Phytochrome Signaling in Green Arabidopsis Seedlings: Impact Assessment of a Mutually Negative phyB–PIF Feedback Loop

Abstract

The reversibly red (R)/far-red (FR)-light-responsive phytochrome (phy) photosensory system initiates both the deetiolation process in dark-germinated seedlings upon first exposure to light, and the shade-avoidance process in fully deetiolated seedlings upon exposure to vegetational shade. The intracellular signaling pathway from the light-activated photoreceptor conformer (Pfr) to the transcriptional network that drives these responses involves direct, physical interaction of Pfr with a small subfamily of bHLH transcription factors, termed Phy-Interacting Factors (PIFs), which induces rapid PIF proteolytic degradation. In addition, there is evidence of further complexity in light-grown seedlings, whereby phyB–PIF interaction reciprocally induces phyB degradation, in a mutually-negative, feedback-loop configuration. Here, to assess the relative contributions of these antagonistic activities to the net phenotypic readout in light-grown seedlings, we have examined the magnitude of the light- and simulated-shade-induced responses of a pentuple phyBpif1pif3pif4pif5 (phyBpifq) mutant and various multiple pif-mutant combinations. The data (1) reaffirm that phyB is the predominant, if not exclusive, photoreceptor imposing the inhibition of hypocotyl elongation in deetiolating seedlings in response to prolonged continuous R irradiation and (2) show that the PIF quartet (PIF1, PIF3, PIF4, and PIF5) retain and exert a dual capacity to modulate hypocotyl elongation under these conditions, by concomitantly promoting cell elongation through intrinsic transcriptional-regulatory activity, and reducing phyB-inhibitory capacity through feedback-loop-induced phyB degradation. In shade-exposed seedlings, immunoblot analysis shows that the shade-imposed reduction in Pfr levels induces increases in the abundance of PIF3, and mutant analysis indicates that PIF3 acts, in conjunction with PIF4 and PIF5, to promote the known shade-induced acceleration of hypocotyl elongation. Conversely, although the quadruple pifq mutant displays clearly reduced hypocotyl elongation compared to wild-type in response to prolonged shade, immunoblot analysis detects no elevation in phyB levels in the mutant seedlings compared to the wild-type during the majority of the shade-induced growth period, and phyB levels are not robustly correlated with the growth phenotype across the pif-mutant combinations compared. These results suggest that PIF feedback modulation of phyB abundance does not play a dominant role in modulating the magnitude of the PIF-promoted, shade-responsive phenotype under these conditions. In seedlings grown under diurnal light–dark cycles, the data show that FR-pulse-induced removal of Pfr at the beginning of the dark period (End-of-Day-FR (EOD-FR) treatment) results in longer hypocotyls relative to no EOD-FR treatment and that this effect is attenuated in the pif-mutant combinations tested. This result similarly indicates that the PIF quartet members are capable of intrinsically promoting hypocotyl cell elongation in light-grown plants, independently of the effects of PIF feedback modulation of photoactivated-phyB abundance.

INTRODUCTION

Plants monitor and respond to informational light signals from the environment using a set of sensory photoreceptors that include the phytochrome (phy) family (phyA to phyE in Arabidopsis) (Rockwell et al., 2006; Schafer and Nagy, 2006; Franklin and Quail, 2010; Quail, 2010). The phys track the relative levels of incident red (R) and far-red (FR) light by virtue of a capacity to switch reversibly between the biologically inactive Pr and active Pfr, conformers of the molecule, upon sequential absorption of R and FR photons. In dark-germinated seedlings, the inaugural conversion of Pr to Pfr upon initial exposure to light induces the familiar deetiolation process. In fully deetiolated, light-grown seedlings, exposure to vegetative shade imposes a variable reduction (but not abolition) of Pfr in the cell, because of the depletion of R, but not FR, photons (a reduction in R/FR ratio) from radiation filtered through, or reflected from, neighboring vegetation. This color-change-imposed reduction in Pfr levels induces the Shade-Avoidance Syndrome (SAS) in affected seedlings, displayed as accelerated extension-growth rates in hypocotyls, internodes, and petioles, retarded expansion rates in cotyledons, and retarded chloroplast development (Child and Smith, 1987; Smith and Whitelam, 1997; Franklin, 2008; Ballare, 2009; Ruberti et al., 2011). Under diurnal, light–dark cycles, the level of Pfr established at the end of the light period persists in the subsequent darkness, continuing to exert its regulatory activity for a defined period. Experimentally, a pulse of FR light, administered at the termination of the light period, a so-called End-of-Day-Far-Red (EOD-FR) treatment, that effectively removes this residual Pfr from the cell for the duration of the dark period is frequently used to examine this activity (Smith and Whitelam, 1997; Franklin, 2008; Franklin and Quail, 2010).

