Phytochrome coordinates Arabidopsis shoot and root development

Authors


(fax +44(0)131 650 6556; e-mail karen.halliday@ed.ac.uk).

Summary

The phytochrome family of photoreceptors are potent regulators of plant development, affecting a broad range of responses throughout the plant life cycle, including hypocotyl elongation, leaf expansion and apical dominance. The plant hormone auxin has previously been linked to these phytochrome-mediated responses; however, these studies have not identified the molecular mechanisms that underpin such extensive phytochrome and auxin cross-talk. In this paper, we show that phytochrome regulates the emergence of lateral roots, at least partly by manipulating auxin distribution within the seedling. Thus, shoot-localized phytochrome is able to act over long distances, through manipulation of auxin, to regulate root development. This work reveals an important role for phytochrome as a coordinator of shoot and root development, and provides insights into how phytochrome is able to exert such a powerful effect on growth and development. This new link between phytochrome and auxin may go some way to explain the extensive overlap in responses mediated by these two developmental regulators.

Introduction

In plants, sophisticated signalling pathways have evolved to interconnect sensory input and developmental pathways. These molecular channels coordinate development in the natural environment where external parameters are in constant flux. Communication between the shoot and the root is particularly important as these signals influence the relative growth and development of aerial and underground structures. Key triggers for this inter-organ signalling are nutrient and water availability, stress and temperature. However, recent findings have demonstrated that light influences the developmental course of the root, providing the intriguing possibility that it acts via a long-distance signal.

In Arabidopsis, the light environment is monitored by at least three major groups of photoreceptors, comprising red/far-red (R/FR) light-absorbing phytochromes, UVA/blue light-absorbing cryptochromes, and phototropins (Chen et al., 2004). The phytochromes are well known for triggering seedling de-etiolation, but are also influential throughout the plant life cycle, controlling vegetative architecture, apical dominance and the timing of reproductive development. In Arabidopsis, the phytochromes are encoded by a small gene family, of which there are five members (PHYAPHYE) (Clack et al., 1994; Sharrock and Quail, 1989). The levels and timing of expression for individual phytochromes are subject to differential control by light and the circadian oscillator (Hall et al., 2001; Sharrock and Clack, 2002; Toth et al., 2001). Thus, the active pool of phytochrome is highly dynamic and acutely responsive to the changing light environment. Unique amongst light receptors, phytochromes exists as two photoconvertible isomers, Pr, which absorbs maximally in the red regions, and Pfr, which absorbs maximally in the far-red regions of the electromagnetic spectrum (Nagatani, 2004). Absorption of red light promotes Pr conversion to Pfr, the ‘active’ form, whilst far-red light reverses this process, converting Pfr back to the inactive Pr form of the molecule. This property of phytochrome has enabled experimental manipulation of active phytochrome levels (Franklin and Whitelam, 2004; Whitelam et al., 1998). By providing seedlings grown under white light with varying amounts of supplementary far-red light, or end-of-day far-red light, the total seedling Pfr content can be adjusted.

Recent studies have revealed that activation of phytochrome is accompanied by changes in the cellular location of the molecule (Nagatani, 2004). Upon activation by light, the phytochrome molecule undergoes a conformational change that exposes nuclear localization signals in the PAS domain and facilitates its nuclear translocation (Chen et al., 2005). This is thought to be important for phytochrome activity; indeed, red light-induced nuclear localization of phyB–GFP was shown to be reversible by far-red light (Kircher et al., 1999). Several studies have revealed that, in the nucleus, phytochrome molecules aggregate in sub-nuclear foci (speckles), whilst speckling intensity has been shown to correlate with the severity of the response (Chen et al., 2003; Kircher et al., 2002). The precise function of the sub-nuclear speckling is not yet fully understood, although it has been proposed as the site where phytochrome regulates downstream signalling events such as transcription. Several lines of evidence support this view (Ang et al., 1998; Bauer et al., 2004; Hardtke et al., 2000; Mas et al., 2000). Recent work has demonstrated that speckle formation is not essential for all phyB responses, as biological activity has been demonstrated for N-terminal phyB dimers that localize to the nucleus but do not form speckles, and for diffuse phyB–GFP nuclear staining that occurs at low light fluence rates (Chen et al., 2003; Matsushita et al., 2003).

