Light induces phenylpropanoid metabolism in Arabidopsis roots


For correspondence (fax +1 765 496 7213; e-mail


Experiments have shown that many phenylpropanoid genes are highly expressed in light-grown Arabidopsis roots. Studies employing reporter gene constructs have indicated that the expression of these genes is localized not only to the lignifying root vasculature, but also to non-lignifying tissues, such as the root cortex, suggesting that the proteins encoded by these genes may be involved in aspects of phenylpropanoid metabolism other than lignification. Consistent with this hypothesis, roots of etiolated and soil-grown plants contain almost no soluble phenylpropanoids, but exposure to light leads to the accumulation of flavonoids, as well as high levels of coniferin and syringin (coniferyl and sinapyl-4-O-glycosides), compounds not previously reported to be accumulated in Arabidopsis. To elucidate the mechanism by which light induces root secondary metabolism, extracts of mutants defective in light perception and light responses were analyzed for phenylpropanoid content. The results of these assays showed that phytochrome (PHY)B and cryptochrome (CRY)2 are the primary photoreceptors involved in light-dependent phenylpropanoid accumulation, and that the hypocotyl elongated (HY5) transcription factor is also required for this response. The presence of phenylpropanoids in etiolated roots of cop (constitutively photomorphogenic)1, cop9, and det (de-etiolated)1 mutants indicate that the corresponding wild-type genes are required to repress root phenylpropanoid biosynthesis in the absence of light. Biochemical analysis of root cell walls and analysis of phenylpropanoid gene expression suggest that coniferin and syringin accumulation may be the result of both increased biosynthesis and decreased conversion of these compounds into other phenylpropanoid end products. Finally, our data suggest that the accumulation of coniferin, syringin, and flavonoids in Arabidopsis roots is a high-irradiance response (HIR), and suggest that comparative analysis of light- and dark-grown Arabidopsis roots may provide new insights into both phenylpropanoid biosynthesis and light signaling in plants.


Light plays an important role in a number of plant developmental processes, including the initiation of cell differentiation in vegetative meristems, chloroplast development, hypocotyl elongation, leaf expansion, and flowering (Dale, 1988; Gruissem, 1989; Halliday and Fankhauser, 2003). At the same time, light exposure also affects both primary and secondary plant metabolism. It induces primary metabolism, such as amino acid and chlorophyll biosynthesis (Chory, 1991; Gruissem, 1989), as well as influences the levels of secondary metabolites, including alkaloids and phenylpropanoids (Mancinelli, 1985; Mancinelli and Rabino, 1978; Vazquez-Flota and De Luca, 1998).

The phenylpropanoid pathway is responsible for the biosynthesis of many metabolites, including flavonoids, hydroxycinnamic acid esters, and the precursors for lignin (Boudet, 2000; Lewis and Yamamoto, 1990). Light induction of flavonoid biosynthesis has been well characterized (Jenkins et al., 2001; Mancinelli, 1985; Mancinelli and Rabino, 1978), primarily through the analysis of light-dependent anthocyanin accumulation (Jenkins et al., 2001; Mancinelli, 1985). The accumulation of these compounds is, in part, because of the transcriptional regulation of the gene encoding chalcone synthase (CHS), an enzyme that catalyzes the first committed step in flavonoid biosynthesis (Jenkins et al., 2001; Sakuta, 2000). Other phenylpropanoids, including sinapate esters in Arabidopsis, have also been shown to accumulate in a light-dependent manner (Randhir and Shetty, 2003; Ruegger et al., 1999); however, in contrast to what is known about the regulation of flavonoid biosynthesis, little is known about the proteins involved in the light-dependent regulation of other classes of phenylpropanoids.

Although much work has been carried out to characterize the metabolism of soluble phenylpropanoids in aerial tissues, phenylpropanoid metabolism in Arabidopsis roots remains largely unstudied. A surprising result of experiments examining the tissue specificity of phenylpropanoid gene expression is that many of these genes are strongly expressed in roots, sometimes to levels higher than those found in all other plant tissues. Although this expression has been explained by the fact that roots lignify, in many cases, phenylpropanoid genes have been found to be highly expressed in cells that do not normally deposit lignin (Bell-Lelong et al., 1997; Lacombe et al., 2000; Nair et al., 2002; Ohl et al., 1990). These results suggest that this expression may be involved in the production of metabolites other than lignin.

Like phenylpropanoid genes, many genes encoding light receptors in Arabidopsis are highly expressed in roots (Goosey et al., 1997; Ruppel et al., 2001; Somers and Quail, 1995a; Tóth et al., 2001). Arabidopsis contains at least three classes of light receptors: the phytochromes, the cryptochromes, and phototropin (Neff et al., 2000; Quail, 2002). The roles of these proteins in regulating photomorphogenesis have been investigated through the characterization of photoreceptor-deficient mutants (Chamovitz and Deng, 1996; Wang and Deng, 2003). Except for phototropin, which is required for blue-light-induced root-negative phototropism, the role of these proteins in roots remains to be determined, although there is evidence that phytochrome is required for a slight positive phototropic response in Arabidopsis roots (Jarillo et al., 2001; Quail, 2002; Ruppel et al., 2001).

