The tomato mutant spr1 is defective in systemin perception and the production of a systemic wound signal for defense gene expression

Authors


* For correspondence (fax +517 353 9168; e-mail howeg@msu.edu).

Summary

Wound-induced systemic expression of defensive proteinase inhibitor (PI) genes in tomato plants requires the action of systemin and its precursor protein prosystemin. Although it is well established that systemin induces PI expression through the octadecanoid pathway for jasmonic acid (JA) biosynthesis, relatively little is known about how systemin and JA interact to promote long-distance signaling between damaged and undamaged leaves. Here, this question was addressed by characterizing a systemin-insensitive mutant (spr1) that was previously identified as a suppressor of prosystemin-mediated responses. In contrast to JA biosynthetic or JA signaling mutants that lack both local and systemic PI expression in response to wounding, spr1 plants were deficient mainly in the systemic response. Consistent with this phenotype, spr1 plants exhibited normal PI induction in response to oligosaccharide signals that are thought to play a role in the local wound response. Moreover, spr1 abolished JA accumulation in response to exogenous systemin, and reduced JA accumulation in wounded leaves to approximately 57% of wild-type (WT) levels. Analysis of reciprocal grafts between spr1 and WT plants showed that spr1 impedes systemic PI expression by blocking the production of the long-distance wound signal in damaged leaves, rather than inhibiting the recognition of that signal in systemic undamaged leaves. These experiments suggest that Spr1 is involved in a signaling step that couples systemin perception to activation of the octadecanoid pathway, and that systemin acts at or near the site of wounding (i.e. in rootstock tissues) to increase JA synthesis to a level that is required for the systemic response. It was also demonstrated that spr1 plants are not affected in the local or systemic expression of a subset of rapidly induced wound-response genes, indicating the existence of a systemin-independent pathway for wound signaling.

Introduction

Many plants respond to insect attack and wounding by modulating the expression of genes involved in various defense-related processes. The synthesis and deployment of wound-induced phytochemicals is regulated by signal transduction pathways that operate both locally at the site of wounding and systemically in undamaged leaves throughout the plant (Green and Ryan, 1972; Karban and Baldwin, 1997). Wound-inducible defensive proteinase inhibitors (PIs) in Solanaceous plant species provide an attractive model system in which to study the mechanism of long-distance wound signaling, and several ideas have been proposed regarding the identity of the systemic signal transmitted from wound sites (reviewed by Bowles, 1998; León et al., 2001; Malone, 1996; Ryan, 2000; Walling, 2000). Among the proposed intercellular signals for wound-induced PI gene expression are systemin, an 18-amino acid peptide that is produced from cleavage of a larger precursor protein called prosystemin, and the octadecanoid pathway-derived hormone jasmonic acid (JA) (Farmer and Ryan, 1992; Li et al., 2002a; McGurl et al., 1992; Pearce et al., 1991). A wealth of biochemical and genetic evidence indicates that systemin and JA work together in the same signaling pathway to activate expression of PI and other defense-related genes (Li et al., 2001; Ryan, 2000).

The systemin/JA signaling pathway is activated upon binding of systemin to a 160-kDa plasma membrane-bound receptor called SR160 (Meindl et al., 1998; Scheer and Ryan, 1999). This receptor was recently identified as a member of the leucine-rich repeat (LRR) receptor kinase family (Scheer and Ryan, 2002). Binding of systemin to the cell surface is associated with several rapid signaling events including increased cytosolic Ca2+ levels, membrane depolarization, inhibition of a plasma membrane proton ATPase, and activation of a MAP kinase activity (Felix and Boller, 1995; Moyen and Johannes, 1996; Moyen et al., 1998; Schaller and Oecking, 1999; Stratmann and Ryan, 1997). The systemin signaling pathway leading to PI expression is thought to culminate in activation of a phospholipase that releases linolenic acid, the metabolic precursor of JA, from membrane lipids (Farmer and Ryan, 1992; Narváez-Vásquez et al., 1999). Chitosan oligomers and oligogalacturonides (OGAs) derived from fungal and plant cell walls, respectively, also activate PI expression via the octadecanoid pathway (Bishop et al., 1981; Doares et al., 1995; Walker-Simmons and Ryan, 1984). The presence of wound-inducible polygalacturonase activity in tomato leaves (Bergey et al., 1999), together with the relative immobility of OGAs in the plant vascular system (Aldington and Fry, 1996), suggests that these compounds induce PI expression at or near the site of wounding. JA synthesized in response to wounding, systemin, and OGAs acts in concert with ethylene (O'Donnell et al., 1996) and hydrogen peroxide (Orozco-Cárdenas et al., 2001) to coordinate the induction of PI gene expression. Recent studies in Arabidopsis indicate that JA signaling depends upon assembly of ubiquitin–ligase complexes that presumably target transcriptional repressors of JA-responsive genes for proteolytic degradation (Xie et al., 1998; Xu et al., 2002).

