UV and blue light signalling: pathways regulating chalcone synthase gene expression in Arabidopsis

Authors

  • Gareth I. Jenkins,

    Corresponding author
    1. Plant Molecular Science Group, Division of Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, Bower Building, University of Glasgow, Glasgow G12 8QQ, UK;
    • Author for correspondence: Gareth I. Jenkins Tel: +44 141330 5906 Fax: +44 141330 4447 Email:G.Jenkins@bio.gla.ac.uk

    Search for more papers by this author
  • Joanne C. Long,

    1. present address: Department of Botany, North Carolina State University, Box 7612, Raleigh, NC 27695–7612, USA;
    Search for more papers by this author
  • Helena K. Wade,

    1. Plant Molecular Science Group, Division of Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, Bower Building, University of Glasgow, Glasgow G12 8QQ, UK;
    Search for more papers by this author
  • Matthew R. Shenton,

    1. Plant Molecular Science Group, Division of Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, Bower Building, University of Glasgow, Glasgow G12 8QQ, UK;
    Search for more papers by this author
  • Tatiana N. Bibikova

    1. present address: Biology Department, Pennsylvania State University, University Park, PA 16802, USA
    Search for more papers by this author

Summary

UV-B, UV-A and blue light control a variety of aspects of plant development via distinct photoreceptors and signalling pathways. The known photoreceptors for UV-A/blue light are cryptochrome (cry)1 and cry2, and the phototropism photoreceptor, phototropin. Redox processes are important in cry and phototropin signal transduction. A specific photoreceptor for UV-B has not been identified and there appear to be several possible UV-B signalling pathways. We are investigating the UV and blue light regulation of transcription of the chalcone synthase gene (CHS) in Arabidopsis. Experiments with photoreceptor mutants show that distinct UV-A/blue (cry mediated) and UV-B photoreception systems control CHS expression. Experiments with an Arabidopsis cell suspension culture show that the UV-B and cry1 signalling pathways differ kinetically and pharmacologically. In contrast to some other UV-B responses, the UV-B induction of CHS does not appear to involve oxidative stress signalling. Promoter elements and candidate transcription factors that effect CHS induction have been identified. Interactions within a network of UV-B, cry and phytochrome signalling pathways regulate CHS expression. Synergistic interactions between the UV-B pathway and distinct UV-A and blue-light pathways maximize the response. In addition, specific phytochromes positively control the cry1 pathway via distinct potentiation and coaction effects, and negatively regulate the UV-B pathway.

Introduction

Light has numerous regulatory effects on plant growth and development and these are mediated by several classes of photoreceptors. The phytochrome photoreceptors are well characterized and absorb principally red and far-red light (Quail et al., 1995). Phytochromes also absorb UV and blue light, but most responses of plants to these wavelengths are mediated by several additional photoreceptors. These specific UV and blue light photoreceptors are involved in controlling a range of responses in plants, including stem extension, phototropism, leaf and chloroplast development, chloroplast movement, stomatal opening, flowering time and the expression of various genes (Briggs & Huala, 1999; Cashmore et al., 1999; Lin, 2000).

Excellent progress has been made in recent years in characterizing three of the UV/blue light photoreceptors: cryptochromes 1 and 2 (cry1 and cry2) and the phototropism photoreceptor phototropin (Lin, 2000). These photoreceptors absorb UV-A (320–390 nm), blue (390–500 nm) and, to some extent, green light. UV-B (280–320 nm), which has the potential to cause macro-molecular damage (e.g. to DNA), but also elicits physiological responses, is not detected by these photoreceptors. As discussed below, the mechanisms of UV-B perception are not clear.

In order to understand how UV and blue light elicit specific responses it is necessary to define the mechanisms of signal trans-duction. Although information is accumulating for different photoreceptors and downstream responses, a clear sequence of signalling events has not been delineated for any UV/blue light-regulated response. It is therefore a priority to obtain insights into the relevant cellular and molecular mechanisms.

Several recent reviews discuss UV and blue light photoreceptors and signal transduction and the reader is referred to these for detailed information (Briggs & Huala, 1999; Cashmore et al., 1999; Lin, 2000). A further review is beyond the scope of this short article and so we will confine ourselves to a summary of some key points. This brief introduction to UV and blue light perception is followed by an account of our recent research into the UV and/blue light regulation of gene expression.

UV/blue light photoreceptors and signalling processes

Cryptochromes

The cryptochromes, cry1 and cry2 (Cashmore et al., 1999; Lin, 2000), have amino acid sequence similarity to microbial DNA photolyases and have binding sites for two chromophores, most likely a flavin and a pterin based on experiments with heterologously expressed proteins. However, neither cryptochrome has DNA photolyase activity.