We have characterized soluble phenylpropanoid metabolism in roots of Arabidopsis, and we report here the identification of the monolignol glucosides (MGs), coniferin and syringin, and show that MG and flavonoid accumulation is light-induced. Further, we have shown that this response requires proteins involved in photomorphogenesis and is repressed by proteins that act in skotomorphogenesis, and involves increases in phenylpropanoid gene expression and changes in the cell specificity of this expression in response to light. We also show that growth in light leads to opposite changes in insoluble and soluble phenylpropanoid levels in roots, suggesting that MG accumulation may be because of altered utilization of these compounds in lignin and hydroxycinnamic acid biosynthesis, as well as from generally increased phenylpropanoid biosynthesis. Finally, our results suggest that the accumulation of MGs and flavonoids is regulated at least partially independently.


Light-grown Arabidopsis roots contain monolignol and flavonoid glycosides

Although the soluble phenylpropanoids found in aerial tissues of Arabidopsis have been well documented (Bloor and Abrahams, 2002; Chapple et al., 1994; Graham, 1998), the products that accumulate as a result of phenylpropanoid gene expression in roots remain predominantly uncharacterized. To determine whether Arabidopsis roots accumulate soluble phenylpropanoid metabolites, methanolic extracts from roots of 2-week-old plate-grown plants were analyzed by HPLC. Extracts of light-grown roots contained a number of UV-absorbing metabolites (Figure 1a). The absorbance spectra of the later-eluting compounds suggested that they were flavonoids, and the absence of these compounds in tt (transparent testa)4 roots supported this preliminary identification (Figure 1a). LC–MS analysis showed that two of these compounds have masses identical to those of the glucosyl-rhamnosyl disaccharides of quercetin (Glu-Rha-Quercetin; neg. m/z 609) and kaempferol (Glu-Rha-Kaempferol; neg. m/z 593), flavonoids known to be accumulated in Arabidopsis leaves (Graham, 1998), and are consistent with the identification of quercetin and kaempferol conjugates in Arabidopsis roots (Saslowsky and Winkel-Shirley, 2001).

Figure 1.

HPLC analysis of soluble metabolites accumulated in roots of wild-type and mutant plants.

(a) Methanolic root extracts of light-grown wild-type, fah1-2 (lacking F5H activity), C4H-F5H (overexpressing the F5H gene under the control of the C4H promoter), tt4 (lacking CHS activity), and ref8 (defective in the gene encoding p-coumarate 3′-hydroxylase) plants analyzed by HPLC.

(b) Mass spectra of coniferin and syringin in wild-type root extracts as compared to mass spectra of coniferin and syringin standards. Mass spectra of standards show that coniferin and syringin appear as negative ion adducts with M + 44.

(c) HPLC analysis of methanolic root extracts of 2-week-old wild-type plants grown in the dark on plates, and in the soil.

The UV spectra of the two early-eluting compounds (Figure 1, retention times 14.3 and 16.9 min) were distinct from those of the flavonoids, and this result suggested that these compounds could be derivatives of coniferyl and sinapyl alcohol (Figure 2). As an initial test of this identification, roots of mutant and transgenic plants, known to have altered phenylpropanoid biosynthesis, were analyzed for the accumulation of the putative coniferyl and sinapyl alcohol derivatives. Light-grown roots of the fah (ferulic acid hydroxylase)1-2 mutant, defective in ferulate 5-hydroxylase (F5H) activity, lacked detectable levels of the putative sinapyl alcohol derivative and accumulated only the early-eluting compound (Figure 1a). In contrast, in roots of transgenic plants overexpressing F5H under control of the cinnamate 4-hydroxylase (C4H) promoter (Meyer et al., 1998), the putative coniferyl alcohol derivative was below detectable limits, and levels of the later-eluting compounds were substantially elevated. Root extracts of ref (reduced epidermal fluorescence)8, a mutant blocked upstream of both coniferyl and sinapyl alcohol biosynthesis (Franke et al., 2002a,b), contained neither compound (Figure 1a). The changes in phenylpropanoid biosynthesis seen in light-grown fah1-2, C4H-F5H, and ref8 roots mirror the changes observed in lignin composition in stems of these plants (Chapple et al., 1992; Franke et al., 2002a; Meyer et al., 1998). Thus, the altered metabolite levels in the roots of these plants supported the initial identification of the two compounds as being derivatives of coniferyl and sinapyl alcohol.

Figure 2.

Model of the phenylpropanoid pathway leading to flavonoids, lignin, and MGs.