Whilst the role of phytochromes in the shoot has been extensively studied, little attention has been directed to the role of these photoreceptors in root development. However, there is evidence for phytochrome activity within the Arabidopsis root system. Roles have been identified for phytochromes in the control of phototropism and gravitropism in roots (Correll et al., 2003; Kiss et al., 2003; Ruppel et al., 2001). However, phytochrome action in roots does not appear to be confined to the tropic responses. Work by Reed et al. (1993) demonstrated a role for phyB in the control of root hair elongation. More recently, phytochromes A, B and D have been shown to control red light-mediated elongation of the primary root (Correll and Kiss, 2005). The hy5 mutant, known to be defective in phytochrome signalling, also has a pleiotropic root phenotype, including altered lateral root production (Cluis et al., 2004; Oyama et al., 1997). HY5 has been shown to control these aspects of root growth by altering signalling through the cytokinin and auxin pathways (Cluis et al., 2004; Sibout et al., 2006). Thus, HY5 has been proposed as a signal integration point in the light and hormone signalling networks.

It is well established that auxin exerts a major influence on root growth and development. Several studies have shown that auxin is essential for lateral root production. aux1 auxin efflux carrier mutants fail to develop auxin gradients, and exhibit deficiencies in lateral root primordia production (Marchant et al., 2002). Furthermore, mutants that are defective in auxin transport, such as tir1 and tir3, produce fewer lateral roots (Ruegger et al., 1997). Recent work has demonstrated that seedling lateral root development is dependent on a shoot-derived auxin pulse (Bhalerao et al., 2002). In the young seedling, this auxin pulse appears to promote lateral root growth and augment root auxin production, which occurs predominantly in the primary and lateral root tips (Ljung et al., 2005).

In this paper, we identify an important new role for phytochrome in the coordination of shoot and root development. We present evidence for new phytochrome–auxin links that contribute to inter-organ signalling. Our findings also provide additional insights into how phytochrome is able to exert such an extensive influence on plant development.

Results

Phytochromes are expressed in roots, and form nuclear speckles in response to light

Previous studies have shown that phytochrome is expressed in roots (e.g. Toth et al., 2001); however, the spatial expression pattern of root phytochrome has not been examined in detail. To do this, we examined expression ofPHYA-, PHYB-, PHYC-, PHYD- or PHYE-promoter::LUC in seedlings grown under a 16 h photoperiod for 10 days to allow lateral root development. Under these conditions PHYC : : LUC was poorly expressed in roots PHYA, PHYB, PHYD and PHYE::LUC activity was observed throughout the root, and, with the exception of PHYB::LUC, higher bioluminescence was observed at the root tips of both primary and lateral roots (Figure 1a). Interestingly, the root bioluminescence of PHYA::LUC, PHYD::LUC and PHYE::LUC was relatively high and PHYD::LUC appeared to be highly expressed throughout the elongation zone of the primary root. The inset in Figure 1(a) shows that similar levels of PHYD::LUC expression are maintained after root tip excision, which establishes that this expression pattern was not simply an artefact of light piping. The apparent high levels of PHY::LUC expression in root tips observed by us and other workers may reflect the increased density of cells rather than a relatively high cellular expression in this area (Toth et al., 2001). Our experiments do not distinguish between these two possibilities.

Figure 1.

PHY::LUC expression and phy–GFP cellular localization in roots.
(a) PHYAE::LUC spatial expression patterns are shown for 10-day-old seedlings. The inset shows PHYD::LUC expression following excision of the root tip.
(b) Cellular location of phyA–E–GFP in roots of 7-day-old seedlings. Upper row: phyB–GFP localization patterns in cell epidermal cells of seedlings grown in darkness and 0.1 or 1 μmol m−2 sec−1 of red light, corresponding to 0, 1 and 16% Pfr, respectively, and sub-nuclear speckling in a root hair cell (10 μmol m−2 sec−1 of red light). Lower row: sub-nuclear speckling of phyA–GFP, phyC–GFP, phyD–GFP and phyE–GFP in root epidermal cells in response to far-red light (phyA) or red light (phyC–E). Dotted lines outline the nucleus.

In Arabidopsis and tobacco shoots, the dynamic properties of phytochrome within the cell provide an activity signature for the molecule. As phytochrome is expressed in roots, we wished to determine whether individual phytochromes exhibited the same light-responsive cellular dynamics in root and shoot cells. We analysed seedlings expressing fusions of GFP with phyA–E under the control of the CaMV35S promoter. In agreement with previous studies, we observed cytosolic phyB–GFP expression in dark-grown root epidermal cells (Figure 1b) (Yamaguchi et al., 1999). Furthermore, as for shoot epidermal cells, we observed diffuse nuclear staining in seedlings grown in red light at 0.1 μmol m−2 sec−1 ( approximately 1% Pfr) and nuclear speckling in seedlings grown at 1.0 μmol m−2 sec−1 (approximately 16% Pfr) (Chen et al., 2003). We noted the formation of phyB–GFP sub-nuclear foci in a variety of cell types, for example root hairs (shown in Figure 1b), epidermal and underlying cortical cells. We also observed far-red light-mediated nuclear localization and speckling of phyA–GFP, and red light regulation of these events for phyC–GFP, phyD–GFP and phyE–GFP (Figure 1b). Collectively, our data demonstrate that the phytochromes are expressed and exhibit similar light-regulated intracellular dynamics in roots as they do in shoots.