Current evidence indicates that the intracellular pathway by which the photoreceptor transduces these light signals involves translocation of the photoactivated phy molecule from the cytoplasm into the nucleus (Nagatani, 2004), where it induces changes in gene expression as a result of direct, physical interaction with members of a subfamily of basic helix-loop-helix (bHLH) transcription factors, called Phytochrome-Interacting-Factors (PIFs) (Castillon et al., 2007; Jiao et al., 2007; Bae and Choi, 2008; Leivar and Quail, 2011). The data indicate that the signal-transfer mechanism from the phy to the PIF molecule involves rapid phy-induced phosphorylation of the bHLH factor, which, in turn, triggers degradation of this factor via the ubiquitin–proteasome system (Bauer et al., 2004; Park et al., 2004; Shen et al., 2005; Al-Sady et al., 2006; Oh et al., 2006; Nozue et al., 2007; Shen et al., 2007; Al-Sady et al., 2008; Lorrain et al., 2008; Shen et al., 2008).

Genetic studies, using loss-of-function mutations in four of the PIFs (PIF1, PIF3, PIF4, and PIF5, designated here as the PIF quartet), have provided compelling evidence that these factors function with overlapping redundancy in dark-grown seedlings, to promote skotomorphogenesis, and that initial exposure to light induces deetiolation (the transition from skotomorphogenic to photomorphogenic development) as a consequence of the rapid, phy-triggered PIF degradation (Leivar et al., 2008a). Transcriptome analysis of the partially constitutively photomorphogenic pif1pif3pif4pif5 quadruple (pifq) mutant has identified the genes genome-wide that are regulated by these PIFs under phy control (Leivar et al., 2009; Shin et al., 2009), and has defined a subset, enriched in transcription-factor-encoding loci, that respond rapidly (within 1 h) to the initial light signal (Leivar et al., 2009). These rapidly light-responsive genes are thus considered candidates for being components of the primary transcriptional network targeted through the phy–PIF system.

Definition of the biological function(s) of the PIFs in fully deetiolated, green plants has been somewhat more complicated. Monogenic and higher-order pif mutants display light-hypersensitive seedling-phenotypes (shorter hypocotyls and larger cotyledons than wild-type (WT)) when grown under constant, prolonged R, FR, or white light (WL) irradiation (Huq and Quail, 2002; Kim et al., 2003; Fujimori et al., 2004; Huq et al., 2004; Monte et al., 2004; Oh et al., 2004; Khanna et al., 2007; Leivar et al., 2008b; Lorrain et al., 2008, 2009). This observation has been taken to indicate that the PIFs function as negative regulators of photomorphogenesis (Duek and Fankhauser, 2005; Castillon et al., 2007; Bae and Choi, 2008). Conversely, pif4, pif5, and pif4pif5-double mutants have reduced responsiveness to simulated shade (reduced hypocotyl elongation and marker-gene responsiveness compared with WT), and PIF4- and PIF5-overexpressors have the opposite phenotype (approaching constitutively long hypocotyls and petioles, and high marker-gene expression) (Lorrain et al., 2008). The mechanism by which these PIF activities might be exerted, in both light and shade, could in principle simply be by partial retention of the intrinsic skotomorphogenic-promotive activity of these factors. Consistent with this possibility, the evidence indicates that the reduction in PIF levels induced in light-grown WT seedlings does not completely abolish these proteins from the cell, but rather establishes a new, lower steady-state level than was present in darkness (Monte et al., 2004). Similarly, the abundance of the PIF4 and PIF5 proteins increases rapidly in white-light-grown WT seedlings upon their transfer to simulated shade, consistent with a function in promoting hypocotyl elongation (Lorrain et al., 2008). An alternative mechanism might involve feedback inhibition of the photomorphogenic-promotive activity of the phy molecule. Consistent with this possibility, the genetically imposed absence of the PIFs has been found to result in higher levels of phyB in the light, thus enhancing the photosensory, and thereby the photomorphogenic-inducing, capacity of the photoreceptor in the pif mutants compared with the WT (Khanna et al., 2007; Monte et al., 2007; Al-Sady et al., 2008; Leivar et al., 2008b). These data have thus been interpreted to indicate the existence of a mutually negative-feedback loop between the phyB and PIF proteins (Leivar and Quail, 2011) and there is evidence that PIF-induced phyB degradation is mediated via the ubiquitin–proteasome system using COP1 as an E3 ligase (Jang et al., 2010). The elevated phyB levels observed in the pif mutants have the capacity both to impose the observed light-hypersensitivity and to attenuate the extent of the shade response in these mutants, as has been demonstrated for seedlings engineered to overexpress phyB (Wester et al., 1994; Wagner et al., 1996; Roig-Villanova et al., 2006). Arguing against this possibility for the shade response is that the pif4 and pif5 mutations suppress the long-hypocotyl and high marker-gene phenotypes of the phyB mutant in R and WL (de Lucas et al., 2008; Lorrain et al., 2008). Collectively, these considerations indicate that both these suggested interwoven mechanisms may be operative simultaneously in regulating light-directed seedling development, but that there is ambiguity as to the relative quantitative contributions of each.

