caused by a block in salicylic acid synthesis. Although salicylic acid and salicylic acid signal transduction pathways are involved in resistance to many pathogens, in some cases PR gene expression and resistance can be activated in a salicylic acid-independent manner. For Fig. 1. The hypersensitive response and systemic acquired resistance. Tobacco cultivars that carry a dominant resistance gene [e.g. N ('Nicotiana')] are able to restrict the spread of tobacco mosaic virus to a small zone of tissue around the point of entry, where a necrotic lesion later appears (right). This resistance phenotype, the hypersensitive response, is subsequently accompanied by the induction throughout the plant of systemic acquired resistance. Consequently, a secondary infection with the virus, occurring several days after the initial infection, results in much smaller lesions (left) as compared with those induced by the primary infection. The leaves are shown 4 d after infection. The activation of systemic induced resistance in Arabidopsis by root inoculation with the biocontrol bacterium P. fluorescens is not associated with increases in endogenous salicylic acid or PR gene expression 7. Additionally, the systemic resistance induced by P. fluorescens is manifested equally well in transgenic Arabidopsis expressing nahG. Similarly, the presence of the nahG gene does not compromise either Cf-2 or Cf-9 gene-mediated resistance to Cladosporium fulvum in tomato s. Is salicylic acid the mobile signal? It has been known for some time that the signal for establishing systemic acquired resistance is transported from the pathogen-inoculated leaf to uninoculated leaves via the phloem. It was suggested that salicylic acid might be the systemic signal for systemic acquired resistance following reports of salicylic acid accumulation occurring in parallel to or even preceding PR gene activation, and the development of systemic acquired resistance in uninfected leaves of TMVinoculated tobacco, combined with the detection of salicylic acid in the phloem of pathogen-infected tobacco or cucumber 2'3's. However, while these experiments clearly demonstrated a correlation between salicylic acid and systemic acquired resistance, they do not prove that salicylic acid is the long-distance mobile signal. Currently, strong evidence that salicylic acid may be this long-distance signal has come from an elegant experiment in which the translocation of labeled salicylic acid was monitored in TMV-infected tobacco 9. This analysis made use of the fact that the final step in salicylic acid biosynthesis in tobacco is the O2-dependent hydroxylation of benzoic acid, catalyzed by benzoic acid 2-hydroxylase. It was therefore possible to label the salicylic acid synthesized in TMV-inoculated lower leaves by enclosing July 1997, Vol. 2, No. 7 267 reviews Systemic;, f'" 1 /.."acquired i ~t j/ resistance~ i ................... Hypersensitive i ~__-~ Me~yl ~ / j , .................. i , } response i ~ ~ ~ sal!~ylate Salicylic i ~lnfection---"-0~ ",. ,~ / ~t ac,d / Methl , '. / l i " / "~ Salicylic ...... [~Y. "~.. -- Sa eylic acid \L acid sa,icyl~m ~/ ...................................... ~]glucoside I \. t',. / Salioyli~ acid \. 'k.~, ~ , B_glucos'l-de ~ ~-~ _~...,,J I i Fig. 2. Defense responses to pathogen infection. The infection of resistant plants by pathogens generally results in the hypersensitive response - the formation of necrotic lesions and restricted pathogen growth and spread. A variety of defense responses is induced locally around the sites of infection. An oxidative burst precedes the formation of necrotic lesions. Additional defense responses in surrounding cells include the induction of genes for pathogenesisrelated (PR) proteins, peroxidases and enzymes involved in cell wall strengthening and the biosynthesis of phytoalexins. Some of these genes are also activated systemically, and are believed to play a role in the development of systemic acquired resistance. The synthesis and accumulation of salicylic acid appear to be necessary for the activation of several of these defense responses, both locally and systemically. A substantial amount of salicylic acid is converted to salicylic acid ~-glucoside, a probable storage form 2''~. It is still unclear whether salicylic acid is a long-distance mobile signal in systemic acquired resistance. Most recently, methyl salicylate, which is synthesized from and metabolized to salicylic acid, was shown to act as an airborne signal that activates defense mechanisms in distal leaves and possibly even neighboring plants 11. However, at room temperature, methyl salicylate is a liquid and could be translocated through the vascular system of the plant, just as salicylic acid. them in an lSQ-rich environment. Subsequent analysis of the upper uninoculated leaves indicated that almost 70% of the salicylic acid was 1SO-labeled and had therefore been synthesized in and transported from the TMV-inoculated leaf. The biosynthesis and transport of salicylic acid have been studied by administering 14C benzoic acid to cucumber cotyledons infected with C. lagenarium ~°. In these experiments, 14C-labeled salicylic acid was detected in upper uninoculated leaves before the development of systemic acquired