Phytochrome controls the rate of lateral root production

Whilst the role of phytochrome in the shoot has been well characterized, its role in root development has received little attention. Our analysis of seedlings null for individual phytochromes demonstrates that phytochromes participate in the regulation of lateral root production (Figure 2a). The phyD mutant has slightly enhanced lateral root production, and phyA, phyB and phyE mutants show reduced lateral root production. Mutants deficient in phyA in addition to phyB also exhibit a reduction in the rate of lateral root outgrowth (Supplementary). This indicates that phyA, phyB and phyE promote lateral root production, and, as for the shoot, phyB has the most prominent role. phyD appears to antagonize this action. We and others have also shown that reduced phytochrome activity impairs root gravitropism in primary and lateral roots (data not shown; Correll and Kiss, 2005). Thus, in addition to their well documented roles in shoot development, the phytochromes appear to play a central role in controlling root development.

Figure 2.

 Phytochrome regulation of root development in light- and dark-grown roots.
(a) Lateral emergence in phytochrome null mutant seedlings. Lateral root number in Ler (WT), phyA, phyB, phyD and phyE seedlings counted between 7 and 11 days after germination. Data represent mean values from at least 30 seedlings. SE bars are shown.
(b,c) Root phenotypes (b) and mass (c) of 5-week-old, soil-grown Arabidopsis plants. Pie charts show the root (white) expressed as a proportion of the total seedling mass.
(d) Localization of phyB–GFP in soil-grown plants with sub-nuclear speckling in guard cells of the shoot, diffuse staining in the upper root, and absence of staining in cells at the root tip.

The phyB root phenotype is retained in soil-grown plants

Recent work has shown that, when directly irradiated with red light, root elongation is inhibited (Correll and Kiss, 2005). This suggests that phytochromes can act locally within the root system to control its growth. It also raised the possibility that our observed root phenotypes were an artefact of our experimental conditions where roots grown on agar plates were exposed to light (Figure 2a). To test this, we grew our phyB seedlings in a more natural situation. When grown in a compost/sand mix, phyB seedlings produced far fewer lateral roots than the Ler wild-type (Figure 2b,c). The less bulky phyB roots had reduced total FW when compared to wild-type roots (data not shown). Furthermore, when expressed as percentage of shoot FW, root mass was proportionally lower in phyB seedlings compared with wild-type seedlings (Figure 2c). These observations demonstrate that compost-grown phyB roots are phenotypically similar to those of phyB seedlings grown on agar plates.

Root-localized phyB is not activated by axially conducted light

Previous work has demonstrated that light is conducted axially from the shoot to the root via the vascular tissue, with wavelengths in the 710–940 nm range being transmitted with the greatest efficiency (Sun et al., 2005). These findings suggest that, in some species, light can penetrate a proportion of the root system to trigger phytochrome action. We have shown that light induces nuclear localization and speckling of root phy–GFP fusion proteins (Figure 1b). Furthermore, we have demonstrated that root-localized phyB–GFP has comparable fluence rate-induced localization properties to shoot-localized phyB (Chen et al., 2003). Thus, light activation of root-localized phytochrome by conductance was a formal possibility. To establish whether this occurred in Arabidopsis, we examined the localization of phyB–GFP in the roots of soil-grown seedlings. We were able to detect nuclear-localized phyB–GFP which aggregated in sub-nuclear foci in hypocotyl and leaf epidermal cells. Figure 2(d) illustrates this in leaf guard cells. In contrast, we consistently observed diffuse nuclear phyB–GFP just below the root–shoot junction, but never sub-nuclear foci (Figure 2d). We were unable to detect nuclear-localized phyB in root cells that were more than 1 cm below the soil line. These data suggest that, for the bulk of the root system, axially conducted light is unable to trigger nuclear localization of phyB. This suggests that either phyB can act independently of light to regulate root physiology or that shoot-localized phyB exerts its control on the root through a long-distance signal.