Here, using a phyBpif1pif3pif4pif5 pentuple (phyBpifq) mutant and various multiple pif-mutant combinations, coupled with Western blot analysis of phyB and PIF3 levels, we have examined the role of the phyB–PIF negative-feedback loop in modulating the magnitude of the deetiolation, shade-avoidance, and EOD-FR responses of Arabidopsis seedlings, with a focus on defining the relative quantitative contributions of the dual intrinsic-promotive and negative-feedback regulatory components of the PIFs’ activities.

RESULTS

PIFs Concomitantly Transduce and Modulate phyB-Regulated Photoresponses in Light-Grown Seedlings under Prolonged, Continuous R-Light (Rc) Irradiation

To extend our understanding of the interplay between phyB and the PIF quartet in light-exposed seedlings, we generated a pentuple phyBpif1pif3pif4pif5 (phyBpifq) mutant and compared its responsiveness to increasing fluence rates of continuous R-light (Rc) with that of WT and the parental phyB and pifq mutants grown in parallel. The data provide a number of important insights into this dynamically interactive system (Figure 1A and 1B).

Figure 1.

A Negative Regulatory Loop between phyB and the PIFs Regulate Growth Responses during Seedling Deetiolation. (A) Visible phenotypes of pifq, phyB, and phyBpifq pentuple mutant in the dark and under continuous R (Rc). Wild-type (WT) Col-0 and mutant seedlings were grown for 4 d in the dark or in Rc (28 μmol m⁻² s⁻¹).(B) Rc-Fluence response curves of pifq, phyB, and phyBpifq. WT and mutant seedlings were grown for 4 d under the indicated Rc fluence rates and hypocotyl lengths were measured. Data represent mean and standard error from at least 20 seedlings.(C) phyB and PIF3 protein levels (arrowheads) in WT, phyB, and pifq. Seedlings were grown for 4 d in the dark (4d-D) or in 1.26 μmol m⁻² s⁻¹ of Rc (4d-Rc). Samples were immunoblotted with antibodies against phyB and PIF3, and tubulin was used as a loading control. n.s., non-specific.(D) Quantification of the PIF3 and phyB protein levels from the blots in (C). Data were normalized to tubulin and presented as a percentage of value of WT-Dark samples.(E) Simplified model of regulation of seedling deetiolation consistent with the genetic and molecular data presented in panels (A)–(C) and elsewhere (reviewed in Leivar and Quail, 2011).

First, the lack of any discernable effect of Rc-light, across the full fluence-rate range, on the hypocotyl length of phyB seedlings, compared to that of completely dark-grown phyB or WT seedlings, reaffirms the conclusion that phyB is the dominant, if not exclusive, phy family member responsible for the Rc-light-imposed suppression of hypocotyl elongation observed in the WT. None of the other four phys, collectively or alone, can substitute for phyB in this activity.

Second, the similarly shorter hypocotyls of dark-grown pifq and phyBpifq mutants than those of the dark-grown WT (or phyB) reaffirm the intrinsic, photoreceptor-independent function of the PIF quartet in promoting seedling skotomorphogenesis in darkness.