resistance. Recently, it was shown that methyl salicylate, produced from salicylic acid upon TMV infection of tobacco, may function as an airborne signal '1. Alternatively, it may be translocated through the vascular system. After the conversion of methyl salicylate back to salicylic acid, it activates defense responses in uninfected tissues and possibly even neighboring plants (Fig. 2). Despite the strong correlations from these studies on salicylate biosynthesis and transport, they do not rule out the possibility that salicylates are simply translocated in parallel with an unknown signal molecule. This possibility is supported by the observation that the signal for the development of systemic acquired resistance moved out of P. syringae-infected cucumber leaves before any increase in salicylic acid level could readily be detected in the phloem sap 12. Grafting experiments between NahG and wild-type tobacco have also suggested that salicylic acid is not the long-distance signal s. When a NahG rootstock (which is unable to accumulate salicylic acid) was inoculated with TMV, the uninoculated leaves of the wild-type scion still showed systemic acquired resistance. However, these results need to be interpreted with caution. Although it has been shown that catechol, produced from salicylic acid via salicylate hydroxylase, cannot substitute for salicylic acid 13, it is unclear whether the residual salicylic acid in NahG plants is able to act as a long-distance messenger. Studies of transgenic tobacco expressing the cholera toxin gene (a known modulator of signaling pathways dependent on heterotrimeric G proteins) also suggest that salicylic acid is not the translocated systemic acquired resistance signal '4. These plants constitutively accumulate high levels of salicylic acid, express PR genes, show enhanced resistance and develop spontaneous lesions. However, systemic acquired resistance was not observed when a wild-type scion was grafted onto a transgenic rootstock, even though the rootstock accumulated high levels of salicylic acid. Thus, further work, such as the isolation of salicylic acid biosynthesis genes and the identification of mutations targeting the pathways for salicylic acid metabolism and transport, are required to clarify whether salicylic acid functions as a long-distance signal. Mechanisms of action Whether or not salicylic acid emerges as the mobile systemic acquired resistance signal, it does appear to be required for establishing and maintaining systemic acquired resistance. This conclusion is based on results from the NahG grafting experiments already described ~. Systemic acquired resistance did not develop in an NahG scion after infection of the wild-type rootstock with TMV. However, the mechanism by which salicylic acid induces systemic acquired resistance is still unclear. Previous studies have demonstrated that salicylic acid binds and inhibits tobacco catalase activity both in vitro and in vivo 1~'16. Thus, one possible function of salicylic acid is to inhibit the hydrogen peroxide (H202)-degrading activity of catalase, thereby leading to an increase in the endogenous level of H2Q, which is generated by photorespiration, photosynthesis, oxidative phosphorylation and the hypersensitive responseassociated oxidative burst. The H2Q , and other reactive oxygen species derived from it, could then serve as second messengers to activate the expression of plant defenserelated genes, such as PR-1. This hypothesis is currently the subject of intense debate. Reactive oxygen species and plant defense In plants, H202, superoxide radicals (02" ) and hydroxyl radicals (OH') are thought to play key roles in defense responses. Following infection, plants resistant to the invading pathogen develop a sustained increase in reactive oxygen species. In a manner analogous to their participation in macrophage or neutrophil action, these reactive .oxygen species might be involved in directly killing invading pathogens. In addition, increases in H202 have been shown to induce the crosslinking of cell wall proteins ~7 and to enhance the peroxidase-catalyzed synthesis of lignin, thereby creating a physical barrier against pathogens ~. Reactive oxygen species can also serve as second messengers for the activation of defense gene expression. For example, elevated reactive oxygen species levels induce the genes for glutathione-S-transferase, glutathione peroxidase and polyubiquitin, as well as peroxidases, catalases and other enzymes involved in scavenging reactive oxygen species, chilling tolerance and pathogen resistance. Currently, the mechanism(s) by which redox signaling activates these genes is a matter of debate. The ability of reactire oxygen species, and thus the cellular redox state, to activate plant defenses may parallel the mechanism by which oxidative stress induces the genes associated with animal immune and inflammatory responses. Activities of at least two transcription factors, NF-KB and AP-1, have been shown to be regulated by the cellular redox state; whether these proteins are activated directly by H~O2, or indirectly by thiol metabolites, such as glutathione, is unclear. To date, the only genes shown to be regulated directly by reactive oxygen species are those