The most significant developments in shade avoidance research over the past decade have undoubtedly resulted from the identification of multiple R : FR ratio-regulated genes and elucidation of molecular mechanisms by which phytochromes control their expression. Early observations reporting stem elongation to occur within 10–15 min of FR supplementation appeared to preclude the involvement of gene expression changes in the initiation of these responses (Morgan et al., 1980). Quantitative kinetic analyses of low R : FR-mediated gene expression later revealed similar rapidity, presenting an exciting opportunity to investigate the molecular mechanisms controlling these responses (Salter et al., 2003). The combined research efforts of both the photomorphogenesis and gibberellin signalling fields have recently converged, providing a more holistic understanding of how plants regulate elongation growth at the molecular level. In this review, I present a (simplified) hypothetical model of how plants may initiate elongation growth following perception of low R : FR (Fig. 3). The mechanisms presented are based on evidence from published literature and are described in detail throughout this section.
Figure 3. Hypothetical model depicting molecular mechanisms controlling red to far red ratio (R : FR)-mediated elongation growth. (a) In daylight, phyB exists predominantly in the Pfr form (Pfr B). Phytochrome A Pfr (Pfr A) is degraded in both the cytosol and the nucleus. Following import into the nucleus, Pfr B binds PIF4 (and PIF5 proteins), resulting in their degradation via the 26S proteosome. The binding of DELLA and PIF proteins simultaneously results in PIF inactivation. An enhancement of DELLA activity relative to PIF4 activity results in the suppression of genes involved in elongation growth. (b) In vegetational shade, a reduction in R : FR results in conversion of phyB Pfr to the inactive Pr form, which dissociates from PIFs and exits the nucleus, increasing PIF stability. Antagonism of shade avoidance is provided by phyA operating in the high irradiance response (HIR) mode. In low R : FR, a small pool of cycled phyA is protected from degradation and binds some PIF4 protein, targeting it for degradation. The overall increase in PIF4 activity results in increased expression of genes involved in elongation growth. These include genes involved in the biosynthesis of gibberellin (GA). (c) In prolonged vegetational shade, GA concentrations may increase, thus enhancing DELLA degradation. Increased DELLA degradation would further enhance PIF stability, resulting in prolonged elongation growth.
Download figure to PowerPoint
1. R : FR ratio-regulated genes
Research efforts to identify R : FR ratio-regulated genes progressed considerably following the development of microarray technology (Devlin et al., 2003; Salter et al., 2003; Sessa et al., 2005). Before this, the only genes reported to be reversibly regulated by changes in R : FR were the transcription factors ATHB2 (formerly HAT4) and ATHB4 (Carabelli et al., 1993, 1996). Both are members of the homeodomain leucine zipper (HDZip) family of transcription factors and interact with the DNA binding site CAATNATTG (Henriksson et al., 2005). Transcript abundance of ATHB2 was shown to increase rapidly following low R : FR and EOD-FR treatments (Carabelli et al., 1996). This response was partially retained in phyB mutants, suggesting the involvement of additional phytochromes (Carabelli et al., 1996). Subsequent analysis of multiple phytochrome-deficient mutants revealed phyE to act redundantly with phyB to suppress ATHB2 expression (Franklin et al., 2003). A role for ATHB2 in the regulation of shade avoidance was proposed following analyses of transgenic plants, expressing altered abundances of ATHB2 transcript (Schena et al., 1993; Steindler et al., 1999). Plants expressing elevated abundances of ATHB2 displayed reduced apical dominance and earlier flowering, phenotypes reminiscent of the shade avoidance syndrome (Schena et al., 1993). A reciprocal phenotype was observed in plants expressing an antisense ATHB2 construct, which displayed shorter stature and a reduced developmental rate (Schena et al., 1993). Plants expressing elevated abundances of ATHB2 were subsequently observed to display reduced growth of secondary tissues and decreased lateral root formation, phenotypes associated with altered auxin transport (Steindler et al., 1999). Such observations suggested the involvement of auxin in shade avoidance signalling, a notion which has gained increased credence in recent years (see Section V).