The cryptochromes mediate a range of responses to blue and UV-A light. Their functions have been analysed in studies of cry1 and cry2 mutants and transgenic plants over-expressing CRY1 and CRY2. cry1 is the principal photoreceptor involved in the suppression of hypocotyl or stem extension by blue light (Ahmad & Cashmore, 1993). It also controls extension and expansion of other organs, such as leaf petioles, cotyledons and the leaf lamina (Jackson & Jenkins, 1995). cry2 is involved in the control of stem extension, but principally in low fluence rates of blue light; high fluence rates of blue light promote the post-translational destruction of cry2 (Lin et al., 1998). A major function of cry2 is in the regulation of flowering time (Guo et al., 1998). Arabidopsiscry2 mutants flower much later than wild-type in long days, but not short days.

As discussed below, cry1 mediates the UV-A/blue light induction of several genes involved in flavonoid biosynthesis and anthocyanin accumulation in Arabidopsis. However, cry2 has a minor role in this response (Wade et al., 2001; Ahmad et al., 1998; see ‘Genetically distinct UV-A/blue and UV-B phototransduction pathways’); cry1 is also involved in the circadian regulation of chlorophyll a/b-binding protein gene expression in Arabidopsis (Somers et al., 1998).

Thus, cry1 and cry2 have distinct functions in controlling plant development, although there is overlap in such responses as the suppression of hypocotyl extension. Given their range of functions it is likely that the cryptochromes will be coupled to branching signal transduction pathways. Whether cry1 and cry2 have similar primary mechanisms of action is not yet known. Because they resemble microbial DNA photolyases in their chromophore binding region and in possession of flavin and pterin chromophores, it is possible that they function via a similar primary mechanism, namely electron transfer. There is evidence for the involvement of electron transport and plasma membrane redox activity in a number of blue light responses (Jenkins et al., 1995; Long & Jenkins, 1998, see ‘Pharmacological dissection of the distinct UV-B and UV-A/blue light signalling pathways controlling CHS expression’). Rapid plasma membrane depolarization is observed following blue light exposure in hypocotyls (Spalding & Cosgrove, 1989) and other systems (Jenkins et al., 1995) and blue light has been shown to induce the rapid opening of an anion channel in Arabidopsis hypocotyls (Cho & Spalding, 1996). Thus plasma membrane redox activity and associated ion fluxes appear to be important early events in blue light signalling. There is evidence that calcium is involved in UV-A/blue light signal transduction in the regulation of gene expression (Christie & Jenkins, 1996; Frohnmeyer et al., 1997; see ‘Pharmacological dissection of the distinct UV-B and UV-A/blue light signalling pathways controlling CHS expression’), but there is very little information on events coupling plasma membrane activities to calcium signalling and to processes in the nucleus.

Phototropin

In recent years a photoreceptor for phototropism has been characterized by Briggs and coworkers (Briggs & Huala, 1999). This protein is encoded by the Arabidopsis NPH1 gene and was identified following the isolation of several nonphototropic hypocotyl (nph) mutants (Liscum & Briggs, 1995). Previously it had been reported that blue light induced the reversible phosphorylation of an approximately 120 kDa plasma membrane protein in several species. Liscum & Briggs (1995) demonstrated that this protein was NPH1.

The NPH1 protein contains a serine/threonine protein kinase domain and two LOV (light, oxygen, voltage) domains that bind a flavin chromophore (Huala et al., 1997; Christie et al., 1999). The heterologously expressed NPH1 flavoprotein becomes autophosphorylated upon exposure to blue light and has spectral properties consistent with it functioning as a photoreceptor for phototropism (Christie et al., 1998). NPH1 has now been renamed phototropin.

Flavin-mediated redox activity initiated by blue light absorption is evidently central to phototropin signal transduction. However, it is not yet clear how redox activity and consequent phototropin autophosphorylation initiate cellular signal transduction and ultimately phototropic bending. It appears that phototropin functions in partnership with at least one other protein, NPH3 (Motchoulski & Liscum, 1999), in association with the plasma membrane. Moreover, a recent report (Baum et al., 1999) indicates that phototropin signal transduction involves cytosolic calcium. However, details of the mechanism of transduction are lacking at present.

UV-B perception

A key event in UV-B signal transduction in animal cells is the generation of a ‘DNA damage signal’ emanating from the nucleus which initiates cytoplasmic signalling events leading to the induction of various genes (Bender et al., 1997). There is little evidence as yet for an equivalent process in plants. If there were such a mechanism, processes that remove DNA damage would be expected to diminish UV-B-induced responses. These processes would include the blue light mediated photorepair of DNA via DNA photolyase. It is therefore significant that blue light given together with UV-B hyperstimulates, rather than reduces, expression of the gene encoding chalcone synthase (CHS) in Arabidopsis (Fuglevand et al., 1996; see ‘Synergistic interactions between UV and blue light pathways’). This observation argues against a DNA damage signal being involved in this response. For different reasons, Conconi et al. (1996) concluded that a DNA damage signal was unlikely to be involved in the UV-B induction of gene expression in tomato.