in the bacterial oxyR and sox regulons ~s. Another line of early plant defense that may be triggered by reactive oxygen species is cell death. Treatment of soybean suspension cells with high concentrations of H20 2 (6-10 raM) was shown to cause cell death, which could be enhanced by the addition of salicylic acid or the catalase inhibitor 3-aminotriazole ~9. In contrast, another study has suggested that O z" , but not H202, is crucial for the induction of cell death. To determine the mechanism by which spontaneous lesions develop on the leaves of the Arabidopsis lsdl ('lesion simulating disease') mutant in the absence of pathogen infection, these plants were treated with O2"-or H20 2 generating or scavenging systems 2°. Strikingly, elevated levels of O2"-, but not H20 2, were able to induce lesion formation. Despite repeated suggestions that reactive oxygen species are involved in the signaling pathways leading to apoptosis and/or programmed cell death in animals ~s, there is still no conclusive evidence that they are required for actually killing the cells. Indeed, it has recently been suggested that reactive oxygen species may be associated with, but not directly responsible for, apoptosis in animal cells 2~. 1-1202 and salicylic acid: which is the source and which is the signal? It has been hypothesized that salicylic acid binds to catalase, inhibits its activity and thereby increases the intracellular concentration of H2Q, which might then serve as a second messenger for the induction of a defense response 1~. In contrast, recent reports have suggested that PR gene induction during the hypersensitive response and systemic acquired resistance may not be activated by salicylic acidmediated increases in H=O 2. At the site of infection, salicylic acid levels can reach 150 llM, a concentration sufficient to cause substantial inhibition of catalase and ascorbate peroxidase, the other major H2Q-scavenging enzyme 4'1~'~6'22'~3. However, no decrease in catalase activity could be detected in pathogen-inoculated leaves ~3'24. In contrast, the concentration of salicylic acid in uninfected systemic tissue is probably too low to increase H2Q levels through the inhibition of catalase or ascorbate peroxidase, unless salicylic acid is concentrated in a subcellular compart/nent. Recent studies using transgenic tobacco plants have also suggested that the salicylic acid-mediated inhibition of catalase and increased level of H20 2 are not involved in the activation of defense responses. When catalase expression was suppressed in leaves of transgenic plants through sense cosuppression or antisense suppression, most plants failed to show constitutive PR gene expression 25'26. Additionally, H20 2 and H2Q-inducing chemicals were unable to induce PR expression in NahG plants, although they could activate PR-1 genes in wild-type tobacco 13'27. Based on these results, salicylic acid appears to act downstream of H20 2, rather than the reverse. Furthermore, it was recently demonstrated that high levels of H20~, as well as ozone or ultraviolet treatment, stimulate salicylic acid biosynthesis 24'27'2s. Thus, H20 2 might play a role in the activation of PR genes by increasing salicylic acid levels. It is believed that all organisms utilize signaling cascades to transduce oxidative stress, and that they frequently respond to stress by strengthening their anti-oxidative systems. Whether salicylic acid plays a role in either process in plants is unknown. It has been proposed that alterations in the cellular redox state, as a result of changes in the glutathione or plastoquinone pools or the levels of metabolites such as nicotinamide or salicylic acid, serve as redox messengers in plants 15,1s'2~. On the other hand, phenol-based anti-inflammatory drugs (including salicylic acid and aspirin) are thought to act, at least in part, as anti-oxidative compounds and direct scavengers of reactive oxygen species 3°. Thus, it is possible that salicylic acid functions, in part, by acting as an anti-oxidant. In this capacity, it could help contain the oxidative damage associated with lesion formation and/or spread during the hypersensitive response. This would closely resemble the anti-oxidative role of salicylic acid in inflamed mammalian tissues. It is noteworthy that, in NahG plants, lesions that develop upon TMV infection are larger than in wild-type plants ~. However, an anti-oxidative role of salicylic acid in the living plant remains to be proven. If the predominant mechanism by which salicylic acid induces defense responses is not through increased H20 2 levels caused by the inhibition of catalase and ascorbate peroxidase, how is the salicylic acid signal perceived and transmitted? One possibility is through the generation of salicylic acid radicals, a likely by-product of the interaction of salicylic acid with catalase and peroxidases 31 (see Box 1). Free radicals derived from phenolic compounds can induce lipid peroxidation, and the products of this reaction, such as lipid peroxides, are potent signaling molecules in animals and possibly also in plants. For example, it is known that salicylic acid induces lipid peroxidation in tobacco suspension cells, and that lipid peroxides activate PR-1 genes in these cells 