Crosstalk among stress responses in plants: Pathogen defense overrides UV protection through an inversely regulated ACE/ACE type of light-responsive gene promoter unit

Abstract

Plants often have to cope with two or more environmental hazards simultaneously. Such coincidences require instantaneous decisions on relative severity and consequential crosstalk between the respective signaling cascades. Among the frequently encountered threats are pathogen infections and UV irradiation, both of which trigger specifically targeted defense responses by means of changes in gene transcription rates. In Petroselinum crispum, pathogen defense has been shown to be associated with extensive metabolic reprogramming, including strong repression of the UV-protective flavonoid biosynthetic pathway. Here we show that one of the involved genes, encoding acyl-CoA oxidase, responds positively to UV light and negatively to a pathogen-derived elicitor through an inversely regulated promoter unit consisting of two almost identical ACGT-containing elements (ACEs). This unit, when either introduced into an unrelated promoter or generated by mutation of a differently composed unit, confers the same type of response pattern on the recipient genes, confirming its general functionality at a convergence site of two largely distinct signaling pathways. Similarly large, rapid, and partly inverse effects of UV light and elicitor were observed for several mRNAs encoding common plant regulatory factors (CPRFs) that exhibit distinct dimerization and DNA-binding properties. This striking coincidence suggests a major role of common plant regulatory factors in mediating the apparent switch in the function of ACGT-containing elements from positive UV light to negative elicitor or pathogen responsiveness.

Gene activation is a major, probably decisive, component of numerous elaborate defense responses evolved by plants as sessile, permanently stress-exposed organisms. Particularly well studied stress conditions triggering active, transcriptionally regulated defense measures are pathogen attack (1), UV irradiation (2), extreme temperatures (3, 4), herbivory (5, 6), wounding (6), and drought (4). If these conditions are causally and functionally related, such as herbivory and mechanical wounding, or wounding and potential pathogen ingress, certain degrees of overlap of the respective defense programs are to be expected. However, how does a plant respond if two unrelated threats coincide? To address this question, we analyzed the crosstalk between pathogen defense and UV protection at the acyl-CoA oxidase (ACO) gene promoter as one of two previously identified sites of inverse signal entry (7).

For most of the major plant stresses, including pathogen attack and UV exposure, the molecular mechanisms of gene activation have been studied more or less extensively. By contrast, the simultaneously occurring gene silencing has been poorly investigated, though the reallocation of resources from temporarily dispensable metabolic activities may be essential for successful defense (8). Preferred experimental tools for such studies are suspension-cultured cells or protoplasts, e.g., from parsley (Petroselinum crispum) (9, 10), in combination with pathogen-derived signal molecules (elicitors) (11, 12) or UV-containing white light (13). In parsley cells treated in such a manner, two major, widely occurring, plant-specific, positively cis-acting promoter regions, as well as the corresponding trans-acting factors, have been identified: one involved in pathogen defense and the other in UV protection. The pathogen- or elicitor-response element contains one or multiple copies of the TGAC or W-box core sequence (14), which is specifically regulated by the WRKY transcription factor family (15), whereas several UV-responsive genes are activated through a light-regulatory promoter unit (13) comprising one basic leucine zipper (bZIP)-binding, ACGT-containing element (ACE) (16) and one MYB-recognition element (MRE) (17).

Although a few examples of strong and rapid gene silencing by elicitor have been observed (7, 8, 18), the responsible promoter elements have not been identified. Among the silenced genes are those encoding ACO (7) and chalcone synthase (CHS) (18), catalyzing, respectively, a major supply reaction and the first committed step of UV light-induced flavonoid biosynthesis. Both genes respond positively to UV light and, subsequent to activation by UV light, respond negatively to elicitor/infection. Here we analyze the parsley ACO gene promoter, which features several different types of sequence motif with similarity to established or putative signal response elements (top line in Fig. 1A) (7).

Figure 1.