The mechanism of UV-B perception in plants is ill-defined and it appears that several different mechanisms may exist (Jenkins, 1999a). There is increasing evidence that some UV-B signalling pathways inducing gene expression overlap with wound-response and pathogen-defence pathways. UV-B causes the generation of reactive oxygen species (ROS), which have been shown to stimulate the expression of pathogenesis-related genes such as PR1 in tobacco (Green & Fluhr, 1995). In addition, UV-B can stimulate the production of salicylic acid, jasmonic acid and ethylene, all of which function as wound and/or defence signalling molecules. Conconi et al. (1996) reported that the UV-B induction of proteinase inhibitor gene expression in tomato was mediated by jasmonic acid. Moreover, there is recent evidence that distinct pathways involving jasmonate, salicylic acid or ethylene mediate the UV-B induction of several genes in Arabidopsis (Surplus et al., 1998; A.-H.-Mackerness et al., 1999). However, it would be premature to conclude that all UV-B responses are mediated via wounding or pathogenesis signalling pathways. Our studies indicate that the UV-B induction of CHS gene expression in Arabidopsis does not involve oxidative stress signalling or jasmonate (J.C. Long & G.I. Jenkins, unpublished; Jenkins, 1999a; see ‘The UV-B induction of CHS expression does not appear to be mediated by oxidative stress’ below). Hence there appear to be pathways of UV-B perception that are independent of wound and pathogenesis signalling. Whether plants possess a specific UV-B photoreceptor remains an open question.

Regulation of chalcone synthase gene expression by UV and blue light

The aim of our research is to understand the cellular and molecular mechanisms involved in the UV and blue light regulation of gene expression. The focus is on transcription of the CHS gene in Arabidopsis. CHS is the first enzyme in the flavonoid-specific branch of the phenylpropanoid biosynthesis pathway (Dixon & Paiva, 1995; Weisshaar & Jenkins, 1998). As such it is a key enzyme in secondary metabolism. Products of the pathway are important in several ways, including in protection against abiotic stresses. Some flavonoids absorb UV-B and function as a protective shield in the epidermal tissues (Bornman et al., 1997; Jenkins et al., 1997).

CHS expression is stimulated by light and various other environmental and endogenous stimuli. There is an extensive literature detailing the effects of different light qualities on flavonoid (e.g. anthocyanin) accumulation and the expression of CHS and other flavonoid biosynthesis genes in a range of species (Beggs et al., 1986; Hahlbrock & Scheel, 1989; Dixon & Paiva, 1995; Mol et al., 1996). Parsley (plants and cell culture) and Sinapis alba (white mustard) were used for much of the research that established the principles of light regulation of CHS (Bruns et al., 1986; Ohl et al., 1989; Batschauer et al., 1991; Ehmann et al., 1991). Expression is stimulated by UV and blue light and this can be seen as a protective mechanism as it promotes the accumulation of UV-absorbing flavonoids in the epidermis. In addition, depending on the species and the stage of development, expression may be stimulated by red and/or far-red light absorbed by phytochromes. In several species there is a change of photoreceptor usage during development, in that phytochrome control is diminshed or lost in older tissue (Batschauer et al., 1991; Frohnmeyer et al., 1992). The effect of light on CHS expression is principally transcriptional, as initially demonstrated in parsley cells (Chappell & Hahlbrock, 1984).

In the last decade Arabidopsisthaliana has been used increasingly for studies of CHS regulation. Ausubel and coworkers were first to demonstrate that Arabidopsis CHS is induced by UV and blue light (Feinbaum & Ausubel, 1998; Feinbaum et al., 1991; Kubasek et al., 1992). Induction by phytochrome is confined to very young seedlings (Kaiser et al., 1995; Batschauer et al., 1996). Since Arabidopsis has a single CHS gene there is no question of differential expression within a gene family in response to different stimuli. As outlined below, we have started to dissect the UV and blue light signalling pathways regulating CHS expression in Arabidopsis.

Genetically distinct UV-A/blue and UV-B phototransduction pathways

It is important to define which photoreceptors mediate the UV and blue light induction of CHS. This can be achieved by studying the responses of photoreceptor mutants to different inductive light qualities. In our experiments we have used mature Arabidopsis leaf tissue rather than seedlings, because in leaves CHS is induced only by UV and blue wavelengths. Seedlings less than 6-d-old also show induction by far-red light and, to a small extent, red light, mediated by phytochromes (Kaiser et al., 1995). The plants we use routinely are grown in a low fluence rate of white light that does not induce CHS for 3 wk prior to transfer to inductive light qualities.

UV-A and blue light, both of which are absorbed by cryptochrome, induce CHS expression. In addition CHS is induced by UV-B light. A cry1 null mutant has much reduced blue light induction of CHS and undetectable UV-A induction (Jackson & Jenkins, 1995; Fuglevand et al., 1996), indicating that cry1 is the principal photoreceptor mediating the UV-A/blue light response. A cry2 null mutant is unaltered in UV-A/blue light induction (Wade et al., 2001). Nevertheless, cry2 appears to be capable of mediating induction when cry1 is absent, since the cry1 cry2 double mutant shows less induction of CHS in UV-A/blue light than the cry1 single mutant. Significantly, mutants lacking one or both cryptochromes are unaltered in the UV-B induction of CHS, demonstrating that some other photoreception system must be responsible for this response (Fuglevand et al., 1996; Wade et al., 2001). Although phytochrome can detect UV-B light, the absence of red and far-red light induction of CHS in mature leaves rules out phytochrome as the UV-B photoreceptor for this response.