4CL, 4-(hydroxy)cinnamoyl CoA ligase; C3′H, p-coumaryl-shikimate/quinate 3′-hydroxylase; C4H, cinnamate 4-hydroxylase; CAD, cinnamyl alcohol dehydrogenase; CCoAOMT, caffeoyl CoA O-methyltransferase; CCR, cinnamoyl CoA reductase; CHS, chalcone synthase; COMT, caffeic acid/5-hydroxyferulic acid O-methyltransferase; CST, hydroxycinnamoyl CoA:shikimate hydroxycinnamoyltransferase; F5H, ferulate 5-hydroxylase; HCALDH, hydroxycinnamaldehyde dehydrogenase; PAL, phenylalanine ammonia-lyase; SAD, sinapyl alcohol dehydrogenase; UGT, UDP-glucose:alcohol glucosyltransferase.

Final identification of the putative coniferyl and sinapyl alcohol derivatives was determined by comparison to phenylpropanoid standards. The retention time and absorption spectra of the unknown compounds were identical to those of the coniferin and syringin, the 4-O-glucosides of coniferyl and sinapyl alcohol (data not shown). The identity of these compounds was confirmed by LC–MS analysis (Figure 1b).

Root phenylpropanoid content is light- and developmentally regulated

Although light-grown roots accumulate both MGs and flavonoids, roots of etiolated plants and young soil-grown plants (Figure 1c) accumulate only very low levels of phenylpropanoids. This relationship between light exposure and phenylpropanoid accumulation was observed in all ecotypes tested, including Columbia, Landsberg, and Wassilewskija (data not shown). These data indicate that soluble phenylpropanoid accumulation in roots is induced by light, with levels of MGs in light-grown roots being 100–400-fold higher than that in dark-grown roots (Figure 1a,c). To determine if the accumulation of these secondary metabolites is also developmentally regulated in a light-independent manner, root extracts of field-grown plants and 4-week-old pot-grown plants were analyzed. Roots of these older plants accumulate 20–40% of the coniferin found in light-grown roots, but contain only low levels of syringin and flavonoids (data not shown).

When 2-week-old etiolated plants were transferred to light, MGs accumulated within 24 h (data not shown). To test the ability of roots to autonomously detect light and accumulate MGs, roots of 2-week-old etiolated plants were excised in the light, transferred to new plates, and exposed to light for 48 h, or returned to the dark for an equivalent period of time. MG levels increased only in light-treated roots (Figure 3a); however, the level of MGs in excised light-treated roots increased by about only fourfold as compared to the 20-fold increase in non-excised roots (Figure 3a). These data suggest that the shoot may play a role in light-induced root secondary metabolism, either by perceiving the light signal, or by providing phenylalanine or downstream phenylpropanoid intermediates for MG synthesis.

Figure 3.

Accumulation of coniferin and syringin in roots after transfer to light or dark conditions.

(a) Roots of 2-week-old dark-grown plants were excised and exposed to 48 h of white light. Two-week-old dark-grown control plants were also excised, during which they were exposed to around 30 min of light and were then returned to darkness for 48 h. Coniferin (black bars) and syringin (gray bars) levels in excised roots were quantified along with non-excised roots exposed to similar light conditions.

(b) Coniferin (black bars) and syringin (gray bars) in roots of 1-week-old light-grown plants transferred to light or darkness for 2 and 4 days. Control roots were kept in the light.

Error bars represent SE for at least triplicate assays.

Light exposure was necessary for MG accumulation in roots (Figure 3a), and was also required for the maintenance of high MG levels (Figure 3b). MG content decreased significantly in roots of light-grown plants transferred to the dark, with coniferin content decreasing at a faster rate than syringin (Figure 3b). These data indicate that the accumulation of coniferin and syringin requires constant light exposure, and that the levels of these metabolites measured in all the experiments presented here likely represent steady state levels established by both MG synthesis, and subsequent utilization or catabolism.

Phenylpropanoids accumulate throughout light-grown roots

A survey of MG levels in roots during plant development showed that these compounds accumulate rapidly after germination in light, but level-off after around 10 days (Figure 4a). Three experiments were conducted subsequently to determine where the MGs are accumulated in the root. First, to examine MG accumulation as the root differentiates and matures, a 1-cm region of root tissue, beginning 1 cm below the hypocotyl/root junction, was excised from plants of various ages and analyzed for phenylpropanoid content. HPLC analysis of these samples showed that MGs are found in the youngest tissue examined, and continue to accumulate as the root matures (Figure 4b). In a complementary experiment, extracts from a series of 1-cm sections of 2-week-old roots were analyzed by HPLC. The results of these analyses showed that all sections of light-grown roots accumulate MGs, and that the level is the highest in the basal regions of the root (Figure 4c). In a final analysis, secondary roots were found to contain MG levels comparable to those found in primary roots (Figure 4d), although flavonoid levels in secondary roots were roughly twice that in primary roots (data not shown). These data indicate that root phenylpropanoid content is likely to represent an integration of root age and root light exposure.

Figure 4.

Characterization of the accumulation of coniferin and syringin in wild-type roots.

(a) Accumulation of coniferin (filled circles) and syringin (open circles) in whole roots over time.

(b) Accumulation of coniferin (filled circles) and syringin (open circles) in a 1-cm region of root tissue, spanning the region 1–2 cm below the hypocotyl/root junction, from plants of increasing age.