32. Salicylic acid may also interact with other effector proteins, besides those involved in redox regulation. Recently, a soluble, high-affinity, salicylic acid-binding protein (SABP2) was identified, which reversibly binds biologically active, but not inactive, analogs of salicylic acid in vitro 3~. Additionally, it has a 15-fold higher affinity for the plant protecting agent benzothiadiazole (BTH), which is consistent with the greater efficacy of BTH in inducing plant defense responses ~3'34 (Box 2). However, whether SABP2 is a true receptor or another member of the class of salicylic acid-binding metalloproteins is currently unclear. Moreover, it is likely that future analyses will identify other July 1997, Vol. 2, No. 7 269 reviews Box 1. Proteins that interact with salicylic acid The discovery that salicylic acid (SA) inhibits tobacco catalase(s ~ stimulated discussion concerning both the biological significance of this interaction and the mode of inhibition. The complexity of the redox chemistry of the enzyme allows for many possibilities. H20 + 02 H202 (a) Ferric enzyme . "\"-- J = Compoundl / H202 H20 (b) Ferric enzyme ~-- ~ = Compound I ~\ H202 H20 // SA" \ ~ SA (+H20)~"~ / \, / Compound I1 The issue as to how salicylic acid could inhibit catalase Was recently clarified by the demonstration that salicylic acid acts as a one electron-donating substrate that siphons catalase from the extremely rapid catalytic cycle (a) into a much slower peroxidative cycle (b), which is a secondary acti,dty of catalase ~1 (compound II is another enzyme intermediate of different oxidation state from ferric enzyme and compound I). A consequence of the interaction of salicylic acid with catalase and peroxidases is the formation of a salicylic acid radical (SA'). The ability of salicylic acid to serve as an electron donor for heme proteins is not restricted to catalase. Many protein~ with peroxidase function have been shown to interact with salicylic acid (although this does not necessarily imply inhibition), Some of the potential targets of salicylic acid in mammalian cells include: ..... • Prostagtandin H synthetase. • Lactoperoxidase*. • Myeloperoxidase% • Catalase* • Aconitase~ • Methemoglobin*, • Metmyoglobin*. • Potential targets in plants include: • Catalase*. • Aconitase. • Leghemoglobin*. (a) The catalytic cycle of catalase, in which hydrogen peroxide • Aminocyclopropane carboxylic acid (ACC) oxidase. (H~Q) is converted to H20 and 02 Ferric enzyme and compound I are enzyme intermediates of different oxidation states. Among An asterisk denotes proteins for which salicylic acid has been the suggestions as to how salicylic acid could inhibit catalase shown or suggested to act as an electron donor; all of these are have been chelation of the home iron and a novel atlosteric home proteins. In other cases, salicylic acid might chelate the binding site at the surface of catalase s1"5~, iron of an FeS protein or simply block a substrate binding site. salicylic acid effector proteins that might play roles in disease resistance and/or other salicylic acid-mediated responses (e.g. thermogenesis). Salicylic acid and gene expression. In plants, salicylic acid has been shown to induce the expression of many defense-related genes, as well as to potentiate the production of H202, the induction of cell death and the activation of several genes induced by fungal elicitors and wounding 19'35'~6. The genes induced by salicylic acid can be grouped into two broad classes. The first class consists of genes whose expression is insensitive to protein synthesis inhibitors, such as the glutathione-S-transferase genes, the 35S promoter of cauliflower mosaic virus and the nopaline and octopine synthase genes of Agrobacterium. Promoters of this class of genes contain copies of as-l-like cis elements, which mediate salicylic acid-induced expression. Several transcription factors belonging to the TGA family of bZIP proteins have been identified and shown to bind these elements 37. It was also recently demonstrated that salicylic acid or cycloheximide treatment of tobacco leaves increases an as-1 binding activity, and that phosphatase treatment of nuclear extracts decreases it 38. From these results, it was proposed that the as-1 binding activity is sequestered by an inhibitory protein that is released after salicylic acid treatment, probably via a phosphorylation event(s). This in turn leads to activation of promoters containing the as-l-like element. A MAP kinase that can be activated by salicylic acid and TMV has been identified and purified from tobacco extracts 39, but it is not known whether this kinase is involved in the salicylic acid-mediated activation of this DNA-binding protein. The second class of salicylic acid-inducible genes includes the acidic (class II) PR genes, whose induction by salicylic acid is sensitive to inhibitors of protein synthesis. Promoters of the tobacco PR-la and PR-2 genes have been studied by several groups. However, no common cis elements involved in the salicylic acid-inducible expression of these genes have yet been defined. A 10 bp TCA element that is common to the promoters of several tobacco PR genes, as well as several stressinduced genes, was shown to bind a 40 kDa nuclear