Kinetically distinct UV-B and UV-A/blue light signalling pathways control CHS expression

To identify components of the UV-B and UV-A/blue (cry mediated) light signalling pathways regulating CHS expression we have used an Arabidopsis cell suspension culture system. The advantage of cell cultures is that the cells are largely homogeneous, in contrast to intact plants where different tissues show differential gene expression and may possess distinct signalling pathways. Moreover, with cell cultures it is easier to introduce compounds such as pharmacological inhibitors and agonists.

The Arabidopsis cell culture is routinely grown in a low fluence rate of white light (20 µmol m−2 s−1) that does not induce CHS expression, and it responds similarly to mature leaf tissue when transferred to inductive light qualities. As with mature leaves, UV-B and UV-A/blue light are inductive, whereas red and far-red light are not (Christie and Jenkins, 1996).

Kinetic experiments with the cell culture indicate that distinct UV-B and UV-A/blue phototransduction pathways regulate CHS expression. Transcript accumulation is detectable within a few hours of continuous illumination with each of these light qualities (Christie & Jenkins, 1996). However, much less exposure to UV-B than UV-A/blue light is required to induce CHS. In the experiment shown in Fig. 1, cells were illuminated with UV-B or UV-A/blue light for various times and then returned to noninductive white light. They were harvested for RNA extraction 6 h after the start of illumination. As shown in the Figure, more than 30 min UV-A/blue light is needed to induce CHS transcripts and 60 min gives only a weak response. However, as little as 5 min UV-B is effective and 15 min gives a strong response. In fact it has been reported that less than a second of UV-B exposure is sufficient to induce an increase in CHS promoter – luciferase expression in cultured parsley cells (Frohnmeyer et al., 1999). Thus it appears that sufficient signal is generated by a brief exposure to UV-B to promote transcription of CHS whereas the cry1 signalling pathway needs more prolonged stimulation. This implies that the pathways are mechanistically different.

Figure 1.

Different kinetics of UV-B and UV-A/blue light-mediated induction of CHS expression. Arabidopsis cell suspension culture was grown in a low (20 µmol m−2 s−1) fluence rate of white light (LW) that does not induce CHS expression, as described by Long & Jenkins (1998). Cells were illuminated with 3 µmol m−2 s−1 UV-B or 80 µmol m−2 s−1 UV-A/blue light for the durations indicated and then returned to LW. All samples were harvested 360 min after the start of illumination. Total RNA was isolated and equal amounts fractionated on a denaturing agarose gel. A blot of the RNA was hybridized to a radioactively labelled CHS DNA probe, washed and autoradiographed. The blot was subsequently re-hybridized to a control H1 DNA probe (Christie & Jenkins, 1996). All methods were as described by Long & Jenkins (1998).

The UV-B induction of CHS expression does not appear to be mediated by oxidative stress

As mentioned in ‘UV-B perception’ above, there is evidence that the induction of some genes by UV-B is mediated by ROS. We have obtained evidence that the UV-B induction of CHS expression in Arabidopsis cells is not mediated by oxidative stress.

Several biotic and abiotic stresses involve oxidative stress. In general this entails the excessive production of ROS, in particular the superoxide radical, which is converted to hydrogen peroxide via the action of superoxide dismutases. Hydrogen peroxide then accumulates until removed by the action of catalase or peroxidases. A classic example is the oxidative burst involved in plant defence responses (Lamb & Dixon, 1997). Hydrogen peroxide induces the expression of genes encoding enzymes involved in cellular protection. We therefore examined the effect of hydrogen peroxide on CHS expression in Arabidopsis cells. Addition of the compound to cells in noninductive white light failed to induce CHS expression compared with the effect of an inductive illumination (Fig. 2, upper panel). In addition, we examined the effect of 3-amino-1,2,4-triazole (ATZ) on the response. ATZ inhibits catalase and would therefore allow hydrogen peroxide produced during normal cell metabolism to accumulate in the cells. However, ATZ failed to stimulate CHS expression (Fig. 2, lower panel). It therefore appears that oxidative stress mediated by hydrogen peroxide does not stimulate CHS expression in our cell culture system. Similar findings were reported for soybean cells (Levine et al., 1994). However, additional experiments, using different generators of ROS, could be undertaken to further address this issue.

Figure 2.

Hydrogen peroxide and 3-amino-1,2,4-triazole do not induce CHS expression. Arabidopsis cell suspension culture was grown in a low (20 µmol m−2 s−1) fluence rate of white light (LW) that does not induce CHS expression, as described by Long & Jenkins (1998). Hydrogen peroxide or 3-amino-1,2,4-triazole (ATZ) were added at the concentrations indicated and the cells kept in LW for 6 h. Control cells were illuminated with 80 µmol m−2 s−1 UV-A/blue light for 6 h without any additions. Cells were harvested, total RNA was isolated and equal amounts fractionated on a denaturing agarose gel. A blot of the RNA was hybridized to a radioactively labelled CHS DNA probe, washed and autoradiographed. The blot was subsequently re-hybridized to a control H1 DNA probe (Christie & Jenkins, 1996). All methods were as described by Long & Jenkins (1998).