(c) Coniferin (black bars) and syringin (gray bars) content of a series of 1-cm sections of 2-week-old light-grown roots.

(d) Coniferin (black bars) and syringin (gray bars) content of primary and secondary roots of 2-week-old light-grown seedlings.

Error bars represent SE of at least triplicate assays.

Light-induced phenylpropanoid accumulation in roots requires proteins involved in photomorphogenesis

To dissect genetically how light induces phenylpropanoid biosynthesis, roots of mutants blocked in light perception and response were analyzed for altered phenylpropanoid content. Of the phytochrome mutants assayed, light-grown roots of plants lacking phytochrome (PHY)B contained substantially less MGs and flavonoids than wild type, whereas phyA roots accumulated only slightly lower levels of flavonoids (Figure 5) and, in some experiments, coniferin (data not shown). Although these results indicate that PHYB may be the primary phytochrome involved in this response, the presence of less MGs and flavonoids in the phyA/phyB double mutant than in either single mutant suggests that at least one of these proteins must be active for the accumulation of substantial amounts of phenylpropanoids (Figure 5). Decreased MG levels in roots of a leaky hy2 allele, partially blocked in the biosynthesis of the phytochrome chromophore (Kohchi et al., 2001), confirm that active phytochrome is required for phenylpropanoid accumulation. Analysis of MG levels in mutants blocked in the cryptochromes 1 (hy4) and 2 (cry2) showed that the loss of CRY2 has a greater effect on phenylpropanoid accumulation than that of CRY1 (Figure 5). Taken together, these results suggest that PHYB and CRY2 may be the primary photoreceptors involved in the accumulation of MGs, consistent with the relatively high level of PHYB and CRY2 expression throughout the roots of light-grown Arabidopsis seedlings (Goosey et al., 1997; Tóth et al., 2001).

Figure 5.

Accumulation of soluble metabolites in roots of 2-week-old light-grown mutants that are blocked in light perception or response.

Quantification of coniferin (black bars), syringin (gray bars), and Rha-Glu-Quercetin (dark gray bars) in wild-type, phyA, phyB, phyA/B, hy2, hy4, and hy5 plants analyzed by HPLC. Error bars represent SE for at least triplicate assays. nd, Not detected.

Previous work has shown that phytochromes and cryptochromes are required for a number of photomorphogenic processes where the light stimulus is perceived by other photoreceptors (Folta and Spalding, 2001; Mancinelli et al., 1991). An example of this ‘co-action’ of different photoreceptors is the requirement of phytochrome for cryptochrome-mediated Arabidopsis hypocotyl elongation and anthocyanin production (Ahmad and Cashmore, 1997). To determine if both phytochromes and cryptochromes act as photoreceptors in roots leading to phenylpropanoid accumulation, MG levels were assayed in roots of plants grown in either red or blue light. Compared to roots exposed to white light of similar intensity, roots grown in blue or red light accumulated almost no soluble phenylpropanoid metabolites (Figure 6). In a complementary experiment, 2-week-old etiolated plants were transferred to dim white, red, or blue light for 48 h. Very low levels of MGs were observed in red and blue-light-incubated roots, as compared to control plants (data not shown). These results suggest that light perception by both red-light-absorbing phytochromes and blue-light-absorbing cryptochromes may be required for phenylpropanoid accumulation.

Figure 6.

Accumulation of soluble metabolites in roots of plants grown in white, dim-white, red, and blue light. Quantification of coniferin (black bars), syringin (gray bars), and Rha-Glu-Quercetin (dark gray bars) in wild-type Columbia and Landsberg 2-week-old roots analyzed by HPLC. Error bars represent SE for at least triplicate assays. nd, Not detected.

Finally, roots of the hypocotyl elongated (hy5) mutant contained very low levels of MGs and flavonoids (Figure 5), indicating that the HY5 transcription factor is required for the accumulation of all these compounds. Further, considering what is known about light signaling from previous studies (Schäfer and Bowler, 2002), these data suggest that HY5 acts downstream of both phytochromes and cryptochromes in the regulation of soluble phenylpropanoid levels. Residual levels of MGs in hy5 roots, however, suggest that other transcription factors may also be involved in their accumulation.

Mutations in genes involved in skotomorphogenesis also alter root phenylpropanoid metabolism

The involvement of HY5 in the induction of root metabolism by light strongly suggested that proteins that repress HY5 activity in other tissues may also act to repress phenylpropanoid biosynthesis in dark-grown roots. To test this hypothesis, cop (constitutively photomorphogenic)1, cop9, det (de-etiolated)1, det2, and det3 plants were grown in the dark for 2 weeks, after which extracts of their roots were analyzed by HPLC. Although det2 and det3 roots contained the same low amounts of phenylpropanoids seen in etiolated wild-type plants (data not shown), roots of the cop1 and cop9 mutants contained significant levels of flavonoids and MGs (Figure 7). In contrast, roots of the det1 mutant accumulated flavonoids, but little, if any, coniferin (Figure 7). These results indicate that COP1, COP9, and DET1 act to repress phenylpropanoid accumulation in the dark, although the different phenylpropanoid profiles of these mutants suggest complex and partially independent regulation of secondary metabolism in etiolated roots. The virtual absence of syringin in extracts of these mutants also suggests that syringin accumulation in the dark may be regulated by an as yet unidentified factor(s).