Phytochrome coordinates root and shoot growth by controlling auxin transport and response

Recent work has provided evidence that seedling lateral root emergence is triggered by a pulse of auxin derived from the shoot (Bhalerao et al., 2002). We wished to establish whether phytochromes regulate lateral root growth by manipulating this auxin pulse. To explore this possibility, we grew seedlings expressing the auxin-responsive DR5::GUS construct in either high R:FR ratio or low R:FR ratio light, which severely depletes active phytochrome (Pfr) levels. We found that seedlings subjected to low R:FR ratio light had higher levels of DR5::GUS in shoots, particularly at the base of the hypocotyl, and lower levels in roots (Figure 3b,c). As these seedlings had more elongated hypocotyls and produced fewer lateral roots (Figure 3a,b), our data suggest that the phytochromes may control hypocotyl elongation and lateral root production either by altering seedling auxin distribution and/or the response of auxin-regulated genes. We also examined the influence of phyB on this response by assessing phyB (phyB-5) mutants expressing DR5::GUS. These seedlings exhibited alterations in DR5::GUS expression that were comparable to those of wild-type seedlings treated with low R:FR ratio light, suggesting a contributing role for phyB (Figures 3 and 4).

Figure 3.

 Lateral root production and DR5::GUS expression patterns in seedlings grown under low and high R:FR ratio light.
(a) Lateral root emergence in Ler (WT) and Col (WT) seedlings expressing DR5::GUS and grown under high or low R:FR ratio light.
(b) DR5::GUS expression in seedlings grown under either high or low R:FR ratio light.
(c) DR5::GUS expression (expressed in arbitrary units), quantified by fluorometric assay, in shoots and roots of 5-day-old seedlings.
In all experiments, seedlings were grown under 16 h photoperiods at 18°C. SE bars are shown. The difference between DR5::GUS levels in roots treated with high or low R:FR ratio light is significant (**P < 0.022, t = 2.42).

Figure 4.

 Auxin-induced DR5::GUS expression in WT and phyB seedlings.
DR5::GUS expression in Col (WT) and phyB seedlings following 24 h treatment with 0, 0.1 or 1 μm IAA or NAA applied in a strip of agar just above the shoot–root junction. Representative seedlings are shown (n = 30). The inset shows DR5::GUS expression at the root tip of untreated wild-type and phyB seedlings.

We next conducted auxin feeder experiments in which NAA (freely diffusible) or IAA (which requires auxin transport) was applied in a band of agar to wild-type or phyB shoots (Figure 4). We found that 0.1 μm NAA induced expression of DR5::GUS in both wild-type and phyB roots. The same concentration of IAA induced root DR5::GUS expression in wild-type roots, but was quite ineffective in the phyB null mutant, which exhibited only mild GUS staining at the application site and at the root tip. It is possible that the moderate levels of DR5::GUS expression at the application site may reflect reduced IAA uptake in phyB. When combined with the data from Figure 3, these data support the notion that phyB, and possibly other phytochromes, regulates auxin transport into the root system and/or reduces the response of auxin-inducible genes.

To explore the possibility that phytochrome controls auxin distribution, we conducted auxin transport assays using tritiated IAA. In our experiments, whilst we always observed small reductions in auxin transport in phyB null seedlings when compared to the wild-type, the data were not statistically significant (Figure 5). However, in phyA phyB mutants, we consistently observed significant reductions in auxin transport when compared with wild-type seedlings (Figure 5). These data suggest that the combined action of phyA and phyB controls the distribution of auxin between the shoot and root.

Figure 5.

 Auxin transport assay in WT, phyB and phyA phyB seedlings.
Auxin transport was measured in Ler (WT), phyB and phyA phyB seedlings. 0.1 μm3H-IAA was applied in a strip of agar just above the shoot–root junction. After 24 h, radioactivity (Bq) was assessed in 5 mm root tip sections. SE bars are shown. The difference between WT and phyA phyB is statistically significant (**P > 0.0045, t = 2.35). Similar results were also obtained 12 h after 3H-IAA application (data not shown).

IAA1 and IAA3 transcript levels are altered in phyAphyB shoots and roots

To further test the hypothesis that phytochrome moderates auxin transport, we measured the mRNA levels of genes known to be involved in auxin response. We focused on IAA1 and IAA3/SHY2 as these genes have been previously shown to be phytochrome-regulated (Devlin et al., 2003; Tian et al., 2003). In these experiments, we compared transcript levels in shoot and root tissue. Figure 6 illustrates that, when compared to wild-type, IAA1 and IAA3 are expressed at higher levels in phyA phyB seedling shoots. In wild-type root tissue, although we detected reduced expression in roots, IAA1 and IAA3 transcript levels were moderately lower in phyA phyB roots. These data demonstrate that loss of phyB and phyA action enhances shoot IAA1 and IAA3 mRNA and leads to a modest reduction of IAA1 and IAA3 transcripts in the root. When combined with our auxin feeder data, these results support a role for phyA and phyB in reducing auxin transport or availability in hypocotyl tissue, and enhancing the transport of shoot-derived auxin into the root.