Microarray-based investigations have enabled additional genes displaying robust R : FR reversibility to be identified. In a study by the Whitelam laboratory, transcripts eroding the bHLH transcription factors PIL1 (PIF3 LIKE1) and a close homologue, PIL2, were shown to display rapid up-regulation by low R : FR treatment (Salter et al., 2003). These proteins were named based on their close sequence similarity to a group of phytochrome-binding bHLH transcription factors termed PHYTOCHROME INTERACTING FACTORs (PIFs) (see later). Differences in the responsivity of PIL1 and PIL2 to low R : FR were recorded at different times of day, suggesting the possible involvement of the circadian clock in their regulation. In a series of free-running experiments, the authors established circadian gating of both gene expression and hypocotyl elongation responses to low R : FR. Intriguingly, plants expressing altered abundances of PIL1 displayed a phase shift in the timing of maximal hypocotyl elongation response. Further connections between PIL1 and the circadian clock are provided by the reported physical interaction of PIL1 with the central oscillator component TOC1 (Makino et al., 2002) and observations that pil1 null mutants display reduced light-regulated expression of the putative clock component PSEUDO RESPONSE REGULATOR 9 (PPR9) (Khanna et al., 2006). A confusing observation from the Salter et al. study involved the relative timings of different low R : FR ratio-mediated responses. Gene expression responses to low R : FR were observed to occur maximally post-subjective dawn, whereas hypocotyl elongation displayed maximum responsivity pre-subjective dusk (Salter et al., 2003). Elongation responses to R : FR therefore mapped on to the period of maximum seedling elongation observed in free-running experiments (Dowson-Day & Millar, 1999), yet were at variance with the timing of gene expression responses. This apparent temporal paradox has recently been addressed in a seminal study by Nozue et al. (2007) and is discussed later in this section.
Global analysis of low R : FR ratio-mediated gene expression has additionally revealed a number of genes associated with cell wall loosening and auxin signalling to be up-regulated in low R : FR, including a number of the indole-3-acetic acid (IAA) family of auxin-regulated transcription factors (Devlin et al., 2003). In a later study, the bHLH transcription factor HFR1, the auxin responsive gene IAA29 and the xyloglucan endotransglycosylase XTH15 were shown to display significant up-regulation following simulated shade treatment (Sessa et al., 2005). Together, these results present further support for the involvement of auxin in low R : FR signalling and provide potential candidates for the regulation of cell wall metabolism during elongation responses. The latter study additionally reports enhanced shade avoidance responses in hfr1 null mutants. Based on these observations, the authors suggest the existence of a HFR1-mediated negative feedback loop controlling the magnitude of shade avoidance responses (Sessa et al., 2005). Given the well documented role of HFR1 in phyA signalling (Fairchild et al., 2000; Fankhauser & Chory, 2000; Soh et al., 2000), the possibility exists that reported phenotypes may represent a defect in the phyA-mediated ‘antagonism’ of shade avoidance (see Section III).
2. Phytochrome interacting factors (PIFs)
Following photoconversion to the active Pfr form, phytochromes undergo translocation to the nucleus (Sakamoto & Nagatani, 1996; Kircher et al., 1999, 2002). Nuclear-localized Pfr binds directly to a family of bHLH transcription factors termed PIFs (reviewed in Duek & Fankhauser, 2005; Monte et al., 2007). The first PIF protein (PIF3) was identified by yeast-two-hybrid screening and shown to display photoreversible binding to phyA and phyB Pfr in vivo (Ni et al., 1998, 1999). A number of closely related PIFs have since been identified which, together with PIF3, perform varied and overlapping regulatory roles in seedling de-etiolation (Halliday et al., 1999; Huq & Quail, 2002; Kim et al., 2003; Huq et al., 2004; Khanna et al., 2004, 2007; Monte et al., 2004; Leivar et al., 2008). Binding of phyB and PIFs has been shown to occur via a conserved APB (Active Phytochrome Binding) motif present in PIF proteins (Khanna et al., 2004). In a series of de-etiolation studies, binding of photoactivated phytochromes to PIFs has been shown to result in rapid PIF phosphorylation and degradation via the 26S proteosome (Bauer et al., 2004; Park et al., 2004; Shen et al., 2005, 2007; Al-sady et al., 2006; Lorrain et al., 2008).