Figure 7.

Accumulation of soluble metabolites in dark-grown roots of mutants that are blocked in skotomorphogenesis.

Methanolic root extracts from 2-week-old light-grown and dark-grown wild-type, cop1, cop9, and det1 plants analyzed by HPLC. Con, coniferin; Syr, syringin.

Light exposure leads to changes in the phenylpropanoid gene expression

Many phenylpropanoid genes are stress- or light-inducible, and characterization of shoot development have revealed that light-dependent regulation of gene expression is an important component of photomorphogenesis. With this in mind, we tested the hypothesis that light-dependent MG and flavonoid accumulation in Arabidopsis roots might, in part, be because of light-dependent induction of phenylpropanoid gene expression (Figure 8). Surprisingly, quantitative RT-PCR (qRT-PCR) revealed that most of the genes examined exhibited comparable levels of expression in light- and dark-grown roots. Phenylalanine ammonia-lyase (PAL)3 and putative cinnamyl alcohol dehydrogenase-like protein (ELI)3-2 were both slightly downregulated in light-grown roots, whereas the expression of 4-(hydroxy)cinnamoyl CoA ligase (4CL)3, caffeic acid/5-hydroxyferulic acid O-methyltransferase (COMT), cinnamyl alcohol dehydrogenase (CAD)-1, and HY5 was increased two- to fivefold. CHS expression was induced in the light approximately 70-fold. In contrast, when RNA from dark-grown roots exposed to 5 h of light was analyzed by qRT-PCR, more widespread changes in phenylpropanoid gene expression were detected, consistent with the relatively rapid accumulation of MGs after light exposure (Figure 3a). Under these conditions, the expression of all three Arabidopsis PAL genes was induced, as was C4H, 4CL3, p-coumaryl-shikimate/quinate 3′-hydroxylase (C3′H), and CAD-C. CHS and HY5 were also highly induced in this experiment. These results suggest that the induction of a number of phenylpropanoid genes may be involved in the short-term accumulation of soluble phenylpropanoids, but that high levels of gene expression is not necessarily required for the maintenance of high MG levels.

Figure 8.

Light-dependent expression of phenylpropanoid genes in roots of 2-week-old plants.

Quantitative RT-PCR was used to determine the relative transcript levels of phenylpropanoid genes in roots of light-grown plants versus dark-grown plants (black bars), and dark-grown plants transferred to light for 5 h versus dark-grown plants (gray bars). 18S rRNA was used as a standardization control, and expression levels were normalized to dark-grown plants. Error bars represent SE for duplicate biological samples. The genes are as follows: 4CL, 4-(hydroxy)cinnamoyl CoA ligase; C3′H, p-coumaryl-shikimate/quinate 3′-hydroxylase; C4H, cinnamate 4-hydroxylase; CAD, cinnamyl alcohol dehydrogenase; CCR, cinnamoyl CoA reductase; CHS, chalcone synthase; COMT, caffeic acid/5-hydroxyferulic acid O-methyltransferase; DHS, 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase; ELI, putative cinnamyl alcohol dehydrogenase-like protein; F5H, ferulate 5-hydroxylase; HY5, hypocotyl elongated; PAL, phenylalanine ammonia-lyase; UGT, UDP-glucose:alcohol glucosyltransferase.

As a complementary approach to qRT-PCR analysis, light-dependent changes in phenylpropanoid gene expression were investigated in a transgenic line carrying a C4H-GUS reporter construct (Bell-Lelong et al., 1997). As with the qRT-PCR analysis of C4H expression, C4H-GUS showed a similar expression pattern in dark- and light-grown roots. The transgene was highly expressed in the vascular tissue (Figure 9a,b), and some GUS expression was also detected in the root cortex of both light-grown and etiolated primary and secondary roots (Figure 9c,f,g). In further support of the qRT-PCR analysis, C4H-driven GUS expression was also strongly induced in the cortex of etiolated roots that were transferred to the light for 24 h (Figure 9d,e). These results suggest that light-induced increases in C4H expression, although relatively modest when compared to genes like CHS, probably contribute to MG and flavonoid accumulation.

Figure 9.

GUS expression in roots driven by the C4H promoter.

(a) C4H-GUS expression in a dark-grown root tip.

(b) C4H-GUS expression in a light-grown root tip.

(c) C4H-GUS expression throughout a 2-week-old light-grown root.

(d) C4H-GUS expression in a dark-grown root section.

(e) Expression in a dark-grown root section transferred to light for 24 h.

(f) C4H-GUS expression in an older light-grown root section close to the hypocotyl/root junction.

(g) C4H-GUS expression in a light-grown root section and in lateral roots.