Figure 6.

 Auxin-responsive gene expression in phyA phyB mutant seedling shoots and roots.
Quantification of IAA1 and IAA3 mRNA levels in Ler (WT) and phyA phyB shoots and roots. Relative transcript levels were determined by real-time quantitative PCR and normalized using AtACT7 mRNA as a reference. SE bars are shown. The differences between IAA levels in WT and phyA phyB roots was significant for IAA1 (*P > 0.027, t = 2.13) and IAA3 (*P > 0.034, t = 2.13).

Recent microarray experiments have suggested that expression of the auxin efflux genes PIN3 and PIN7 is controlled by phytochrome (Devlin et al., 2003). We have shown that PIN3 and PIN7 mRNA levels are elevated in phyB versus wild-type seedlings (Supplementary). This provides the possibility that phytochrome controls IAA1/IAA3 transcription, at least in part, by regulating PIN3/PIN7 levels. Interestingly, the levels of PIN1 mRNA are similar in phyB and wild-type seedlings (data not shown), indicating that PIN1, a major regulator of shoot-to-root auxin transport (see Friml and Palme, 2002), is not subject to control by phyB at the transcriptional level.

Discussion

Phytochrome and auxin are potent regulators of plant development. Light regulates a broad range of responses which include hypocotyl elongation, leaf expansion and phototropism, responses that are also controlled by auxin. It has been known for a number of years that phytochrome and auxin signalling are connected, yet, until now, no firm molecular link has been established. In this paper, we have shown that phytochrome controls growth and development, at least in part, by regulating auxin transport. This appears to enable phytochrome to coordinate shoot and root development. It is presumably through this mechanism that shoot and root development can be reciprocally adjusted to take full advantage of the available light resources.

phyB has a prominent role in controlling the rate of lateral root production

In early seedling development, auxin synthesized in the first leaves is transported to the root where it promotes the outgrowth of lateral roots (Bhalerao et al., 2002). Perturbation of auxin transport or signalling can have dramatic effects on lateral root production (Marchant et al., 2002; Santelia et al., 2005). We have shown that phytochromes are important components of this regulatory process, as loss of phyA and/or phyB, of phyE or reduction of the Pfr:Pr ratio by low R:FR ratio light, reduces the rate at which lateral roots emerge in young seedlings. Of the monogenic phy null mutants, this phenotype is most severe in phyB, indicating a prominent role for phyB in this response.

Shoot-localized phytochrome controls root physiology

Previous work has shown that roots exposed to red light exhibit increased inhibition of primary root elongation (Correll and Kiss, 2005). It was therefore possible that the root phenotypes of our agar plate-grown seedlings were caused, at least partly, by local phytochrome action within the root. However, we demonstrated that the phyB lateral root phenotype was retained in compost-grown seedlings. These findings support the notion that this response was controlled from the shoot.

Recent work has demonstrated that, in some plant species, light can penetrate the root through vascular conductance. We have shown that root-localized phytochrome exhibits comparable light-regulated localization characteristics to shoot phytochrome. However, we only ever observed diffuse nuclear localization, and never speckling, of phyB–GFP in cells located just below the root–shoot junction. We were unable to detect nuclear phyB–GFP in root cells more than 1 cm below the soil line. This illustrates that, under our light fluence rates (100 μmol m−2 sec−1), phyB activation by axially conducted light is unlikely to play a significant role in shaping root development. Combined, our data provide strong support for a role of phyB and other phytochromes in modulating root growth in response to light signals perceived in the shoot. Our experiments do not eliminate the possibility that cytosolic-localized Pr plays a role in controlling this response. However, in experiments where seedlings grown on sucrose plates were transferred to darkness on day 3, prior to the early shoot–root auxin pulse, lateral root production is arrested (data not shown). This reinforces the importance of the shoot-derived auxin pulse for the lateral root phenotype. It does not preclude the action of Pr in roots following delivery of the auxin to the root, but does suggest that primary control of this response is achieved by manipulating the auxin pulse.