Third, the residual responsiveness of the pifq mutant to increasing Rc-light fluence rates (compared to the phyBpifq mutant) indicates the existence of an additional transduction channel or mechanism mediating light-imposed hypocotyl growth inhibition that is not dependent on the presence of the PIF quartet.

Fourth, the complete absence of this residual Rc-light response in the phyBpifq pentuple mutant, across the full fluence-rate range, indicates, conversely, that the presence and photoactivation of phyB are selectively necessary for this additional suppression of hypocotyl elongation in the pifq mutant, as is the case for the light-induced response in WT. None of the other four phys can substitute for phyB in this activity.

Fifth, the observation that the differential in hypocotyl length between phyB and phyBpifq seedlings is the same (about twofold) in the dark and across all fluence rates of Rc-light is evidence that the PIF quartet members are necessary to support the maximal WT hypocotyl growth rate, and that they contribute quantitatively equivalently in the light as well as the dark, in the absence of phyB. This finding indicates that the intrinsic growth-promotive capacity of the PIF quartet remains constant in the dark and light, as long as phyB is not there to perceive and mediate the light signal. The data also reaffirm the existing conclusion (Leivar et al., 2008a; Shin et al., 2009; Leivar and Quail, 2011) that the phyB-imposed suppression of hypocotyl elongation in WT seedlings is mediated through reduction in the abundance of members of the PIF quartet. These conclusions suggest, in turn, that the level of the PIFs will be higher in the phyB mutant than in the WT in the light, due to the lack or reduced extent of degradation in the mutant. To test this prediction, we measured the endogenous PIF3 levels in 4-day-old Rc-grown phyB mutant seedlings (Figure 1C). The data show that, in contrast to WT seedlings, where PIF3 levels are reduced in Rc to very low levels (frequently below detection by immunoblot), phyB mutant seedlings retain about 40% of the level of the PIF3 protein present in the WT dark control (Figure 1C and 1D). Altogether, the data strongly suggest that the incomplete degradation of PIF3, and possibly the other PIFs, explain the etiolated hypocotyl phenotype of the phyB mutant under prolonged Rc conditions. The data also suggest that phys other than phyB are promoting a partial (about 60%) degradation of PIF3 under prolonged Rc conditions, but that this reduction in PIF3, and possibly other PIFs, is insufficient to reverse the etiolated hypocotyl phenotype of the phyB mutant.

Sixth, the hypersensitivity of the pifq mutant to Rc-light compared to the WT, coupled with the evidence that this hypersensitivity is dependent on the presence of phyB, suggests that the effectiveness of phyB signaling is somehow enhanced in the absence of the PIF quartet. In principle, this apparent enhanced activity could be the result of at least two possible alternative mechanisms. (1) Given that prolonged light apparently quantitatively reduces the levels of the PIFs to a new lower steady-state level, rather than removing them entirely from the cell in the WT (Monte et al., 2004), the complete absence of the PIF quartet in the pifq mutant might result in additional growth inhibition simply because of removal of the residual growth-promotive activity of the other PIFs that remain in the WT. (2) The absence of the PIF quartet in the pifq mutant could result in higher phyB levels, thereby enhancing the total photosensory sensitivity of the phyB population to any given incoming fluence rate of light. This latter alternative is predicted from previous data on pif3, pif4, and pif5 monogenic and various double-mutant combinations (Khanna et al., 2007; Al-Sady et al., 2008; Leivar et al., 2008b) and was directly demonstrated in 2-day-old pifq mutants by Leivar et al. (2009), as well as here in 4-day-old seedlings (Figure 1C and 1D). (3) A combination of both these alternatives is also possible. Regardless, the data suggest that all of these alternatives would require an additional channel through which phyB can act negatively on hypocotyl elongation. This, in turn, suggests the existence of additional phyB-supressible, growth-promotive, or phyB-inducible growth-inhibiting factor(s) (depicted as X in the schematic model in Figure 1E).