Black bars in (a), (b), (d), and (f) represent 100 µm, and in (c) and (g) represent 1 mm.

Light exposure alters the hydroxycinnamic acid ester content in root cell walls

In addition to MGs and flavonoids, plant tissues also contain insoluble products of the phenylpropanoid pathway, including lignin and cell-wall-esterified hydroxycinnamic acids. Levels of these metabolites were measured in dark- and light-grown roots to determine if they were also higher in roots grown in the light. Surprisingly, alkaline hydrolysis of dark- and light-grown roots showed that although the levels of esterified p-coumaric acid were not affected by light exposure, light-grown roots contained 10-fold less esterified ferulic acid than dark-grown roots (Table 1). These results suggest that light exposure may lead to opposite changes in soluble and insoluble phenylpropanoid levels, and that re-direction of phenylpropanoid flux may, in part, be responsible for MG accumulation in light-grown roots.

Table 1.  Hydroxycinnamic acid content of root cell walls (n = 4)
Root sampleEsterified hydroxycinnamic acid ester contenta
p-Coumaric acidFerulic acid
  • a

    Picomole milligram per cell wall ± SD.

Dark-grown93 ± 11501 ± 67
Light-grown83 ± 949 ± 9


Light alters secondary metabolism in Arabidopsis roots

Arabidopsis roots accumulate a number of phenylpropanoids in response to light, including dihydroflavonol glycosides and the coniferyl- and sinapyl-alcohol-4-O-glucosides coniferin and syringin. Arabidopsis enzymes that catalyze the glucosylation of coniferyl and sinapyl alcohol have been identified through a large-scale screen of glycosyltransferases (Lim et al., 2001). In trees, MGs have been found to accumulate in conifer xylem prior to lignification of the cambium, and have been implicated as soluble storage forms of the lignin precursors coniferyl and syringyl alcohol (Dharmawardhana et al., 1995). The presence of these compounds in Arabidopsis roots suggests that they may also serve as lignin precursors in Arabidopsis. Further, MGs and other phenylpropanoid glycosides exhibit antibacterial and antiviral activities (Pan et al., 2003), suggesting that the presence of high levels of these compounds in roots may help protect the plant from microorganisms found in the surrounding soil. Finally, given that we know relatively little about the biological functions and activities of MGs, our data suggest that light-grown Arabidopsis roots may exhibit MG-mediated differences from dark-grown roots in ways that have yet to be elucidated.

The induction of MG and flavonoid biosynthesis in roots exposed to light is dramatic, resulting in a 100–400-fold increase in the concentration of these metabolites within 2 weeks of light exposure. In fact, after 10 days of growth in light, the levels of these compounds are equivalent to the levels of soluble phenylpropanoids in leaves, such as sinapate esters and flavonoids. Light-induced flavonoid accumulation has been documented in roots of Alnus glutinosa (Hughes et al., 1999); however, the genetic resources available in Arabidopsis, particularly mutants defective in light sensing and perception, make it the ideal system for the systematic dissection of root phenylpropanoid biosynthesis.

Phenylpropanoid accumulation is regulated by proteins involved in photomorphogenesis and skotomorphogenesis

Anthocyanin accumulation in seedlings of a number of Brassicaceae species is controlled by two light responses: one that induces pigment accumulation after a single, brief irradiation with red light, and one that leads to anthocyanin accumulation after several hours of light exposure (Mancinelli, 1985; Mancinelli and Rabino, 1978). Responses that fall into the second category, called high-irradiance responses (HIRs), are characterized by a requirement for light exposure at high fluence rates, continued light irradiance, and a lack of photo-reversibility (Mancinelli, 1985; Mancinelli and Rabino, 1978). The fact that Arabidopsis roots accumulate phenylpropanoids after days of light exposure, levels of these compounds decrease in roots transferred back to the dark, and short-term irradiation does not induce their accumulation strongly suggests that the induction of phenylpropanoid accumulation in roots is an HIR.

In Arabidopsis, phytochromes and cryptochromes are expressed in roots (Goosey et al., 1997; Sakamoto and Briggs, 2002; Somers and Quail, 1995a; Tóth et al., 2001), although the roles of these proteins in roots are poorly characterized. These data show that phytochromes and cryptochromes are required for the light-dependent accumulation of MGs and flavonoids. Consistent with the observation that accumulation of phenylpropanoids in roots requires a long period of light exposure, the light-stable PHYB photoreceptor may be quantitatively more important for metabolite accumulation than PHYA, which is light-labile (Nagy and Schäfer, 2002). Phytochrome has also been shown to play the major role in anthocyanin HIRs in seedlings of other species (Mancinelli, 1985), and PHYB has been implicated as the primary photoreceptor involved in red-light Arabidopsis HIRs (Nagy and Schäfer, 1999). The lack of MG accumulation in roots exposed to continual red light suggests, however, that CRY2 and possibly CRY1 may also act as photoreceptors in the induction of phenylpropanoids, rather than functioning only as signal transducers downstream of the phytochromes.