Phytochrome adjusts the distribution of auxin in the shoot and root

Several reports have shown that phytochrome regulates a subset of auxin-responsive genes, including IAA1, IAA3/SHY2 (Devlin et al., 2003; Tian et al., 2002) and components of the complex auxin transport machinery including MDR1/PGP11, PGP1, PIN3 and PIN7 (Devlin et al., 2003; Lin and Wang, 2005; Sidler et al., 1998). We have shown that subjecting seedlings to low R:FR ratio light, which reduces active phytochrome (Pfr) levels, induces a change in the pattern of DR5::GUS expression. These seedlings have enhanced DR5::GUS expression in the lower third of the hypocotyl and reduced expression in the roots. Furthermore, phyA phyB null seedlings have higher levels of shoot IAA1 and IAA3 mRNA, whilst phyA phyB roots have slightly reduced levels. DR5::GUS has been shown to be regulated by brassinolide in addition to auxin, so this may reflect altered regulation of one or both these pathways in seedlings exposed to low R:FR ratio light (see Halliday, 2004). However, auxin transport in hypocotyls has been shown to be phytochrome-dependent, and mutants that are null for phyA and phyB have reduced temperature-mediated hypocotyl extension (Jensen et al., 1998; Mazzella et al., 2000). Furthermore, shoot-derived auxin has been shown to be important for lateral root production, a response that is perturbed by treatment with low R:FR ratio light (Bhalerao et al., 2002; Marchant et al., 2002). Thus, our findings most likely reflect a role for phytochrome in regulating auxin distribution. IAA feeder experiments provide further support for this proposition. Our auxin transport assays using tritiated IAA demonstrated that depletion of both phyA and phyB attenuates shoot–root auxin transport, whilst depletion of phyB alone does not. Interestingly, the lateral root phenotypes of phyB and phyA phyB are similar, thus it is possible that our tritiated IAA transport assays were not sufficiently sensitive to detect subtle differences in auxin status. Indeed, auxin feeder experiments using shoot-applied IAA were less effective at triggering DR5::GUS expression in phyB roots than in wild-type roots. It is therefore possible that DR5::GUS, which comprises eight tandem copies of the 11 bp natural AuxRE from the GH3 promoter, may be a more sensitive indicator of auxin distribution (Ulmasov et al., 1997). Alternatively, phyB may also control lateral root production by an alternative mechanism. Collectively, our results suggest that phytochrome effects at least a subset of its responses by manipulating seedling auxin distribution. This phytochrome–auxin link may account for a proportion of the extensive overlap between phytochrome- and auxin-regulated responses.

What is the role of root-localized phytochrome?

Recent work has provided insights into roles for root-localized phytochrome in regulating phototropic responses and primary root elongation growth (Correll and Kiss, 2005; Correll et al., 2003). In support of these findings, we have shown that phytochromes A, D and E are highly expressed in primary and lateral root tips, and that phyD is expressed throughout the elongation zone of the primary root. Furthermore, we have demonstrated that root-localized phy–GFP shows light-regulated cellular localization characteristics similar to those described for shoot phytochromes. Indeed, the role of phytochrome in the root appears to be similar to that in the shoot where it also regulates cell elongation and phototropism. However, in contrast to the shoot, phytochrome acts as a phototropic receptor in roots. Given the role of shoot phytochrome in shaping root growth via a long distance signal, it will be interesting to establish whether root-localized phytochrome signals to the shoot.

Possible mechanisms of phytochrome–auxin cross-talk

This paper has demonstrated that the phytochromes act collectively to control lateral root outgrowth. The phytochromes appear to regulate this response at least partly by moderating the shoot–root auxin pulse in early seedling development. These findings provide a means to coordinate shoot and root development in response to the external light environment. Interestingly, a recent study provides support for a role of the cryptochromes in regulating aspects of root growth by modulating auxin transport (Canamero et al., 2006). This may therefore be a common control mechanism for both the phytochromes and cryptochromes.

The extensive overlap between phytochrome- and auxin-induced responses suggests their tight linkage in the signalling network. At present, we do not know precisely how the phytochromes may be acting to control auxin transport; however, our work and recent studies provide some leads. In line with previous microarray studies, we have shown that transcript levels of the auxin efflux effectors, PIN3 and PIN7, are phytochrome-regulated (Devlin et al., 2003). Aspects of the pin3 mutant phenotype are reported to be light-specific (Friml et al., 2002); furthermore, we have shown that pin3 mutants have accelerated lateral root production (Supplementary). Therefore, phytochrome may control auxin transport by altering the levels and/or the cellular location of PIN proteins. Recent work has implicated the p-glycoproteins PGP1 and MDR1/PGP11 in auxin transport; indeed, PGP1 has been shown to directly catalyse the active efflux of IAA (Geisler et al., 2005). PGP1 and MDR1/PGP11 have previously been shown to be regulated by phytochrome, and mdr1 and pgp1 loss-of-function mutants have light-specific hypocotyl phenotypes (Lin and Wang, 2005; Sidler et al., 1998). Furthermore, MDR1 has been shown to operate in a phyA pathway in seedlings grown under far-red light (Lin and Wang, 2005). These observations suggest that MDR1/PGP1 activity is tightly coupled to light signalling, and they are therefore good candidates for phytochrome action in the manipulation of shoot–root auxin distribution.