Promoter/GUS constructs used for protoplast transformation and measurements of UV-light and elicitor responsiveness (fold-induction values as indicated by color code). (A) Block mutation and deletion series using either direct translational fusions with the GUS reporter gene (Upper) or with the cauliflower 35S minimal promoter in between (Lower). (B) Mutational and deletion analysis of PcACO-IV. (C) Comparison of PcACO-IV with native or mutated light-regulatory units from the PcCHS and AtCHS gene promoters. (D) Comparison of PcACO-IV with native and mutated PcH3–7 gene promoter fragments. Where applicable, UV light and elicitor were given simultaneously. GUS activities were recorded 8 h after onset of treatment. Each value is the mean from at least eight independent measurements, with the SE ranging from 8% to 35%, except for the first and last lines in A (≈48%).

Materials and Methods

Cultured Cells and Conditions for Treatments.

Suspension-cultured parsley cells (Petroselinum crispum cv. Hamburger Schnitt) were propagated and treated either with UV-containing white light or with a crude elicitor preparation (50 μg/ml) from the oomycete Phytophthora sojae (previously named Phytophthora megasperma), as described elsewhere (18). Where applicable, the relative timing of UV light and elicitor application for combined treatments is indicated in the figure legend.

Cloning of DNA.

All materials and methods used for DNA cloning and analysis have been described (19).

Transient Expression Assay.

Again, all relevant materials and methods, including treatments with UV light (20) or with the P. sojae-derived synthetic peptide elicitor, PEP25 (11, 19), used here at 0.33 μg/ml, have been described. Each assay contained 5 μg Sca-linearized plasmid per ≈2 × 10⁶ parsley protoplasts. Two such standard assays were mixed, divided into four equal portions, and used for three different treatments (UV light, elicitor, and UV light + elicitor) and an untreated control. Incubations were carried out for 8 h.

RNA-Blot Analysis.

RNA was isolated, separated on agarose gels (15 μg per lane), blotted onto Nylon membrane (Boehringer Mannheim), and analyzed as described (16).

Results

The ACO Gene Promoter.

We used the TATA-proximal part of the ACO gene promoter as reference to inactivate all sequences containing the ACE or MRE consensus, either individually by mutation or in a deletion series, and then analyzed the resulting promoter fragments in conjunction with the β-glucuronidase (GUS) reporter gene for UV-light and subsequent elicitor responsiveness (Fig. 1A). Only those mutations or deletions affecting the two most TATA-proximal ACE sequences abolished induction by UV light and consequently repression by elicitor. Conversely, the promoter fragment containing these two sequences alone (fragment PcACO–IV in Fig. 1A) retained full positive UV-light and negative elicitor responsiveness and therefore was used as a more precisely defined reference in the following series of experiments (Fig. 1 B–D).

The ACE/ACE Light-Regulatory, Elicitor-Responsive Unit.

As PcACO-IV represented an unusual type of light-regulatory unit, consisting of two nearly identical ACE motifs instead of the established ACE/MRE combination (13, 16, 17), we next tested two small mutations by generating fully identical ACE motifs, either with or without a C after the ACGT core (fragments PcACO-IV-D5 and -D6). In contrast to the poorly active version containing the additional C, the C-less version responded even more strongly to UV light than the reference unit, and was as strongly repressed by elicitor, demonstrating that two completely identical ACE motifs can constitute a functional, positively acting UV-response element and negatively acting elicitor-response element. Mutation of a single, although nonclassical (21), W-box core sequence (PcACO-IV-D7) occurring in reverse order (GTCA) within the second ACE motif did not significantly alter the response pattern, indicating its functional irrelevance in this context. By contrast, shortening of the reference unit from 72 to 42 bp (PcACO-VI) reduced, and an additional major internal deletion (PcACO-VII) largely abolished, the UV-light response. Even insertion or deletion of a single C in the central, C-rich region of PcACO-IV, as well as a few other such small changes at different positions between the two ACE motifs, greatly reduced the response factor (data not shown), suggesting that both the spacing and the nature of the intervening base pairs are critical for quantitative performance of the unit as a whole.