A key experiment is to remove ROS and investigate the effect on the UV-B induction of CHS. We found that the addition of ROS scavengers to Arabidopsis cells failed to prevent the UV-B induction of CHS. We used pyrrolidine dithiocarbamate (PDTC) and N-acetylcysteine (NAC), which were found to strongly inhibit the UV-B induction of pathogenesis-related protein 1 (PR-1) accumulation in tobacco (Green & Fluhr, 1995). The compounds were added to the Arabidopsis cell culture prior to illumination with UV-B, analogous to the approach taken by Green & Fluhr (1995). As shown in Fig. 3, neither compound affected the UV-B induction of CHS in the cells. Similarly, there was no effect on the UV-A/blue inductive response. Since we used up to 2 µM NAC, the highest concentration nontoxic to cells, and 100 M PDTC, the same concentration that Green & Fluhr (1995) found was effective when sprayed on plants, it seems very unlikely that the lack of effect of these compounds was due to a lack of uptake. We conclude that the UV-B signalling pathway inducing CHS in Arabidopsis does not appear to involve ROS, in contrast to the UV-B pathway inducing expression of some other genes, such as PR-1 in tobacco.

Figure 3.

Active oxygen scavengers do not prevent the UV-B induction of CHS expression. Arabidopsis cell suspension culture was grown in a low (20 µmol m−2 s−1) fluence rate of white light (LW) that does not induce CHS expression, as described by Long & Jenkins (1998). Pyrrolidine dithiocarbamate (PDTC) or N-acetylcysteine (NAC) were added to aliquots of the cells at the concentrations indicated. Treated and untreated control cells were left to incubate for 1 h in LW. Cells were then illuminated with 80 µmol m−2 s−1 UV-A/blue or 3 µmol m−2 s−1 UV-B light for 6 h and then harvested. Total RNA was isolated and equal amounts fractionated on a denaturing agarose gel. A blot of the RNA was hybridized to a radioactively labelled CHS DNA probe, washed and autoradiographed. The blot was subsequently re-hybridized to a control H1 DNA probe (Christie & Jenkins, 1996). All methods were as described by Long & Jenkins (1998).

Pharmacological dissection of the distinct UV-B and UV-A/blue light signalling pathways controlling CHS expression

We have used a pharmacological approach to characterize the UV-B and UV-A/blue phototransduction pathways regulating CHS. The use of inhibitors has to be undertaken with considerable caution. The absence of an effect may simply mean that a compound has not entered the cells and so a positive control is needed. Similarly, if an inhibitor is effective it must be shown to have no effect on a control response to ensure that its action is specific rather than general. We employed several controls to ensure that the compounds we used were having specific effects (Christie & Jenkins, 1996; Long & Jenkins, 1998).

In our initial experiments (Christie & Jenkins, 1996) we demonstrated that the UV-B and UV-A/blue light signalling pathways were pharmacologically distinct and, furthermore, that they were different from the phytochrome signalling pathway reported to regulate CHS in tomato hypocotyl and cultured soybean cells (Neuhaus et al., 1993; Bowler et al., 1994). Evidence was presented that the UV-B and UV-A/blue light signalling pathways involved calcium and protein phosphorylation (Christie & Jenkins, 1996). The experiments suggested the involvement of an internal calcium pool, rather than flux from outside the cell. The pathways differed in the likely involvement of calmodulin, since the UV-B induction of CHS, but not the UV-A/blue light induction, was inhibited by the calmodulin antagonist W-7.

Long & Jenkins (1998) extended analysis of the signalling pathways. An important question was whether redox processes were involved, given that cry1 has similarity to DNA photolyase which has an electron transfer mechanism (see ‘Cryptochromes’ previously). Ferricyanide, an electron acceptor that does not enter cells, was found to strongly inhibit CHS expression in UV-A/blue and UV-B light. In addition, the flavoprotein antagonist, diphenylene iodonium, inhibited expression. These data suggest that electron transport may be an important component of both the cry1 and UV-B signalling pathways regulating CHS. Moreover, because ferricyanide is cell impermeable, the plasma membrane is implicated as a primary site of action. It is conceivable that cry1 associates with the plasma membrane to initiate electron transport processes, which in turn promote signal transduction. For instance, redox processes may generate a membrane potential and ion fluxes and/or cause the production of second messengers that promote calcium release from an internal pool. As discussed above, (see ‘Cryptochromes’) it is well known that blue light initiates membrane potential changes (Spalding & Cosgrove, 1989, Jenkins et al., 1995) and activates an anion flux associated with hypocotyl elongation (Cho & Spalding, 1996). Moreover, blue light initiates H+ATPase activity in stomatal guard cells (Jenkins et al., 1995). Thus there is ample evidence to suggest that plasma membrane redox processes are involved in blue light signal transduction. This has to be reconciled with a recent report that cry1 fused to green fluorescent protein is targeted to the nucleus in onion epidermal cells (Cashmore et al., 1999). However, immunological experiments with light-grown Arabidopsis indicate that the nucleus contains only a small proportion of the total cellular cry1 (Guo et al., 1999). Perhaps cry1 moves between the nucleus and cytosol in response to light/dark treatments, although it is not clear how this may relate to plasma membrane redox processes.