The accumulation of secondary metabolites in roots of cop1, cop9, and det1 mutants show that phenylpropanoid biosynthesis is repressed by these proteins in dark-grown roots. COP9 encodes a subunit of the COP9 signalosome, which has been shown to degrade HY5 in a light-mediated fashion (Osterlund et al., 2000; Peng et al., 2001), suggesting that the high levels of coniferin and flavonoids in cop9 roots may be a result of decreased HY5 degradation. COP1 encodes a nuclear-localized protein that mediates the degradation of HY5 by the COP9 signalosome (Ang et al., 1998). Interestingly, COP1 has been shown to translocate to the cytosol upon light exposure in all Arabidopsis tissues except roots. This translocation is thought to be a major factor in increased HY5 activity in the light (Ang et al., 1998). The accumulation of secondary metabolites in light-grown roots therefore suggests that HY5 may be able to act in root nuclei even with COP1 present. One possible explanation for this activity is that COP1 requires the presence of one or more other proteins, such as COP10, to mediate HY5 degradation, and that these proteins themselves may be translocated or otherwise repressed in root cells in the light.

The DET1 protein, which is thought to repress photomorphogenesis through chromatin modification (Benvenuto et al., 2002), has also been implicated in repressing HY5 in wild-type seedlings, as HY5 expression is increased in the det1 mutant (Oyama et al., 1997). The fact that det1 mutation leads to increased levels of only flavonoids in dark-grown roots, however, suggests that DET1 may regulate phenylpropanoid accumulation at a point downstream of HY5 in the signal cascade as a loss of HY5 regulation would be expected to increase both MG and flavonoid levels.

Monolignol glucoside accumulation in roots may be the result of changes in both synthesis and subsequent utilization of soluble phenylpropanoids

In contrast to root flavonoid biosynthesis, we have found little evidence that MGs accumulate in roots because of dramatic changes in phenylpropanoid gene expression (Figure 9). Although it is likely that some genes involved in phenylpropanoid biosynthesis were not tested, a number of genes analyzed have been shown by mutant analysis to be required for MG biosynthesis. One explanation for these results is that most of these genes may also be involved in lignin production, and changes in expression leading to soluble phenylpropanoid biosynthesis may be relatively small. This possibility is consistent with the relatively low levels of C4H-GUS expression in the root cortex, as compared to the stronger C4H-GUS expression in the younger root vascular tissue. The high vascular expression of C4H and other phenylpropanoid genes may therefore hide more subtle changes in expression in other root tissues. Consistent with this hypothesis, AtCAD-1, the only MG-specific gene to be upregulated in response to light, is one of the three CAD genes highly expressed in roots (Goujon et al., 2003), but is not thought to be involved in lignin production (Raes et al., 2003). Thus, light-induced changes in AtCAD-1 gene expression may be more apparent because of a lower background of lignin-specific expression. Another possibility is that the accumulation of MGs is the result of altered phenylpropanoid enzyme activity, rather than gene expression. This argument is supported by the recent finding that DAHP synthase, the first enzyme in the shikimic acid pathway, is redox regulated (Entus et al., 2002).

One surprising result of this study was the discovery that two glucosyltransferases, UGT72E2 and UGT72E3, shown to synthesize MGs in vitro, show little change in gene expression in light-grown roots (Figure 9), and that one of them, UGT72E2, is expressed at high levels in dark-grown roots (data not shown). In pine and other plants, coniferin can be hydrolyzed by β-glucosidase(s), such that the monolignol moiety can be used in lignin biosynthesis (Dharmawardhana et al., 1995; Savidge, 1989). These data suggest that MGs may be synthesized at low levels in dark-grown roots, and that changes in the subsequent metabolism of these compounds to lignin or other phenylpropanoid end products may be altered upon light exposure. This hypothesis is consistent with the significantly lower levels of cell-wall-bound ferulic acid in light-grown roots.

Light may induce phenylpropanoid accumulation in roots growing in soil

Studies of light transmission through soil has shown that far-red and red light can travel relatively far through soil, up to 2 cm in loam and 10 cm in sand, whereas blue light is absorbed almost immediately (Mandoli et al., 1990; Tester and Morris, 1987). Thus, in natural conditions, light exposure may lead to phenylpropanoid accumulation in plant roots growing close to the soil surface. Further, plant roots have been shown to transmit light along the root axis, suggesting that light may be propagated to roots relatively deep in the soil (Mandoli et al., 1990; Sun et al., 2003), where it could induce MG accumulation. Although 2-week-old soil-grown roots lack soluble phenylpropanoids, they are present in roots of older Arabidopsis plants, including those grown in soil for 4 weeks and those that have overwintered outdoors. Soluble phenylpropanoid metabolites have also been identified in roots of a number of other species, including Ligularia duciformis (Gao et al., 1998), Ilex rotunda (Wen and Chen, 1996), and Cosmos caudatus (Fuzzati et al., 1995). The conservation of MGs and flavonoid accumulation among diverse plant species suggest that these compounds may serve important roles in root biology, potentially acting in plant–microbe interactions, plant defense, or as precursors to root exudates. MG accumulation may also provide metabolic pools for the later production of lignin and cell-wall-esterified phenolics in roots, whereas light-induced flavonoids may act to protect roots growing close to the soil from pathogens and possible UV damage.