Collectively, the data strongly support the previously proposed bimodal capacity of the PIF proteins (Khanna et al., 2007; Al-Sady et al., 2008; Leivar et al., 2008b; Leivar and Quail, 2011) to concomitantly act to continuously promote elongation growth in both the dark and light, and to negatively modulate phyB activity in the light by reducing its abundance through feedback-induced degradation of the photoreceptor (Jang et al., 2010). We have subjected our data in Figure 1B to a preliminary quantitative analysis in Supplemental Analysis 1 and the associated Supplemental Figure 1, based on the schematic model in Figure 1E. This analysis provides a quantitative estimate of the relative contributions of phyB-induced PIF degradation and PIF-induced phyB degradation to the overall light-induced inhibition of hypocotyl elongation observed in light-grown seedlings.

The Mutually Negative phyB–PIF Feedback Loop Is Dynamically Modulated in Response to Simulated Vegetational Shade

A reduction in the R/FR ratio in incident light signals (vegetative shade) is perceived by the phy system, which then induces the SAS response (Child and Smith, 1987; Smith and Whitelam, 1997; Franklin, 2008; Ballare, 2011; Ruberti et al., 2011). To determine whether, and, if so, to what extent, potential changes in phyB levels, generated by the phyB–PIF negative-feedback loop in response to simulated shade, might contribute to the SAS response, we examined the dynamic changes in phyB levels in the pifq mutant, and various combinations of the individual pif mutants, in parallel with phenotypic analysis of these mutants, in seedlings subjected to shade treatment.

Initially, we examined the genetic interplay between the PIF quartet and phyB in high and low R/FR ratios, by phenotypically characterizing the various pif-mutant combinations using simultaneous irradiation with monochromatic R and FR light (dichromatic irradiations). Preliminary analysis showed that, under these conditions, a reduction in the R/FR ratio (i.e. Rc supplemented with FRc) induced enhanced hypocotyl elongation of WT seedlings characteristic of the SAS, as expected (Supplemental Figure 2A, top). The pifq mutant, however, although responding to the shade signal, showed shorter hypocotyls than the WT under these low R/FR conditions (Supplemental Figure 2A, top). This phenotype was more pronounced than in pif3 and pif4pif5 mutant seedlings (Supplemental Figure 2A, top). Although these data are consistent with members of the PIF quartet having a role in inducing SAS under these dichromatic conditions, the robust phenotypic differences already pre-established under the monochromatic Rc control conditions here (Supplemental Figure 2A, top) complicate the interpretation of the data.

Compared to these data in Rc, previously reported examinations of the responsiveness of pif4 and pif5 mutants to WLc have generated results ranging from minor differences from WT, to robust hypersensitivity (Lorrain et al., 2008; Hornitschek et al., 2009). Because such pre-established phenotypic differences in hypocotyl length in WLc can similarly complicate the interpretation of experiments aimed at defining the role of the PIFs in SAS responses (usually done in WLc supplemented with FRc), we wished to identify conditions that would consistently minimize the pif-mutant phenotypes in WLc. By examining hypocotyl elongation under different WLc fluence rates, we found that, at a fluence rate of 29 μmol m⁻² s⁻¹ (where elongation of WT hypocotyls is minimal (around 1 mm)), the manifestation of the pif3 mutant short-hypocotyl phenotype is strongly reduced, in comparison to a lower WLc fluence rate of 2.5 μmol m⁻² s⁻¹ (Supplemental Figure 2B). PIF3-overexpressor (PIF3-OX) seedlings (Al-Sady et al., 2008) show a converse long-hypocotyl phenotype in Rc and Rc+FRc dichromatic conditions (Supplemental Figure 2A, bottom) and in 2.5 μmol m⁻² s⁻¹ of WLc (Supplemental Figure 2B), but this phenotype is also reduced under higher WLc fluence-rate conditions (Supplemental Figure 2B). Therefore, we opted to standardly use WLc at a fluence rate of 20–30 μmol m⁻² s⁻¹ for all further analysis of responsiveness to simulated shade. In addition, we standardly grew seedlings for 2 d in WLc, and then for 5 additional days in WLc with (WL+FR, Low R/FR, Shade) or without (WL, High R/FR, Light) supplemental FRc—conditions that have been previously used to study SAS responses (Roig-Villanova et al., 2006, 2007).