The importance of PIF stability in shade avoidance has been highlighted in a pivotal recent study by the Fankhauser and Whitelam labs (Lorrain et al., 2008). In this work, the authors demonstrate rapidly increased stability of PIF4 and PIF5 proteins following transfer of de-etiolated plants to low R : FR. Immunoprecipitation assays have shown PIF4 to display strong binding to phyB with weaker binding to phyA (Huq & Quail, 2002). By contrast, PIF5 has only been demonstrated to bind phyB (Shen et al., 2007). This observation does not accord with mutant analyses, showing a clear requirement for phyA in PIF5 degradation (Shen et al., 2007). The authors speculate that either a weak phyA/PIF5 interaction exists in planta, which is unstable in the assay conditions used, or phyA may mediate PIF5 turnover through one or more intermediary factors (Shen et al., 2007). The analyses of null mutants and transgenic plants displaying altered abundances of PIF4 and PIF5 have revealed significant roles for these proteins in mediating shade avoidance responses (Lorrain et al., 2008). Monogenic pif4 mutants displayed short hypocotyls in both high and low R : FR, conditions in which pif5 monogenic mutants resembled WT plants. Double mutants, deficient in PIF4 and PIF5, however, displayed reduced elongation of both hypocotyls and petioles in high and low R : FR conditions. By contrast, plants overexpressing PIF5 were considerably elongated and displayed a ‘constitutive shade avoidance’ phenotype in high R : FR. Similar phenotypes were observed in plants overexpressing PIF4, but difficulties in line maintenance led the authors to focus on lines overexpressing PIF5 (Lorrain et al., 2008). Parallel analyses of gene expression responses revealed a similar pattern. All mutants (pif4, pif5, pif4pif5) displayed reduced expression of shade avoidance marker genes in low R : FR, whilst plants overexpressing PIF5 displayed constitutively elevated amounts of these genes in high R : FR. Transgenic plants expressing PIF4 and PIF5 constructs with a deleted APB motif were additionally investigated in this study. The deletion of this motif prevents phy/PIF interaction (Khanna et al., 2004) and resulted in exacerbation of the ‘constitutive shade avoidance’ phenotype (Lorrain et al., 2008). The authors conclude that PIF4 and PIF5 promote the expression of genes mediating elongation growth. In high R : FR, PIF proteins are degraded, following binding of photoactivated phytochrome. In low R : FR, the photoconversion and subsequent nuclear export of phyB Pfr results in enhanced stability of PIF4 and PIF5, thereby increasing the expression of genes mediating elongation growth (Lorrain et al., 2008). This model suggests that R : FR ratio signalling operates, in part, through the targeted degradation of transcriptional regulators, an emerging regulatory mechanism already established in auxin and gibberellin (GA) signalling (reviewed in Huq, 2006). Further support for this model is provided by observations that pif4 and pif5 mutations can partially restore the constitutively elongated phenotype of a phyB mutant (Lorrain et al., 2008).
Transfer of etiolated seedlings to red light has recently been shown to result in some degradation of the phyB photoreceptor (Khanna et al., 2007). This effect is marginally reduced in pif5 mutants and considerably exacerbated in PIF5 overexpressing lines. The authors suggest that PIF5 may display functional duality, acting as both transcriptional regulator and mediator of photoreceptor abundance. Observations of pronounced elongation phenotypes in plants expressing elevated amounts of PIF4 and PIF5 proteins with a deleted phytochrome-binding APB motif suggests these PIFs can promote shade avoidance independently of phytochrome interaction (Lorrain et al., 2008). Such observations are at variance with the suggestion that PIF5 promotes elongation growth through reducing phyB abundance, and support the notion that phyB acts to target PIF5 for degradation. An alternative explanation for the observed results may be that the light-regulated turnover of PIF5 results in some unavoidable loss of phyB photoreceptor, as a result of incomplete disassociation of the complex, before degradation. As both proteins display light-mediated proteolysis (Khanna et al., 2007; Lorrain et al., 2008), a circular argument could be initiated to explain the functional significance of this interaction in the regulation of elongation growth. Comparative analysis of PIF5 and PHYB protein amounts in different R : FR ratios would therefore be of interest in assessing the functional significance of this interaction in shade avoidance.
The importance of PIF4 and PIF5 in elongation growth is additionally highlighted in recent work from the Maloof and Fankhauser labs (Nozue et al., 2007). The authors demonstrate that in light : dark cycles, the hypocotyl growth of Arabidopsis seedlings is regulated through an external coincidence model, integrating both the circadian regulation of PIF4 and PIF5 transcript abundance and light-mediated proteolysis of PIF4 and PIF5 proteins. Under these conditions, transcript abundance of PIF4 and PIF5 is high from the end of the dark period and peaks in the middle of the day. In agreement with other studies (Shen et al., 2007; Lorrain et al., 2008), the authors show rapid light-mediated degradation of both proteins. The coincidence of high PIF4 and PIF5 transcript abundances and enhanced protein stability at the end of the dark period promotes maximum elongation growth during this time. This observation is in contrast to studies in continuous light, whereby maximum hypocotyl growth occurred at subjective dusk (Dowson-Day & Millar, 1999). Double pif4pif5 mutants displayed loss of the dawn growth peak, confirming their importance in the regulation of this response (Nozue et al., 2007). Furthermore, an advancement of the growth peak was observed in 16 h photoperiods when compared with experiments in 8 h photoperiods (Nozue et al., 2007). In these conditions, maximum hypocotyl growth was observed at approx. 4 h post-dawn, thereby coinciding with maximum low R : FR ratio-induced gene expression (Lorrain et al., 2008). It would therefore appear that previous temporal inconsistencies between low R : FR ratio-mediated gene expression and elongation growth have arisen as a result of the analysis of plants in continuous light (Salter et al., 2003). It is likely that in light : dark cycles, maximum elongation growth in response to transient low R : FR occurs in the early morning, thereby coinciding with maximum gene expression responses. Moreover, the circadian regulation of PIF4 and PIF5 transcript abundances may ultimately provide a molecular mechanism for the gating of rapid shade avoidance.