Mutants affected in the synthesis or perception of prosystemin and JA provide useful tools to understand the mechanism of systemic wound signaling. For example, antisense-mediated depletion of prosystemin expression in tomato plants abrogated wound-induced systemic accumulation of PIs, indicating that this gene is essential for a normal systemic wound response (McGurl et al., 1992). To identify genes involved in the systemin/JA signaling pathway using a forward genetic approach, we took advantage of a transgenic plant (called 35S::prosys) that overexpresses prosystemin from the CaMV 35S promoter and, as a consequence, constitutively expresses PI and other defense-related genes in the absence of wounding (Constabel et al., 1995; McGurl et al., 1994). An ethyl methane sulfonate (EMS)-mutagenized population derived from this line was screened for mutants that are impaired in 35S::prosys-mediated signaling (Howe and Ryan, 1999). Among several ‘spr’ suppressed in prosystemin-mediated responses mutations identified, five independent alleles were shown to define one locus called Spr1. Evidence presented herein indicates that Spr1 is involved in a signaling step that couples systemin perception to activation of the octadecanoid pathway. The results of grafting experiments further indicate that Spr1 function is required at or near the site of wounding to amplify JA accumulation to a level sufficient to promote long-distance signaling. The existence of a wound response pathway that operates independently of Spr1 is also described. These results are discussed in the context of the role of systemin in the wound response of tomato plants.

Results

spr1 preferentially affects wound-induced systemic PI expression

Recessive mutations in Spr1 were previously shown to suppress 35S::prosys-mediated expression of the well-characterized serine PI genes, PI-I and PI-II (Howe and Ryan, 1999). Further characterization of this mutant was conducted using spr1/spr1 homozygous lines in which the 35S::prosys transgene was removed by outcrossing (see Experimental procedures). To determine the effect of spr1 on wound-induced local and systemic gene expression, RNA blot analysis was used to measure PI-I and PI-II transcript levels in leaf tissue located at defined distances from a single wound inflicted at the distal end of the lower leaf (Figure 1a). As previously observed in wild-type (WT) tomato plants (Howe et al., 1996), PI mRNA accumulation in the undamaged section (section 2) of the wounded leaf was significantly greater than that in the damaged section (section 1) of the same leaf (Figure 1b). A relatively strong systemic response was observed in the undamaged leaf (sections 5 and 6), whereas WT petioles (sections 3 and 4) showed little or no PI expression. In the case of spr1, PI mRNA accumulation in the damaged section of the wounded leaf was comparable to that in WT (Figure 1c). Mutant plants also showed PI expression in adjacent unwounded tissue (section 2), albeit at a level lower than in WT. More significantly, however, the steady-state level of PI mRNA in the unwounded leaf (sections 5 and 6) of wounded spr1 plants was <10% of that observed in WT. This result suggests that spr1 impairs a signaling pathway that mediates or amplifies wound-induced systemic PI expression, but contributes less to PI expression near the wound site. This interpretation was supported by measurements of PI-II protein levels in damaged (local response) and undamaged (systemic response) leaves of wounded plants (data not shown). In six independent experiments involving at least six plants per genotype, the local response of spr1 plants ranged between 50 and 75% of the WT response. Systemic PI-II accumulation in spr1 plants ranged between 0 and 35% of that in WT plants, with the average response in the mutant being approximately 15% of WT levels. Analysis of plants homozygous for an independent allele of spr1 (spr1-2) gave very similar results; wound-induced local and systemic PI-II accumulation in spr1-2 plants was 51 and 13%, respectively, of WT levels.

Figure 1.

Spatial pattern of PI-I and PI-II mRNA accumulation relative to the wound site.