protein in a salicylic acid-dependent manner 4°. However, this TCA element was neither sufficient nor required for salicylic acidmediated induction of the tobacco PR-2d promoter in vivo 41. In contrast, in vivo analysis of the PR-2d promoter has identified a 25 bp element that is involved in salicylic acidinducible expression. This element contains the sequence TTCGACC, which is related to the W-boxes present in the promoters of several elicitor- and wound-induced genes 42. Induction of some of these genes by pathogens, elicitors or 270 July 1997, Vol. 2, No. 7 reviews Significant progress has been made in the development of transgenic plants with enhanced resistance to microbial attack. For example, overexpression of genes encoding antifungal enzymes such as chitinase and ~-l,3-glucanase has shown considerable promise. Their introduction and successful use in field crops is anticipated in the near future. Significant advances in the understanding of signal trm~sduction pathways that mediate disease resistance could lead to the next generation oftransgenic plants in which manipulation of key signaling components results in the activation of a broad array of host defenses. Alternatively, the signaling pathway might be altered so that it is primed more rapiclly and effectively to activate these defense arsenals upon infection. Another approach to enhance resistance is through treatmerit with compounds that activate part or all of the host defense arsenals. In addition to salicylic acid and aspirin, two such 'plant defense activators' have been identified and characterized: 2,6-diehloroisonicotinic acid and benzothiadiazole. Both appear to be functional analogs of salicylic acid, and the latter is being used commercially as a plant protecting agent 34. Further research on salicylic acid and its cellular targets should facilitate the development of compounds that mimic endogenous messengers and thereby induce disease resistance. 0 0 H ~ C OH C OH 0 ....OH /O C OH 3 Salicylic acid Asoirin O O H C--OH C S -- OH 3 2.6-Dichloroisonicotinic Benzothiadiazole acid wounding is potentiated by pretreatment with salicylic acid ~9'35'3~. It may be that related factors are involved in the expression of the elicitor-induced defense genes as well as the salicylic acid-induced PR-2d gene. The tobacco PR-la gene promoter contains several binding sites for Myb proteins, redox-regnlated transcription factors found in plants and animals. Some of these sites can be bound by recombinant tobacco Mybl protein in vitro 43. Binding sites for Myb proteins are also present in the promoters of PAL genes, whose expression is potentiated by salicylic acid 35'37. Expression of the tobacco Mybl gene is rapidly induced by salicylic acid, with kinetics similar to those of genes containing the as-1 element. Thus, it is possible that Mybl binding activity is involved in transducing the salicylic acid signal to the PAL and PR promoters. However, Mybl by itself may not be sufficient for salicylic acid inducibility of the PR-la gene, because in vivo analysis of this promoter has suggested that more than one region is involved in the salicylic acid-mediated activation. GT-l-like proteins have also been shown to bind various fragments of the tobacco PR-la promoter in vitro 44. Their binding activity is reduced in extracts from salicylic acid-treated or TMVinfected leaf tissue. Although it is tempting to speculate that the Mybl and GT-l-like proteins might be involved in the salicylic acid-dependent expression of the PR-la gene, there is still no evidence in vivo. Genetic approaches for understanding the role of salicylic acid in defense responses In addition to environmental stress, such as exposure to ultraviolet light and ozone, the inappropriate expression or repression of endogenous or foreign genes in plants can lead to the constitutive expression of defense genes, the activation of systemic acquired resistance and, in several cases, the spontaneous development of lesions like those of the hypersensitive response (Table 1). In most of these cases, the constitutive systemic acquired resistance and spontaneous lesion phenotypes are associated with elevated levels of endogenous salicylic acid. However, none of these transgenes are anticipated to participate directly in salicylic acid biosynthesis. Rather, their expression may induce metabolic stress, which in turn elevates salicylic acid levels, resulting in constitutive systemic acquired resistance. Alternatively, it has been suggested that in some cases expression of these transgenes mimics part of the defense signaling pathway, which then activates salicylic acid biosynthesis s'45. Several Arabidopsis mutants have been identified that constitutively exhibit systemic acquired resistance and contain constitutively high levels of salicylic acid (Fig. 3; Table 2). Interestingly, the Isd, cepl ('constitutive expression of PR genes') and acd2 ('accelerated cell death') mutants also spontaneously develop hypersensitive response-like lesions, while the cprl ('constitutive expresser of PR genes') and cim3 ('constitutive immunity') mutants do not 2'32'4~. Because the constitutive systemic acquired resistance phenotype of these mutants is suppressed by the presence of the nahG gene