Long & Jenkins (1998) obtained further information on the role of calcium in UV-A/blue and UV-B light signal transduction. They obtained evidence that both light qualities may promote the removal of calcium from a cytosolic pool via specific Ca2+ATPases. Addition of the calcium ionophore A23187 plus 10 mM external calcium to the Arabidopsis cells in low fluence rate white light stimulated expression of the calcium-regulated gene TOUCH3 (TCH3). However, exposure to UV-A/blue or UV-B light greatly reduced TCH3 expression under these conditions. Red light was without effect. These data suggest that both light qualities reduced the level of cytosolic calcium in the pool responsible for stimulating TCH3. The removal of calcium from the cytosol occurs via specific Ca2+ATPases. Long & Jenkins (1998) found that a potent Ca2+ATPase inhibitor, erythrosin B, prevented the UV-A/blue light induction of CHS but not the UV-B induction. As would be expected from inhibition of a Ca2+ATPase, there was a concomitant increase in cytosolic calcium as measured by TCH3 expression. As mentioned above, the calmodulin antagonist, W-7, had the opposite effect to erythrosin B, in that it specifically inhibited the UV-B induction of CHS.Long & Jenkins (1998) obtained indirect evidence that this effect may result from the inhibition of a calmodulin sensitive Ca2+ATPase.

Thus the experiments of Christie & Jenkins (1996) and Long & Jenkins (1998) suggest that UV-A/blue and UV-B light regulate the concentration of calcium in one or more internal cellular pools via both efflux and influx. The calcium concentration in the pool(s) may be critical for CHS expression. Clearly it is essential to identify the relevant pool(s) and to obtain direct measurements of the effects of UV-A/blue and UV-B light on the flux of calcium. Work is in progress to address these points. Recent experiments with Arabidopsis seedlings show that blue light does promote a transient increase in cytosolic calcium measured by the calcium-sensitive protein aequorin (Baum et al., 1999). However, this increase is unrelated to the regulation of CHS expression and appears to be associated with the action of phototropin. Clearly it is essential to measure the effects of UV-A/blue and UV-B light on calcium specifically in cells involved in CHS expression. Many questions need to be addressed. Apart from the identity of the proposed internal calcium pool, it is necessary to further examine events at the plasma membrane and to determine how they may be coupled to calcium homeostasis. Also, does cry1 translocate to the nucleus? If so, how? What is the mechanism of UV-B perception in CHS induction? What are the processes downstream of calcium and how do they effect the transcriptional response?

Targets of the signalling pathways

Although we do not yet know how cellular signalling events are coupled to the induction of CHS transcription, progress has been made in identifying the promoter elements and transcription factors involved in the response. Hartmann et al. (1998) used protoplasts obtained from the Arabidopsis cell culture in transient expression experiments to define promoter elements of the CHS gene required for UV-A/blue and UV-B light induction. Promoter sequences were fused to the β-glucuronidase (GUS) reporter and assayed for their ability to direct transcription following transfection into the protoplasts. The Arabidopsis CHS promoter contains a Light Responsive Unit (LRU) which closely resembles elements in other CHS promoters (e.g. parsley; Schulze-Lefert et al., 1989). The LRU alone, fused to a heterologous TATA box, was sufficient to confer responsiveness to UV-B and UV-A/blue light. The LRU contains sequences that bind bZIP and MYB transcription factors and mutation of these elements greatly reduced expressions in response to both UV-A/blue and UV-B light. The identification of the specific factors that mediate the response is now a priority.

Although there is a family of bZIP transcription factors in Arabidopsis, one bZIP is known to be required for expression of CHS: the HY5 protein (Ang et al., 1998). hy5 mutants are altered in several aspects of photomorphogenesis (Chory, 1992), and are defective in the UV-A/blue and UV-B induction of CHS expression (H. K. Wade and G. I. Jenkins, unpublished). The identity of the MYB factor(s) mediating light induction of CHS is not yet known. There are at least 100 MYB genes in Arabidopsis (Kranz et al., 1998) so elucidation of the protein(s) responsible for the light induction of CHS is a challenge. Redundancy of function may be involved and may complicate the assignment of function. Nevertheless, progress is being made in this area.

Signal transduction processes may regulate the amount, activation and/or localization of the relevant transcription factors. HY5 is localized in the nucleus and its abundance increases following illumination (Osterlund et al., 2000). This results from a decrease in the rate of protein turnover. Post-translational destruction of HY5 occurs in darkness, promoted by an interaction with the COP1 protein (Osterlund et al., 2000). It is interesting that the UV-A/blue and UV-B light induction of CHS in Arabidopsis cells requires protein synthesis; Christie & Jenkins (1996) showed that cycloheximide prevented induction but did not affect a control response. Thus, some component(s) must be synthesized rapidly in response to illumination to effect the CHS response. This could be a MYB transcription factor, as some of these are regulated by light (Kranz et al., 1998), or some other component.