A single wound was inflicted at the distal end of the terminal leaflet of the lower leaf of two-leaf-stage plants. Eight hours thereafter, various sections of the leaf blade and petiole were dissected for RNA isolation and analysis. Leaf sections from six plants were pooled for RNA isolation. (a) Schematic drawing illustrating the leaf and petiole sections that were harvested for RNA extraction. Panels in (b) and (c) show the results obtained for analysis of wild-type (WT) and spr1–1 plants, respectively. Lanes 1 through 6 represent analysis of RNA isolated from the corresponding tissue sections shown in panel in (a). Lanes 1′, 2′, and 3′ represent analysis of RNA isolated from various tissue sections of unwounded plants: 1′, pooled tissue from sections 1 and 6; 2′, pooled tissue from sections 2 and 5; 3′, pooled tissue from sections 3 and 4. RNA gel blots containing 5 µg total RNA were hybridized to probes for PI-I and PI-II. To facilitate the comparison between WT and spr1, RNA blots shown in panels in (b) and (c) were hybridized together in the same containers and exposed to autoradiographic film for the same time. Blots were hybridized to a probe for eIF4A as a loading control. Note, however, that eIF4A mRNA abundance is greater in petiole tissue relative to leaf lamina. As an additional loading control, a picture of an ethidium bromide-stained gel of the total RNA (rRNA) is shown. The results shown are representative from three independent experiments.

RNA blot analysis was used to determine the time course of local and systemic expression of various wound-responsive genes in spr1 plants. Two classes of genes that differ with respect to their timing of wound-induced expression have been described in tomato (Ryan, 2000). Transcripts of so called ‘late’-response genes, including PI-I, PI-II, and cathepsin D inhibitor (CDI), accumulate to maximum levels 8–12 h after wounding of WT plants (Figure 2). Consistent with the results shown in Figure 1, spr1 plants were deficient in the amplitude but not the timing of induction of these genes. Genes whose expression is induced rapidly and transiently in response to wounding comprise a second class of ‘early’-response genes. Included among this group are genes encoding signaling-related proteins such as lipoxygenase (LoxD), allene oxide synthase1 (AOS1), and a putative mitogen-activated protein kinase (WIPK). Interestingly, wound-induced local and systemic expression of these early genes was not affected in spr1 plants (Figure 2). These results indicate that spr1 specifically affects the expression of late-response genes (i.e. PI genes).

Figure 2.

Time course of wound-induced gene expression in wild-type (WT) and spr1 plants.

WT and spr1 plants (15-day-old) were wounded once on the lower leaf with a hemostat. Lower damaged (local response) and upper undamaged (systemic response) leaves were harvested at various times (hours) after wounding for RNA isolation and analysis as described in the legend for Figure 1. For each time point, six plants were harvested and pooled for RNA extraction.

spr1 plants are impaired in systemin-mediated signaling

To gain additional insight into the wound response phenotype of spr1, the capacity of the mutant to respond to various PI-inducing compounds was determined (Figure 3). Consistent with the ability of spr1 to suppress 35S::prosys-mediated PI expression, spr1 plants did not accumulate PI-II in response to exogenous systemin or its bioactive precursor, prosystemin (Dombrowski et al., 1999). However, the mutant was responsive to octadecanoid signaling compounds (linolenic acid and JA) and to the polysaccharide elicitors OGA and chitosan. Because exogenous chitosan, OGA, and systemin activate PI expression via the octadecanoid pathway (Doares et al., 1995), these findings suggested that spr1 affects systemin-mediated signaling at a point upstream of the octadecanoid pathway. To further test this hypothesis, the responsiveness of spr1 plants to a range of concentrations of systemin, OGA, and chitosan was compared to that of WT (Figure 4). Parallel analysis of the JA-deficient defenseless1 (def1) that is impaired in PI expression in response to systemin, OGA, and chitosan (Howe et al., 1996) was included as a control. WT plants showed a strong response to systemin concentrations of 1 pmol per plant and greater, whereas def1 accumulated low levels of PI-II in response to high concentrations of systemin, as previously reported (Howe et al., 1996). spr1 plants failed to accumulate significant levels of PI-II (<5% WT levels) in response to all concentrations of systemin tested. As expected, the spr1 mutant responded normally to a range of concentrations of OGA and chitosan, whereas def1 plants were unresponsive to these elicitors (Figures 4b,c). These results support the idea that spr1 specifically affects the systemin branch of the wound-response pathway.