The general functionality of the ACE/ACE unit was tested by according changes in the established ACE/MRE-containing light-regulatory unit of the PcCHS promoter (Fig. 1C). Replacement of the MRE portion with a second ACE, either in the form occurring in the PcACO promoter (PcCHS-m1) or as a direct repeat of the first ACE (PcCHS-m2), yielded in both cases functional units, although with 2- to 3-fold reduced activities relative to the respective original versions. Inclusion of the corresponding Arabidopsis thaliana AtCHS gene promoter fragment (22) in this test series confirmed that the coupling of positive UV and negative elicitor responsiveness is not confined to parsley but is probably a widely occurring phenomenon in plants.

Finally, a gain-of-function experiment using a PcH3–7 promoter fragment from the parsley histone H3–7 gene is shown in Fig. 1D. This promoter contains only a single ACE and, in this form, responds negatively to both UV light and elicitor in cultured cells (8) and not significantly at all in the protoplast assay (Fig. 1D). When an additional ACE, identical in sequence and relative position with the first ACE in PcACO, was introduced by block mutation (PcH3–7m), the previously unresponsive fragment displayed the predicted response pattern. A somewhat lower UV-light response factor than the one observed with PcACO-IV may again indicate an appreciable influence of the remaining promoter sequence on absolute expression levels, whereas the type of response is obviously determined by element specificity.

Induced Changes in Common Plant Regulatory Factor (CPRF) mRNA Levels.

Several independent searches have previously yielded seven ACE-binding proteins in parsley, PcCPRF1–7, all of them from the bZIP (basic leucine zipper) family of transcription factors (16, 23, 24), which bind as hetero- or homodimers with different affinities to various sequences containing the ACGT motif (24, 25). We now attempted to monitor possible UV light- or elicitor-induced changes by immunoblotting of four selected PcCPRF proteins (PcCPRF1, -2, -4, and -5), but found all of them to be too low in abundance to be detected and quantified under the conditions used (data not shown). We therefore made use of the much more sensitive hybridization methods at the RNA level and investigated instead the behavior of the PcCPRF mRNAs.

Fig. 2 demonstrates large UV light- and elicitor-induced changes in the accumulation patterns of nearly all of the seven known PcCPRF mRNAs. In particular, large opposite effects were observed for two of the four most abundant mRNAs, PcCPRF1 and PcCPRF2. PcCPRF1 mRNA was strongly and rapidly induced by UV light, repressed by subsequent addition of elicitor, and unresponsive to elicitor alone. Conversely, PcCPRF2 mRNA was repressed by UV light and induced by elicitor, even when the cells had been pre-exposed to UV light. All other PcCPRF mRNAs, as far as at all detectable, differed greatly from these two isoforms as well as from one another. PcACO mRNA showed a behavior similar to that of PcCPRF1 mRNA, except for a late induction by elicitor (e.g., relative to the induction of PcCPRF2 mRNA).

Figure 2.

Changes in mRNA abundance in suspension-cultured parsley cells treated with UV light, elicitor, or UV light and elicitor (E). Where applicable, elicitor was added 4 h after the onset of irradiation. RNA blots were hybridized with cDNA probes as indicated. C, control.

Discussion

These results demonstrate the existence of two previously unexplored modes of gene regulation in plants: activation by UV light and subsequent repression by elicitor through the same light-regulatory unit, and the action of this unit through two similar or even identical ACE motifs. Whether the switch from positive to negative function of a promoter element in response to an overriding signal is a rare case or a more widely occurring principle remains to be elucidated. At any rate, a plausible explanation for the observed switch in ACE/ACE or ACE/MRE function would be the replacement of an activating with a silencing combination of CPRF isoforms or, in the latter case, of CPRF and/or MYB isoforms. Distinct DNA-binding properties, most probably enhanced by covalent modifications (24, 26), have been demonstrated for several PcCPRF proteins (24, 25), whereas changes in their abundance, as expected from the large changes in mRNA amounts, have yet to be verified.

Irrespective of mechanistic details, the most likely candidates for mediating the switch in ACE function would be such inversely regulated CPRF isoforms as PcCPRF1 and PcCPRF2, with the latter type replacing the former in the putative transcription complex on exposure of previously UV-irradiated cells to elicitor or infections (Fig. 3). A seeming discrepancy of this notion with the recent suggestion that PcCPRF2 acts as a transcriptional activator, and PcCPRF1 acts as a repressor, in transgenic yeast (24) may be explained by greatly differing experimental conditions. Whatever the putative CPRF-containing complex may look like in detail, our model, as depicted in Fig. 3, would in principle explain all PcACO (Fig. 2) and PcCHS (1) mRNA expression modes tested, except for the late induction of PcACO mRNA by elicitor. However, this latter effect, in contrast to all other effects shown here, was not observed in several independent experiments and is probably a cell-line-specific phenomenon.