It is unlikely that phytochrome control of auxin signalling is confined to manipulation of auxin gradients. We and others have shown that phyB null mutants have elongated root hairs (data not shown; Reed et al., 1993). This is the antithesis of the expected phenotype for seedlings with low root auxin levels (Pitts et al., 1998); instead, phyB root hairs are suggestive of enhanced auxin signalling. Interestingly, this is a phenotypic characteristic of hy5, a known phytochrome signalling component (Oyama et al., 1997). HY5 has been shown to regulate transcription by binding to the core G-box sequence CCACGTG (Ang et al., 1998). This binding site is contained within the promoters of the SLR/IAA14/IAA28 and AXR2/IAA7 genes that exhibit reduced transcript levels in hy5 mutants (Cluis et al., 2004). Furthermore, HY5 has been shown to bind the promoter of AXR2 in vitro. Thus, HY5 appears to regulate auxin signalling by targeting negative regulators of the pathway. It is therefore possible that phyB moderates auxin signalling through the HY5 pathway. It is also feasible that phyB regulates AUX/IAA activity by direct interaction. Pull-down assays identified an interaction between phyB and SHY2/IAA3, a known auxin and phyB signalling component (Tian et al., 2003). Further analysis is required to establish whether this is occurs in vivo.

The data in this paper provide insights into how phytochrome coordinates shoot and root development. It appears that phytochrome achieves this, at least in part, by reciprocally regulating auxin gradients in aerial and root structures. It has been known for a considerable time that phytochrome controls multiple developmental processes, yet the molecular basis of this control is not known. It may be through the manipulation of auxin that phytochrome is able to influence such a range of responses.

Experimental procedures

Plant materials and growth conditions

All studies were carried out using the Arabidopsis thaliana Landsberg erecta (Ler) or Columbia (Col) accessions. The phytochrome mutant alleles used in this study were phyA-201 (Nagatani et al., 1993), phyB-1 (Reed et al., 1993), and phyB-5, phyD-1 (Aukerman et al., 1997) and phyE-1 (Devlin et al., 1998), all in the Ler background (Neff and Chory, 1998) in the Col background. Double mutants were created by genetic crossing as described previously (Devlin et al., 1998, 1999). We used the PHYAE promoter::LUC lines and the lines expressing translational fusions of GFP with phyA–phyE, as previously described (Kircher et al., 2002; Toth et al., 2001). The DR5::GUS line, containing a synthetic auxin-responsive promoter fused to the GUS reporter gene, in the Col genetic background, has been described previously by Ulmasov et al. (1997).

For all experiments, seeds were surface-sterilized in 20% v/v bleach for 5 min. After three washes in distilled water, seeds were sown on plates containing Hoaglands No. 2 basal salts medium (Sigma-Aldrich; http://www.sigmaaldrich.com/), pH 5.7, 1% w/v sucrose and 0.5% w/v Phytagel (Sigma-Aldrich). Seeds were stratified in complete darkness for 4 days at 4°C before transfer to specific growth conditions. Plates were positioned vertically to allow root growth along the gel surface. For plant growth, we used plant growth cabinets (Snijders Scientific, http://www.snijders-tilburg.nl) with a 16 h photoperiod (fluence rate of 100 μmol m−2 sec−1) and a temperature of 18 ± 0.5°C. Experiments with high and low R:FR ratio light were performed in climate-controlled growth rooms, also set to 16 h photoperiods at 18 ± 0.5°C, with a photon fluence rate of 70 μmol m−2 sec−1. Supplementary FR light was supplied during the photoperiod by light-emitting diode arrays to create a low R:FR ratio of 0.126. Light quantity and quality were measured using a StellarNet EPP2000 spectroradiometer (Astranet Systems, http://www.astranetsystems.com).

Physiological analysis

Lateral roots were viewed with a stereomicroscope (MZFLIII, Leica Microsystems, http://www.leica-microsystems.com) and were counted daily between 7 and 11 days after transfer to light, following stratification for 4 days. Assessment of root mass in soil-grown seedlings was performed on 5-week-old plants grown in climate-controlled growth rooms under white light (100 μmol m−2 sec−1) with 16 h photoperiods at 20°C. Seeds were sown directly on a 50:50 compost/sand mixture, and stratified for 4 days at 4°C in darkness before transfer to growth conditions.