Figure 3.

PI-II accumulation in wild-type (WT) and spr1 plants in response to exogenous signaling compounds.

WT (filled bar) and spr1 (open bar) seedlings (15-day-old) were excised at the base of the stem and supplied with 15 mm sodium phosphate buffer (Con), systemin (Sys, 15 nm), recombinant prosystemin (PS, 0.1 µg ml−1), chitosan (Chit, 250 µg ml−1), oligogalacturonide (OGA, 250 µg ml−1), linolenic acid (LA, 5 mm), or jasmonic acid (JA, 100 nm). PI-II levels were measured 24 h after treatment. Data points represent the mean and SD (n = 6).

Figure 4.

Response of spr1 plants to exogenous systemin, oligogalacturonides (OGA), and chitosan.

Two-leaf-stage wild-type (WT) (filled bars), def1 (gray bars), and spr1 (open bars) plants were excised at the base of the stem and supplied with either phosphate buffer (‘0’) or buffered solution containing various concentrations of systemin (a), OGA (b), or chitosan (c). PI-II accumulation in leaves was measured 24 h after treatment. Values indicate the mean and SD (n = 6).

RNA blot analysis was used to determine the effect of spr1 on systemin-mediated expression of early and late wound-response genes. Excision of seedlings at the base of the stem resulted in a gradual, low-level increase in PI-II mRNA accumulation in leaves of both WT and spr1 plants (Figure 5a). WT plants accumulated high levels of PI-II mRNA in response to exogenous systemin, whereas spr1 showed no response above that observed in the buffer control. This finding is consistent with the inability of spr1 plants to accumulate PI-II protein in response to systemin (Figure 4). Mock treatment (excision and incubation in buffer) of both WT and spr1 seedlings resulted in rapid and transient expression of three early wound-response genes: LoxD, AOS1, and WIPK. In WT plants, systemin clearly enhanced the accumulation of LoxD and AOS1 mRNA, as previously reported (Heitz et al., 1997; Sivasankar et al., 2000). However, the expression of these genes in spr1 was not enhanced by systemin. In contrast to LoxD and AOS1, the steady-state level of WIPK mRNA in both WT and spr1 plants was not affected by systemin treatment.

Figure 5.

Effect of exogenous systemin on gene expression in wild-type (WT) and spr1 plants.

(a) Excised seedlings (2-week-old) were incubated for 45 min in a solution containing 5 pmol systemin, and then transferred to water. At various times (hours) after the beginning of systemin treatment, leaf tissue was harvested for RNA isolation. Leaf tissue from six plants was pooled for each RNA isolation. RNA gel blots containing 5 µg total RNA were hybridized to probes for PI-II, LoxD, AOS1, WIPK and, as a loading control, eIF4A.

(b) Excised seedlings were pre-incubated in water for 4 h and then transferred either to 300 µl phosphate buffer (buffer) or the same volume phosphate buffer containing 5 pmol of systemin (‘0’). Following uptake of this solution (approximately 45 min), seedlings were transferred to water. At various times (hours) thereafter, leaves were harvested for RNA isolation.

The analysis of systemin-induced gene expression in WT and mutant plants was complicated by the fact that excision of seedlings at the base of the stem induced significant expression of early-response genes (Figure 5a). To determine the effect of systemin in the absence of this excision-induced effect, excised plants were pre-incubated in water for 4 h (to allow mRNA levels to return to basal level), and then transferred to tubes containing either buffer or systemin (Figure 5b). A very low level of LoxD and AOS1 expression was detected in buffer-treated WT and spr1 plants, presumably as a result of handling (i.e. touching) of plants during the transfer procedure. Transfer of pre-incubated WT plants to a systemin-containing solution resulted in a rapid and transient increase in the steady-state level of LoxD and AOS1 mRNAs, and a more gradual, massive accumulation of PI-II transcripts. The level of WIPK mRNA in WT plants was unaffected by systemin treatment, indicating that exogenous systemin stimulates expression of some (e.g. LoxD, AOS1, and PI-II) but not all (e.g. WIPK) wound-response genes. Treatment of pre-incubated spr1 plants with systemin did not increase the accumulation of LoxD, AOS1, PI-II, or WIPK mRNAs above the level observed in buffer-treated plants. In summary, these results show that spr1 impairs systemin-mediated activation of both early- and late-response genes. However, the mutation does not affect the rapid and transient activation of early genes in response to excision or wounding, indicating the existence of an Spr1-independent wound signaling pathway.