3. DELLA proteins
The retention of shade avoidance responses in pif4pif5 double mutants provides evidence for the existence of multiple regulatory mechanisms (Lorrain et al., 2008). Recent studies suggest that GA and phytochrome-mediated signalling pathways may converge to regulate elongation growth. The DELLA family of growth repressing proteins are mediators of the GA signalling pathway (Peng et al., 1997; Silverstone et al., 1998; Dill & Sun, 2001). In Arabidopsis, five DELLA homologues exist: GAI, RGA, RGL1, RGL2 and RGL3 (reviewed in Hussain & Peng, 2003). The relief-of-restraint model proposed by Harberd (2003) suggests that DELLAs act to restrain plant growth whilst GAs promote growth through overcoming DELLA-mediated growth restraint. It is now known that GAs regulate turnover of DELLAs via binding of a receptor, GID1 (Ueguchi-Tanaka et al., 2005; Griffiths et al., 2006; Nakajima et al., 2006; Willige et al., 2007). Interaction of DELLAs with GA-bound GID1 promotes interaction with the F-box protein SLEEPY1 (SLY1). This interaction results in polyubiquitination and degradation of DELLAs via the 26S proteosome, thereby relieving inhibition of growth (McGinnis et al., 2003; Dill et al., 2004).
Numerous reports exist of interactions between light and GA signalling (López-Juez et al., 1995; Reed et al., 1996; Pierik et al., 2004a). More recently, Pierik and colleagues have provided evidence supporting a central role for DELLA proteins in mediating shade avoidance responses (Djakovic-Petrovic et al., 2007). Confocal imaging of transgenic Arabidopsis plants expressing a pRGA:GFP-RGA construct revealed increased DELLA degradation in canopy-grown plants. As expected, these plants displayed greater hypocotyl and petiole elongation than isolated control plants. The identity of the canopy signal regulating DELLA stability was investigated through analyses of plants subject to low R : FR and blue light depletion. Increased DELLA degradation was observed in low R : FR-treated hypocotyls and petioles, confirming the involvement of phytochrome in these responses. Similar degradation was also observed in hypocotyls subject to blue light depletion, suggesting interplay between multiple photoreceptors. The requirement for GA in hypocotyl and petiole growth responses was confirmed by analyses of GA-deficient mutants. These plants displayed short hypocotyls which lacked responsivity to both low R : FR and blue light depletion. All phenotypes were rescued by GA application. A similar result was also observed in petioles treated with low R : FR. Quadruple mutants deficient in four DELLAs (gai/rga/rgl1/rgl2) did not, however, display elongated petioles, when compared with wild-type controls. Hypocotyls of these mutants were elongated, but retained responsivity to both low R : FR and blue light depletion. These data suggest that low R : FR-mediated DELLA degradation is not sufficient to induce shade avoidance responses in petioles, but, together with other signalling mechanisms, performs a more central role in hypocotyl growth (Djakovic-Petrovic et al., 2007).
The importance of light in the regulation of DELLA stability is supported by a parallel study from the Harberd lab (Achard et al., 2007). In this work, the authors show increased GFP-RGA stability in etiolated hypocotyls transferred to light. Comparison of hypocotyl length ratios (wild-type/quadruple DELLA mutant) revealed an increased contribution of DELLAs to growth restraint with increasing photoperiods. Conversely, GA-deficient plants displayed short hypocotyls in the dark, but wild-type hypocotyl lengths at longer photoperiods. Interestingly, transcript abundance of enzymes involved in GA metabolism showed rapid up-regulation upon transfer to light (Achard et al., 2007). The authors conclude that light-mediated growth restraint during photomorphogenesis involves the stabilization of DELLA proteins. Further stabilization may be achieved through light-regulated increases in GA metabolism. This notion is supported by observations in pea, showing decreased bioactive GA1 concentrations upon transfer of etiolated seedlings to light (Gil & García-Martínez, 2000; O’Neill et al., 2000).