Arabidopsis roots: a novel system for characterization of light-response pathways

The relative ease with which phenylpropanoids can be analyzed and quantified, combined with the dramatic difference in metabolism associated with light exposure, suggests that Arabidopsis roots may be a useful model system for further study of light-regulated metabolism. Much has been learned about light signal transduction from investigations into both the light-dependent accumulation of anthocyanins in hypocotyls as well as the light-dependent expression of CHS (Batschauer et al., 1996; Jackson et al., 1995; Jenkins et al., 2001). An advantage of the present study is the identification of a number of flavonoid and non-flavonoid metabolites, all of which show potentially overlapping but dissimilar regulation by light. Thus, the individual metabolic profiles of light perception and response mutants may be useful in determining the exact interplay between members of the light-signal transduction cascade that are expressed in Arabidopsis roots.

Experimental procedures

Plant growth

Arabidopsis thaliana L. Heynh., grown on soil, were cultivated at a light intensity of 100 µE m−2 sec−1 at 23°C under a photoperiod of 16 h light/8 h dark in Redi-Earth potting mix (Scotts-Sierra Horticulture Products; Marysville, OH, USA). Unless otherwise noted, all wild-type Arabidopsis used in experiments were of ecotype Columbia. Light-response mutants used in these studies were ordered from the Arabidopsis Biological Center at Ohio State University ( The stock numbers of these mutants are as follows: cop1-5, CS6259; cop9-1, CS6262; cry2-1, CS3732; det1-1, CS6158; det2-1, CD6159; det3-1, CS6160; hy2-1, CS68; hy4-1, CS70; hy5-1, CS71; phyA-201, CS6219; phyB-1, CS69; and phyA-201/phyB-5, CS6224. The cry2-1, det1-1, det2-1, and det3-1 alleles are in a Columbia background, hy2-1, hy4-1, hy5-1, phyA-201, phyB-1, and phyA-201/phyB-5 are in Landsberg background, and cop1-5 and cop9-1 are in Wassilewskija background.

For experiments with plants grown on plates, seeds were surface-sterilized as previously described by Lehfeldt et al. (2000), and plated on modified Murashige and Skoog medium (Murashige and Skoog, 1962). For root experiments, plates were grown vertically in a controlled-environment growth chamber at 21°C. Plants were illuminated under continuous white light at an intensity of approximately 100–150 µE m−2 sec−1.

For the blue- and red-light experiments, plants were grown as previously described by Wade et al. (2001). In brief, plants grown in red light were cultivated under fluorescent bulbs surrounded by a ‘Deep Gold Amber’ filter (No. 135; Lee Filters, Andover, UK) and plants grown in blue light were cultivated under fluorescent bulbs surrounded by a ‘Moonlight Blue’ filter (No. 180; Lee Filters). To grow plants under dim white light, plates were grown under fluorescent lights surrounded by normal white paper. The intensities of light in these conditions approximately were 80 µE m−2 sec−1 under the red filter, 60 µE m−2 sec−1 under the white paper, and 50 µE m−2 sec−1 under the blue filter.

Soluble metabolite analysis

Soluble metabolites from Arabidopsis tissues were collected and analyzed using the same method as has been previously described by Hemm et al. (2003). Compounds were identified based on their retention times and UV spectra as compared to coniferin and syringin standards. Coniferin and syringin were a generous gift from Dr Brian Ellis at the University of British Columbia.

Peak identity was confirmed by LC–MS (LCMS-2010, single quadrupole mass spectrometer detector, atmospheric pressure chemical ionization source; Shimadzu, Kyoto, Japan) using the same method as described by Hemm et al. (2003). Quantification of coniferin and syringin was performed by comparison to known standards.

GUS staining

Roots expressing C4H-GUS were stained following established protocols (Hemerly et al., 1993).

Quantitative RT-PCR

RNA was collected from 2-week-old roots as previously described by Verwoerd et al. (1989). Quantitative PCR and relative quantification of transcripts were performed as previously described by Rider et al. (2003) using an ABI prism 7000 sequence detection system (Applied Biosystems, Foster City, CA, USA).

Cell wall analysis

Extraction and quantification of cell-wall-esterified hydroxycinnamic acids were performed as previously described by Franke et al. (2002b).


The authors would like to thank Dr Brian Ellis for the generous donation of coniferin and syringin standards. This work was supported by grants from the Division of Energy Biosciences, United States Department of Energy, and the National Science Foundation, and was also supported by a fellowship from the Innovation Realization Laboratory (IRL) at Purdue University's Krannert School of Management. IRL is funded through the National Science Foundation's Integrative Graduate Engineering Research and Training Grant Program. This is journal paper number 17 347 of the Purdue University Agricultural Experiment Station.