In summary, progress is being made in defining the sequence of events from photoreception to initiation of CHS transcription. In particular, with UV-A/blue light induction the photoreceptor is defined and candidate signalling events and transcription factors have been identified. With UV-B, the initial perception mechanism is unknown. In both responses a priority is to identify specific proteins involved in signal transduction.

A network of phototransduction pathways regulates CHS

Although it is important to identify components of the UV-A/blue and UV-B inductive pathways regulating CHS expression, it is becoming increasingly clear that the elucidation of these pathways will provide only part of the information needed to understand the regulation of CHS expression in vivo. The reason is that interactions between different phototransduction pathways regulate CHS expression. Hence, to fully understand the control of CHS it is necessary to define the components that mediate interactions between pathways as well as those that make up the primary pathways.

Interactions between signalling pathways are extremely important (Jenkins, 1999b). Plants have evolved mechanisms to sensitively detect and respond to a wide range of environmental stimuli. In addition, plant cells perceive a range of endogenous metabolic and developmental signals. It is essential for the plant to integrate all this information to produce appropriate responses that enhance its survival. That is, responses must be made in the context of other responses. To achieve integration there must be interaction, or ‘cross-talk’ between signal transduction pathways. Hence we should think in terms of signalling networks rather than isolated pathways.

CHS is regulated by a variety of external and endogenous stimuli, ranging from light and temperature to metabolites and plant growth regulators (Tsukaya et al., 1991; Deikman & Hammer, 1995; Dixon & Paiva, 1995; Leyva et al., 1995). Hence there is considerable scope for interactions between pathways. CHS expression therefore provides an excellent system to study the mechanisms of cross-talk.

Synergistic interactions between UV and blue light pathways

UV-B, UV-A and blue light each stimulate CHS expression up to about 10-fold in mature Arabidopsis leaf tissue (Fuglevand et al., 1996). However, particular combinations of light treatments give a much larger increase. Fuglevand et al. (1996) reported that the level of CHS-GUS expression in transgenic Arabidopsis was approximately 60-fold in plants exposed to UV-B and blue light together, indicating a synergistic interaction between separate phototransduction pathways. A similar synergistic interaction was observed with UV-B and UV-A light. However, blue and UV-A light were not synergistic and gave an additive level of stimulation.

Further experiments showed that blue light given before UV-B synergistically enhanced the level of CHS-GUS expression and CHS transcript accumulation, whereas UV-B given before blue light did not. Therefore, blue light appeared to generate a ‘signal’ that increased the subsequent response to UV-B. The enhancement was still observed when several hours of darkness separated the blue and UV-B treatments, indicating that the blue light signal was relatively stable. Similar findings were reported previously for CHS expression in parsley cell cultures (Ohl et al., 1989). Fuglevand et al. (1996) further observed that the synergistic interaction between UV-A and UV-B was present only when the two treatments were given simultaneously. Hence the signal generated by UV-A appears to be transient. Since the signals derived from UV-A and blue light have different stability we can conclude that the phototransduction pathways that produce them are distinct. Additional evidence in support of this hypothesis is that illumination of plants with blue light followed by UV-A and UV-B together gives approximately twice the level of expression observed with either synergistic combination alone (UV-B plus blue light or UV-B plus UV-A light). Thus it appears that CHS expression is maximized by the additive effect of two synergistic interactions. Clearly this is a very complex system.

Experiments with mutants lacking cry1 and cry2 reveal that neither photoreceptor detects the light involved in the synergistic interactions; cry1 and cry2 single and double mutants retain both synergistic interactions (Fuglevand et al., 1996, Wade et al., 2001). Thus some unknown UV-A and blue light photoreceptors appear to mediate these responses.

The inductive UV-B and UV-A/blue phototransduction pathways and the blue and UV-A synergism-specific pathways that regulate CHS are shown in Fig. 4.

Figure 4.

Model showing interactions within the network of phototransduction pathways controlling CHS expression in mature Arabidopsis leaf tissue. Inductive pathways initiated by UV-B light, cry1 and cry2 are shown along with the distinct blue and UV-A synergism pathways that interact with the UV-B pathway. Positive regulation of the cry1 pathway (potentiation and coaction) by specific phytochromes is indicated; phytochromes with minor roles are placed in brackets. In addition, phyB is shown as a negative regulator of the UV-B pathway, acting upstream of the points of interaction of the synergism pathways. Reproduced with permission from Wade et al. (2001).

Interactions with phytochrome signalling pathways

Recent experiments demonstrate that the network of signalling pathways regulating CHS is in fact more complex than indicated above, in that phytochromes are also involved. In mature Arabidopsis leaf tissue there is no induction by either red or far-red light. Phytochrome induction of CHS is observed only in seedlings younger than about 6-d-old (Kaiser et al., 1995). Nevertheless, we have found that phytochrome signalling pathways interact with both the UV-A/blue (cry1 mediated) and UV-B inductive pathways regulating CHS (Wade et al., 2001).