4. PIF/DELLA interaction – a molecular mechanism for R : FR ratio signalling?
A molecular mechanism integrating light and GA signalling has recently been provided by parallel studies from the Deng and Prát labs (De Lucas et al., 2008; Feng et al., 2008). Both report the physical interaction of PIF and DELLA proteins through yeast-two hybrid and immunoprecipitation assays. The former study focuses on PIF3 and the latter, PIF4. Deletion studies confirmed the PIF4/RGA interaction to require the bHLH DNA recognition domain in PIF4 and a conserved leucine repeat motif in the RGA protein (De Lucas et al., 2008). The physiological significance of PIF/DELLA binding is elegantly demonstrated in both studies through chromatin immunoprecipitation (ChIP) experiments. Here, the physical interaction between PIFs and DELLAs was shown to block transcriptional activity, thereby inhibiting PIF function. These studies highlight PIFs as an integration node for both light and GA signalling. The authors propose that GA acts to target DELLAs for degradation, thereby relieving the inhibition of PIF activity and promoting elongation growth (De Lucas et al., 2008; Feng et al., 2008).
A hypothetical model for the R : FR ratio-mediated regulation of elongation growth is presented in Fig. 3. In daylight (Fig. 3a), photoactivation of phyB (and likely phyD and phyE) results in nuclear import of Pfr and binding to PIF4. A similar role for PIF5 can also be speculated. This physical interaction results in targeted PIF4 phosphorylation and degradation (Lorrain et al., 2008). In high R : FR, DELLA stability would also be enhanced (Djakovic-Petrovic et al., 2007), resulting in PIF4 binding and inactivation (De Lucas et al., 2008). Phytochrome A would be degraded in both the cytosol and the nucleus. Overall, the increased relative enhancement of DELLA to PIF4 activity would restrain elongation growth.
When subject to reduced R : FR (Fig. 3b), phyB is converted to the inactive Pr form. This dissociates from PIF proteins and exits the nucleus. The reduction in phyB Pfr would therefore result in increased abundance of nuclear PIF4. Increased PIF4 activity would result in enhanced expression of genes involved in elongation growth (Nozue et al., 2007; Lorrain et al., 2008). Under these conditions, phyA activity in the FR-HIR response mode would provide some antagonism of elongation growth, possibly through the binding and targeted degradation of PIF4 (Huq & Quail, 2002). Overall, the relative enhancement of PIF4 to DELLA activity would promote rapid elongation growth.
In prolonged vegetational shade (Fig. 3c), an additional role for GA can be proposed. Increased expression of transcripts encoding GA 20-oxidase have been reported following low R : FR, EOD-FR and simulated shade treatment in Arabidopsis (Devlin et al., 2003; Salter et al., 2003; Hisamatsu et al., 2005; Sessa et al., 2005). Such observations suggest GA concentrations may increase in vegetational shade. Increased concentrations of GA would enhance DELLA degradation, thereby relieving DELLA-mediated inhibition of PIF function (Djakovic-Petrovic et al., 2007; De Lucas et al., 2008). It can therefore be speculated that in prolonged low R : FR, increased GA synthesis would increase DELLA degradation, further enhancing PIF4 function and maintaining elongation growth.
Flowering time in Arabidopsis is regulated through the complex integration of multiple environmental signals which act together to regulate the expression of floral meristem identity genes (for review, see Simpson & Dean, 2002). The light-quality pathway mediating acceleration of flowering in low R : FR has been shown to operate through the floral integrator FLOWERING TIME (FT), independently of the transcriptional regulator CONSTANS (CO) (Cerdán & Chory, 2003; Halliday et al., 2003). Indeed, expression of both FT and the meristem identity gene LEAFY (LFY) have been shown to display constitutively elevated expression in early-flowering phyB mutants (Blázquez & Weigel, 1999; Cerdán & Chory, 2003; Halliday et al., 2003). A potential signalling component linking phyB to FT expression was identified in a mutagenesis screen by Cerdán & Chory (2003). A plant displaying late flowering behaviour in both long and short days was isolated. Sequencing of the mutation revealed a nuclear-localized protein with similarity to some transcriptional activators, termed PHYTOCHROME AND FLOWERING TIME 1 (PFT1). Mutants deficient in PFT1 displayed impaired flowering responses to EOD-FR treatments. Double mutants, deficient in phyB and PFT1, displayed petiole lengths similar to phyB controls, but flowered at the same time as wild-type plants. Together, these observations suggest separate signalling mechanisms operate downstream of phytochromes to regulate elongation and flowering responses to low R : FR. This notion is supported by reports of natural Arabidopsis accessions displaying low R : FR-mediated elongation growth but no acceleration of flowering (Botto & Smith, 2002).