Reporter gene analysis

Seedlings expressing PHYA–E::promoter::LUC were sprayed with 5 mm luciferin 5 min before analysis of the LUC expression pattern using an intensified CCD camera (Hamamatsu VIM, http://www.hamamatsu.com). Images were processed using Image J (NIHimage, http://www.rsb.info.nih.gov/nih-image/).

The cellular location and characteristics of 35S::phyA–E–GFP were assessed using an Eclipse confocal microscope (Nikon, http://www.nikon.co.uk). Colour was artificially added using photoshop 8 (Adobe Systems, http://www.adobe.com).

For histochemical analysis of GUS activity, Arabidopsis seedlings were incubated overnight at 37°C in GUS reaction buffer (0.5 mm 5-bromo-4-chloro-3-indolyl-β-d-glucoronic acid in 100 mm sodium phosphate, pH 7). Stained seedlings were cleared with 70% ethanol overnight before being mounted in glycerol and viewed under a DMLB stereomicroscope (Leica Microsystems). Representative seedlings were photographed using a Coolpix 4200 digital camera (Nikon).

GUS staining was quantified by analysis of MUG fluorescence. Seedlings were severed at the shoot–root junction. Shoot and root tissues were incubated in GUS extraction buffer [1 mm MUG (Sigma-Aldrich), 50 mm sodium phosphate, pH 7, 10 mm EDTA, 0.1% SDS, 0.1% Triton X-100] at 37°C for 6 h, before the reaction was stopped with 1 m sodium carbonate. MUG fluorescence was measured on a Fluorolite 1000 fluorometer (Dynatech Laboratories, http://www.dynatechlaboratories.com).

Auxin feeder experiments

To assess auxin transport, a 1% agar band containing 0, 0.1 or 1.0 μm IAA or NAA was applied just above the shoot–root junction of intact, light-grown seedlings expressing DR5::GUS in the Col or phyB-9 backgrounds. Seedlings were treated for 24 h before staining as described above. Comparable results were obtained in three replicate experiments using 15–20 seedlings.

In auxin transport assays using labelled IAA, 0.1 μm3H-IAA was applied in an agar band as described above. After 24 h, 5 mm sections of the root tip were excised and rinsed twice in distilled water to remove residual activity, before being soaked in 80 μl of 80% methanol for 60–90 min. Following this, 920 μl of sterile distilled water and 10 ml of Optiscint HighSafe scintillation fluid (PerkinElmer, http://www.perkinelmer.com) were added, and the radioactivity of samples was counted using a Beckman LS6500 scintillation counter (Beckman Coulter, http://www.beckmancoulter.com). Comparable results were obtained in three replicate experiments using 8–10 seedlings.

Quantitative PCR

Plant total RNA was extracted from 7-day-old, white-light-grown seedlings using RNeasy Plant Mini kits (Qiagen; http://www.qiagen.com/), according to the manufacturer’s instructions, including a DNase I treatment step of column-bound RNA (Qiagen). cDNA was then synthesized using a RevertAid first-strand synthesis kit (Fermentas, http://www.fermentas.com). Quantitative PCR was performed in a Rotor-Gene 3000 (Corbett Life Science, http://www.corbettlifescience.com) using SYBR Green Jump-Start Taq ready mix (Sigma-Aldrich). The primers used were: ACT7for (5′-CAGTGTCTGGATCGGAGGAT-3′), ACT7rev (5′-TGAACAATCGATGGACCTGA-3′), IAA1for (5′-ACATGTTCAAGTTCACAGTA-3′), IAA1rev (5′-TGCCTCGACCAAAAGGTGT-3′), IAA3for (5′-CTGTGGGAGAGTACTTTGAG-3′), IAA3rev (5′-CATATGAACATCTCCCATGGA-3′), PIN3for (5′-CGAGACCAAAGCTGTAGCTC-3′), PIN3rev (5′-GTTTAGACCATTCTCGGCGT-3′), PIN7for (5′-TTGCTTTCAGGTGGGATGTG-3′) and PIN7rev (5′-ACTCACCCAAACTGAACATTGC-3′). Experiments were performed in triplicate on two biological replicates.

Acknowledgements

We thank Professor Ottoline Leyser (University of York, UK) for supplying seeds containing the DR5::GUS construct, and Professors Ferenc Nagy (Biological Research Center of the Hungarian Academy of Sciences, Szeged, Hungary) and Andrew Millar (University of Edinburgh, UK) for supplying seeds of the PHYA–E::LUC and phyA–E–GFP transgenic lines.

Ancillary