Figure 3.

Model illustrating the proposed mechanism of gene activation/inactivation by exchange of transcriptional regulators, most probably including CPRF1, CPRF2, and other CPRF isoforms.

In general terms, these results demonstrate that two distinct signaling cascades can interact through an inversely used promoter element that may be regulated either by positively or by negatively acting members of the same transcription factor family. More specifically, this mechanism would enable a plant perceiving a pathogen-derived signal to rapidly shut off a previously activated, and under these circumstances apparently less urgently needed, UV-protection program by interfering directly with the mode of activation. That the responsible promoter element is to some extent variable in sequence outside the essential core region (Fig. 1 B–D), and that all known genes from the UV-protective flavonoid pathway contain at least one ACE/MYB-based light-regulatory unit and are coinduced by UV light (U. Hartmann and B. Weisshaar, personal communication), suggests that pathogen defense overrides UV protection by repressing all genes of the flavonoid glycoside pathway concomitantly.

However, the apparent variability of the sequence context seems to have major limitations, particularly with regard to the immediate vicinity and the spacing of the ACGT core sequences. Our finding that the nature of a single nucleotide downstream of the core sequence—in this case, either A or C—greatly affects the magnitude of the UV-light response is in close agreement with previous observations in similar contexts (20, 25, 27–29). Furthermore, the present data as well as numerous previous results indicate for most cases a strictly, or at least largely, invariable spacing requirement for the ACE/ACE (this study) or ACE/MYB (B. Weisshaar, personal communication) core sequences.

The major focus of this study was the mode of repression of UV-light-induced genes by elicitor. It should be noted, however, that even in the given case of a seeming lack of overlap of the mechanisms of UV protection and pathogen defense, there are some metabolic activities with multiple, biochemically identical but physiologically divergent functions that have essential roles, and therefore are strongly induced, in either stress condition. A characteristic example in parsley is the small family of genes encoding four distinct isoforms of phenylalanine ammonia-lyase, each with overlapping functions in various branches of phenylpropanoid metabolism, including the initial, common step in UV-light-induced flavonoid and elicitor- or pathogen-induced furanocoumarin biosynthesis (30, 31). Three of the four isoforms are induced by UV light as well as elicitor (31). At least in one established case (isoenzyme 1), activation of the responsible gene involves the same two promoter elements, boxes P and L, that respond positively to both UV light and elicitor and share no sequence similarity with ACEs (30–32).

It may be interesting to note in this connection that all four phenylalanine ammonia-lyase isoforms respond with similar initial timing and to similar extents to wounding and elicitor treatment of parsley leaves (31). This latter result may well be indicative of a more extensive metabolic overlap of pathogen defense and wounding, similar to recent reports on other plant species (33). Analogous examples of extensively overlapping, induced stress responses are those mounted to wounding and herbivory (6) and to low-temperature or high-salinity stress and drought (4, 34, 35). All these examples contrast greatly with the very limited overlap of pathogen defense and UV protection.

Apart from these considerations, the mechanism of crosstalk investigated here appears to be embedded in a more broadly designed hierarchy of stress responses (36), with pathogen defense overriding UV protection by selective transcriptional down-regulation of one or a few metabolic pathways, and heat shock (37), as well as possibly nutrient depletion (38), overriding both by much more generally acting means. It will be interesting to see how the various functionally and metabolically overlapping stress responses, including those mentioned above, are regulated when triggered simultaneously, and whether the hierarchy observed so far extends to such closely interrelated stress situations.

Abbreviations

ACE

ACGT-containing element

ACO

acyl-CoA oxidase

CHS

chalcone synthase

CPRF

common plant regulatory factor

GUS

β-glucuronidase

MRE

MYB-recognition element