Effect of spr1 on wound- and systemin-induced JA accumulation

Because wound- and systemin-induced PI gene expression is dependent upon the synthesis and subsequent action of JA, it was of interest to determine the capacity of spr1 plants to accumulate JA in response to wounding and systemin. JA levels in unwounded leaves of WT and mutant plants were 15.1 ± 1.5 pmol JA g−1 FW and 14.5 ± 4.6 pmol JA g−1 FW, respectively (Figure 6a). In wounded WT plants, JA levels increased 15- and 7.5-fold, 1 and 3 h after wounding, respectively. Wounding also increased JA accumulation in spr1 plants, albeit to levels that were significantly lower (P < 0.05) than WT levels. The amount of JA in wounded spr1 leaves throughout the time course was estimated to be approximately 57% of that in WT leaves. Exogenous systemin induced high levels of JA accumulation in WT, but had no effect in spr1 (Figure 6b). These findings indicate that Spr1 is necessary for maximal levels of JA accumulation in response to wounding, but is strictly required for systemin-induced JA accumulation.

Figure 6.

Jasmonic acid (JA) accumulation in response to systemin and mechanical wounding.

(A) Leaves of 2-week-old wild-type (WT) (filled bar) or spr1 (open bar) plants were mechanically wounded with a hemostat. At various times after wounding (1 or 3 h), wounded leaf tissue was harvested for JA extraction. JA was also extracted from leaves of unwounded plants (‘0’).

(b) Two-week-old seedlings were excised at the base of the stem, pre-incubated in water for 4 h, and then transferred to either buffer unwanted plant (‘0’) or a buffered solution containing systemin (5 pmol per plant). Leaves were harvested for JA extraction 2 h 45 min after systemin application. The amounts of JA in plant extracts were quantified by GC–MS. Data represent the mean and SD of three independent experiments.

spr1 plants are defective in the generation of a systemic wound signal for PI expression

The deficiency of wound-induced systemic PI expression in spr1 plants could result from a defect in production of a long-distance wound signal or a defect in the perception of this signal in distal undamaged leaves. To address this question, wound-induced PI-II expression was analyzed in reciprocal grafts between WT and spr1 plants. Four-week-old plants were grafted such that both the rootstock (stock) and the scion contained at least two healthy leaves. After the graft junction healed, stock leaves were wounded and PI-II mRNA levels were measured 11 h after in both the damaged stock leaves (local response) and the undamaged scion leaves (systemic response). Wounding of WT stock leaves resulted in local and systemic accumulation of PI-II transcripts to levels well above that observed in unwounded control plants that had also been grafted (Figure 7, lanes 1 and 2). This result demonstrates that wounding of WT stock leaves leads to the production of a graft-transmissible signal that is recognized in undamaged scion leaves. Consistent with the pattern of PI-II expression in two-leaf-stage spr1 plants (Figure 1), wounded spr1 stock leaves showed a relatively strong local response and a weak (10% WT) systemic response (Figure 7, lanes 3 and 4). Analysis of spr1/WT hybrid grafts showed that upon wounding of spr1 stock leaves, WT scions failed to activate PI-II expression to levels greater than that in spr1 scions that had been grafted to spr1 stock (Figure 7, lanes 5 and 6). In the reciprocal combination, however, spr1 scions were responsive to a signal emanating from wounded leaves of WT stock (Figure 7, lanes 7 and 8). Taken together, these findings indicate that spr1 impairs wound-induced systemic PI expression mainly by blocking the production of the long-distance wound signal in damaged leaves, rather than the recognition of that signal in systemic, undamaged leaves.

Figure 7.

Wound-inducible PI-II expression in grafts between wild-type (WT) and spr1 plants.

WT and spr1 plants were grafted in the four combinations indicated. The genotypes listed above and below the horizontal line correspond to the scion and rootstock, respectively. For each graft combination, plants were divided into a control (–) and experimental (+) group consisting of four grafted plants per group. For the experimental group, each leaflet on the rootstock was mechanically wounded with a hemostat. Eleven hours after wounding, leaf tissue was harvested separately from wounded rootstock leaves and undamaged scion leaves (scion) for RNA extraction. The control set of plants received no wounding, other than that inflicted by the grafting procedure itself. Levels of PI-II mRNA were analyzed by RNA blot analysis, using an eIF4A cDNA probe as a loading control. The results shown are representative of three independent experiments.