We examined CHS induction in mutants lacking either or both of phytochromes A and B. The cry1-mediated induction of CHS was much reduced in the phyB mutant whereas the phyA mutant responded like wild-type. The double mutant resembled phyB. Therefore phyB is required for maximal induction of CHS by cry1 in mature leaves. This requirement can be termed coaction rather than synergism, as the phyB pathway is itself noninductive. This term has been used previously by Mohr and coworkers (Mohr, 1994) to describe interactions between phytochrome and blue light signalling.

The above findings differ from those reported by Batschauer et al. (1996) for young Arabidopsis seedlings. These authors observed that CHS-GUS induction was unaltered in phyA, phyB and phyA phyB mutant plants. The most likely explanation is that the competence for interactions between pathways changes during development rather like the competence for phytochrome induction of CHS changes.

We have observed a further interaction between phytochrome and the cry1 signalling pathway. Pre-treatment of plants with 1–3 h of red light enhances the subsequent induction mediated by cry1 (Wade et al., 2001). We refer to this effect as potentiation. The effect of red light is undiminished in phyA and phyB single mutants, but much reduced in the double mutant. This indicates that either phyA or phyB can mediate the response and that they are functionally redundant. Nevertheless, an effect of red light pretreatment was still observed in the double mutant, albeit with a lower level of induction, indicating that phytochromes other than phyA and phyB may also mediate the effect.

The fact that the coaction and potentiation responses differ in phytochrome involvement indicates that they are distinct and may involve different mechanisms and signalling pathways.

A further interaction is observed between phytochrome and the UV-B pathway inducing CHS (Wade et al., 2001). In this case the phyB mutant and phyA phyB double mutant showed elevated expression compared to the wild-type, whereas phyA was unaltered. This indicates that phyB acts as a negative regulator of the UV-B signalling pathway.

A model of the light signalling network regulating CHS

The above findings reveal remarkable complexity in the photoregulation of CHS expression. A model showing the interactions between pathways is shown in Fig. 4. The level of expression in natural light environments will depend on the balance of light qualities and the positive and negative interactions within the network. Whether similar interactions regulate the expression of other flavonoid biosynthesis genes remains to be established. At present we know nothing about the signalling components that mediate these interactions and their identification presents a major challenge for the future.

Conclusions and Perspective

In recent years important advances have been made in identifying and characterizing photoreceptors that mediate responses to UV-A/blue light. However, it is likely that plants possess further UV-A/blue photoreceptors (Wade et al., 2001; Lin, 2000) and work should be undertaken to identify these. Similarly, it is important to identify and characterize components involved in UV-B perception and to establish whether a specific UV-B photoreceptor exists.

One of the key approaches that will identify components of light signalling pathways is the isolation and characterization of mutants. The genetic approach has been responsible for much of the recent progress in identification of photoreceptors and signalling components and, crucially, the assignment of function. Indeed, the cloning and functional characterization of cry1 (Ahmad & Cashmore, 1993) and phototropin (Huala et al., 1997), which represent major advances in the field, were facilitated by isolation of the corresponding mutants. More recently, tremendous progress in understanding phytochrome signalling has resulted from the genetic approach (e.g. Hoecker et al., 1999; Ni et al., 1999). We have isolated Arabidopsis mutants altered in the UV and blue light-regulation of CHS expression (Jackson et al., 1995; Jenkins, 1999a; H. K. Wade, M. R. Shenton, G. Fuglevand, R. A. Brown, and G. I. Jenkins, unpublished). These icx (increased chalcone synthase expression) mutants have amplified responses to particular stimuli that regulate CHS and their further characterization should lead to the identification of important regulatory components.

Some information is available on UV and blue light signal transduction, but it relates more to cellular processes than to specific components. The identification of proteins involved in these signalling pathways is therefore a priority. Phototropism, hypocotyl growth suppression and the regulation of flavonoid biosynthesis genes are the best characterized systems and offer the best prospects for elucidation of signalling pathways. Progress is likely to come from mutant isolation and characterization, functional genomics and biochemical approaches in cellular systems. Sequencing of the Arabidopsis genome has revealed the sequences of various putative signalling components but the functions of these proteins need to be examined by reverse genetics and/or expression studies in plants or cell culture systems.

Finally, it is important to remember that UV-A/blue and UV-B signalling pathways do not function in isolation in the intact plant but are integrated with perception mechanisms for other environmental and endogenous signals. It is evident that there are interactions between several UV/blue light signalling pathways and between phytochrome and cryptochrome signalling pathways. At present there is hardly any information on the components that mediate these interactions, but their identification is essential if we are to understand fully UV and blue light signalling in the intact plant.

Acknowledgements

G.I.J. is grateful to the UK Biotechnology and Biological Sciences Research Council for the support of his research. J.C.L. and H.K.W. received BBSRC PhD studentships.

Ancillary