Discussion

spr1 defines a novel class of wound-response mutant

Forward genetic screens have identified two general classes of mutants that are defective in wound-induced systemic PI expression (Howe and Ryan, 1999; Li et al., 2001; Lightner et al., 1993). One group includes JA biosynthetic mutants (e.g. def1 and spr2) that are unresponsive to upstream signals (e.g. systemin, OGA, and chitosan) that activate the octadecanoid pathway, but are responsive to exogenous JA. The second group includes JA-insensitive mutants (e.g. jai1) that are responsive neither to upstream signals nor to JA. Here, we show that spr1 differs from existing wound-response mutants in that it is responsive to OGA and chitosan but unresponsive to systemin and its precursor, prosystemin. Consistent with the fact that OGA and chitosan activate PI expression through JA (Doares et al., 1995), spr1 plants were responsive to exogenous JA and its metabolic precursor, linolenic acid (Figure 3). These findings suggest that the octadecanoid pathway and downstream signaling steps leading to PI expression are intact in spr1 plants. The capacity of the mutant to accumulate significant levels of JA (approximately 57% WT levels) in response to wounding supports this idea. Considered collectively, the most straightforward interpretation of the results is that Spr1 is involved in the perception of systemin or a subsequent systemin-specific signaling event necessary for activation of the octadecanoid pathway. Because systemin perception occurs at the level of the plasma membrane (Meindl et al., 1998; Scheer and Ryan, 1999) and the initial steps of the octadecanoid pathway occur in the chloroplasts, it is possible that Spr1 is involved in relaying a signal from the plasma membrane to the chloroplast. Included among the early signaling events induced by systemin are increased cytosolic Ca2+ levels, membrane depolarization, inhibition of a plasma membrane proton ATPase, activation of a MAP kinase activity, and activation of a phospholipase A2 activity (Felix and Boller, 1995; Moyen and Johannes, 1996; Moyen et al., 1998; Narváez-Vásquez et al., 1999; Schaller and Oecking, 1999; Stratmann and Ryan, 1997). The insensitivity of spr1 plants to both prosystemin and systemin indicates that the mutant is not defective in the synthesis or proteolytic processing of prosystemin. However, the data leave open the possibility that spr1 impairs the interaction of systemin with SR160, or signaling output of the activated receptor complex. Additional work is needed to distinguish these possibilities.

Spr1-independent wound signaling

It is noteworthy that spr1 appears to impair wound-induced systemic PI expression much more than it affects local PI expression. This finding suggests that the signaling pathway for systemic PI expression can be uncoupled from the signaling pathway that operates in tissue adjacent to the wound site. This aspect of spr1 is reminiscent of the wound-response phenotype of prosystemin antisense plants that are compromised in prosystemin production but nevertheless have the capacity to respond to exogenous prosystemin (Dombrowski et al., 1999; McGurl et al., 1992). The robust local wound response of spr1 and prosystemin antisense plants supports the hypothesis that multiple signals generated at the wound site activate the octadecanoid pathway in wounded leaves (Doares et al., 1995; Farmer and Ryan, 1992; Ryan, 2000). This interpretation is consistent with the observation that spr1 plants respond normally to OGAs, and accumulate significant levels of JA in response to wounding (Figures 4 and 6). Whether OGAs are responsible for the entire pool of wound-induced JA in spr1 leaves or whether other mechanisms are involved in initiating the octadecanoid pathway remains to be determined.

Genetic analysis of the wound response in tomato plants indicates that the bulk of wound-induced systemic PI expression requires the systemin/JA signaling pathway. However, we did observe that spr1 plants exhibit a low but significant level of wound-induced systemic PI expression. This residual signaling activity could reflect incomplete loss of Spr1 function or, alternatively, a Spr1-independent pathway for systemic PI expression. Given the complete lack of systemin-induced gene expression in spr1 plants, however, the latter possibility seems more likely. The existence of a systemin-independent wound-response pathway is clearly supported by the observation that spr1 plants are not affected in wound- or cut-induced expression of early wound-response genes such as LoxD and AOS1. The fact that exogenous systemin enhances expression of LoxD and AOS1 in WT plants (Figure 5) suggests that wounding and exogenous systemin may regulate the expression of these genes in somewhat different ways. For instance, it is possible that responses to systemin, when supplied through the transpiration stream, do not accurately reflect the wound-induced activity of systemin produced in vascular bundle cells of intact plants (Jacinto et al., 1997; Ryan, 2000). The expression pattern of WIPK, which was wound and cut inducible in both WT and spr1 plants, provides further evidence for systemin-independent wound signaling. Unlike the LoxD and AOS1 genes, exogenous systemin did not stimulate WIPK mRNA accumulation in WT (or spr1) plants. Several other wound-induced rapid systemic responses have been described in plants (e.g. O'Donnell et al., 1998; Seo et al., 1995; Stratmann and Ryan, 1997), and may involve physical (e.g. hydraulic) signals propagated through the xylem (Malone, 1996) or the phloem (Rhodes et al., 1999).

Role of systemin in wound signaling

The signaling-related phenotypes of spr1 plants are fully consistent with a role for prosystemin in regulating wound-induced systemic PI expression through the octadecanoid pathway (Farmer and Ryan, 1992; Li et al., 2001; McGurl et al., 1992). What is less clear, however, is how systemin and JA (and other signals) interact to effect wound signaling over long distances. Insight into this question was recently provided by means of grafting experiments showing that JA biosynthesis in rootstock leaves is essential for production of a long-distance signal for PI expression, whereas JA biosynthesis in undamaged leaves is not required for PI expression (Li et al., 2002a). Given that prosystemin works through JA, the most straightforward interpretation of these data is that systemin, acting in the rootstock portion of the graft, amplifies the synthesis of JA to levels that are required for long-distance signaling. Support for this model comes from the analysis of systemic PI expression in reciprocal grafts between WT and spr1 (Figure 7). These experiments show that Spr1 function is involved primarily in the production of the long-distance signal, rather than the recognition or processing of that signal in systemic undamaged leaves. The reduced level of wound-induced JA in spr1 plants further suggests that systemin may activate the synthesis of a specific pool of JA that is necessary for the systemic response, which is consistent with the lack of systemin-induced JA accumulation in spr1 (Figure 6). Along these lines, two hypotheses have recently been proposed to explain the role of systemin in long-distance PI expression (Ryan and Moura, 2002). First, systemin may induce localized production of JA that subsequently exits the wounded leaflet and activates PI expression in distal leaves. Alternatively, systemin produced at the wound site may be mobilized in the phloem where it activates JA synthesis in vascular tissues of leaves, petioles, and stems of the rootstock. The latter scenario suggests that a positive feedback loop between systemin and JA may amplify and propagate the systemic signal along the vascular system (Ryan, 2000; Ryan and Moura, 2002). Additional insight into the function of Spr1 and other genes involved in systemic wound signaling may help to distinguish these hypotheses.

Experimental procedures

Plant material and treatments

Tomato (Lycopersicon esculentum Mill cv Castlemart) seedlings were grown in Jiffy peat pots (Hummert International, Earth City, MO) in a growth chamber maintained under 17 h of light (200 µE m−2 sec−1) at 28°C and 7 h of dark at 18°C. Seed for def1 was collected from a def1/def1 homozygous line that was back-crossed four times using Castlemart as the recurrent parent. To simplify and standardize the genetic nomenclature, herein we refer to the previously described spr-1593F and spr-1961E alleles (Howe and Ryan, 1999) as spr1-1 and spr1-2, respectively. Lines homozygous for either of these two alleles were generated as described below. Unless otherwise indicated, spr1-1 was used for all experiments.

Wounding and chemical elicitor experiments were performed using two-leaf-stage plants (14 to 16-day-old) as previously described (Howe et al., 1996). To assay the responsiveness of mutants to PI-inducing compounds, plants were excised at the base of the stem and placed in 0.5 ml microfuge tubes containing 300 µl of the inducing compound. When >75% of the elicitor solution had been imbibed (approximately 45 min), plants were transferred to glass vials containing 20 ml of water, and incubated in a Lucite box for 24 h under continuous light. PI-II levels in leaves were measured by radial immunodiffusion assay (Ryan, 1967). Systemin, oligogalacturonic acid (OGA), and recombinant prosystemin (Dombrowski et al., 1999) were obtained from Dr C.A. Ryan (Washington State University). Chitosan, JA, and linolenic acid were obtained from Sigma. All inducers except linolenic acid were diluted from stock solutions into sodium phosphate buffer (15 mm sodium phosphate, pH 6.5) prior to use. Linolenic acid was diluted into 15 mm sodium phosphate, pH 6.5, containing 0.05% (v/v) ethanol. Control experiments showed that 0.05% ethanol in 15 mm sodium phosphate buffer did not affect the background level of PI expression (data not shown). Grafting experiments were performed as described by Li et al. (2002a).

Genetic analysis

spr1-1 and spr1-2 were originally isolated as recessive suppressors of 35S::prosys-mediated responses (Howe and Ryan, 1999). Segregation of spr1-1 and spr1-2 from the 35S::prosys transgene, for the purpose of isolating homozygous spr1 alleles in an otherwise WT genetic background, was achieved by crossing mutant lines (homozygous for the 35S::prosys transgene and either spr1-1 or spr1-2) to WT (cv Castlemart). Plants in the resulting F2 populations were tested for wound-induced PI-II accumulation with the assumption that spr1 homozygotes would display a significantly reduced systemic response, as is the case for prosystemin antisense plants (McGurl et al., 1992). F2 plants showing a deficiency (<10% of WT levels) in the systemic response comprised approximately one-quarter of the population (data not shown) and were selected as putative spr1 homozygotes. A polymerase chain reaction (PCR)-based assay (Li and Howe, 2001) was used to test these plants for the presence of the 35S::prosys transgene. Individuals lacking the transgene were brought to the greenhouse for collection of F3 seed. To verify homozygosity of the spr1 allele, the deficiency in wound-induced systemic PI-II expression and insensitivity to exogenous systemin were confirmed in F3 seedlings. Southern blot analysis confirmed the absence of 35S::prosys. Selected spr1 homozygotes (either spr1-1 or spr1-2) were back-crossed again to WT (cv Castlemart). Seedlings in the resulting F2 populations were scored for wound-induced systemic accumulation of PI-II, as well as for systemin-induced PI-II accumulation. The results were consistent with the expectation that spr1-1 and spr1-2 impair both responses and behave as single recessive mutations.

RNA gel blot analysis

RNA was isolated from tomato leaves and analyzed by gel blot hybridization as described by Li et al. (2002b). Gels were run in duplicate, with one set stained with ethidium bromide to check for equal loading of the samples and intactness of the RNA. DNA probes were isolated and radiolabeled with [32P-α]dCTP as described by Howe et al. (2000). The tomato EST clones, cLEC9C14 and cLET1D13, were used as probes for detection of AOS1 (Sivasankar et al., 2000) and WIPK transcripts, respectively. Hybridization results were visualized by autoradiography using Kodak XAR-5 film and, when appropriate, quantified using a Phosphorimager (Molecular Dynamics). Hybridization signals were normalized to the signal obtained using a cDNA probe (EST clone cLED1D24) for translation initiation factor eIF4A. To directly compare transcript levels in WT and spr1 plants, blots containing RNA from both plant types were hybridized in the same container, washed under the same conditions, and exposed to film for the same length of time.

Measurement of jasmonic acid

Jasmonic acid was extracted from leaves (10 g FW) of two-leaf-stage plants as previously described (Li et al., 2002b). Dihydrojasmonic acid (DHJA) was used as an internal standard for quantification of JA levels by gas chromatography-mass spectrometry (Li et al., 2002a).

Acknowledgements

We acknowledge Dr Johannes Stratmann for helpful discussions during the course of the work. We also thank Dr Clarence Ryan for providing systemin and OGA, and Dr Jim Dombrowski for providing recombinant prosystemin. Mr Nate Riddle, Mr Benjamin Travis, and Dr Hans Weber provided helpful assistance with JA measurements. Tomato EST clones, cLED1D24, cLEC9C14, and cLET1D13, were obtained from the Clemson University Genomics Institute. This research was supported by grants from the National Institutes of Health (R01Gm57795), the US Department of Energy (DE-FG02–91ER20021), and the Michigan Agricultural Experiment Station at Michigan State University.

Ancillary