Figure 2A and 2B shows a comparison of the hypocotyl elongation phenotypes of pifq, phyB, and phyBpifq mutant seedlings in response to simulated shade. These data show the absence of strong hypersensitivity of the pifq mutant to the WL-control conditions, as intended. Also as expected, the WL-grown phyB mutant shows elongated hypocotyls compared to WT, under these conditions. By contrast, this phenotype is robustly suppressed in the phyBpifq mutant in WLc compared to the phyB mutant (Figure 2A and 2B), similarly to that observed in Rc (Figure 1A and 1B). On the other hand, WT seedlings show elongated hypocotyls and petioles in response to simulated shade (WL+FR), and this phenotype is markedly reduced in the pifq mutant seedlings (WT hypocotyls grow 1.5-fold more compared to pifq) (Figure 2A and 2B). These data suggest that members of the PIF quartet have a role in inducing shade-avoidance growth responses in WL environments enriched in FR, as concluded in a recent similar study focused on transcriptome analysis (Leivar et al., 2012) and consistent with the reported role of PIF4 and PIF5 under similar conditions (Lorrain et al., 2008). In addition, because the pifq mutant still shows a significant elongation response in WL+FR (Figure 2B), the data suggest that phy-regulated factors other than the PIF quartet are also directly involved in inducing this response, in accordance with previous suggestions (reviewed in Franklin, 2008; Stamm and Kumar, 2010; Ruberti et al., 2011).

Figure 2.

The phyB–PIFs Negative Regulatory Loop Is Dynamically Modulated in Response to Changes in R/FR Ratio.

(A) Visible phenotypes of pifq, phyB, and phyBpifq pentuple mutant in high R/FR (Light, WL) or in low R/FR (Shade, WL+FR). Wild-type (WT) Col-0 and mutant seedlings were grown for 2 d in WLc (20 μmol m⁻² s⁻¹) from germination onward (2dWL) and for 5 additional days in the same WL conditions with (2dWL+5d[WL+FR], R/FR ratio 0.006) or without (2dWL+5dWL, R/FR ratio 6.48) supplemental FR light.

(B) Quantification of the hypocotyl elongation phenotypes of pifq, phyB, and phyBpifq. Hypocotyl length was measured from WT and mutant seedlings grown as in (A). Data represent mean and standard error from at least 20 seedlings.

(C) PIF3 protein levels (arrowheads) in WT, phyB, and pifq. Seedlings were grown for 2 d in WL (2dWL) and then transferred to WL supplemented with FR for 1 h (2dWL+1h[WL+FR]). Two-day-old dark-grown seedlings (2dD) were used as a control. Samples were immunoblotted with antibodies against PIF3 and tubulin was used as a loading control. n.s., non-specific.

(D) phyB protein levels in WT and pifq. Seedlings were grown for 2 d in WL (2dWL) and then transferred to WL supplemented with FR (2dWL+h[WL+FR]) for the indicated time (in hours). Samples were immunoblotted with antibodies against phyB and tubulin was used as a loading control.

The inclusion of the phyB and phyBpifq mutants in the present analysis unveils additional complexity in the dynamics of the phy–PIF system. The data in Figure 2B show that the phyB mutant hypocotyls are markedly longer (4 mm) than WT seedlings in our WLc (high R/FR) conditions (Figure 2B), as expected of the well-established constitutively shade-avoiding phenotype of this mutant. This phenotype is largely suppressed by the pifq mutation, since the hypocotyls of phyBpifq mutants are only 0.6 mm longer than WT under these conditions (Figure 2B). These data thus suggest that a large part of the phyB-mediated inhibition of hypocotyl elongation in WLc occurs through the phyB-induced degradation of these four PIFs, as concluded above for Rc-light-grown seedlings. This conclusion is consistent with the data of Lorrain et al. (2008), who reported a partial suppression of the phyB phenotype by a pif4pif5 double mutation under similar conditions. In addition, in contrast to WT and pifq, phyB mutant seedlings show a substantial reduction in hypocotyl length in response to FR supplementation (WL+FR) (phyB mutants grow 5.4 mm in WL but only 3.8 mm in WL+FR) (Figure 2B). This effect has been attributed to phyA-imposed inhibition of hypocotyl elongation in WL+FR in the absence of phyB, indicative of phyA-mediated antagonistic modulation of phyB-induced shade responses (Johnson et al., 1994; Devlin et al., 2003; Roig-Villanova et al., 2006; Franklin, 2008). Interestingly, this behavior is not observed in the phyBpifq mutants, which actually grow 2 mm in response to WL+FR (phyBpifq seedlings are 1.9 mm tall in WL and 3.9 mm in WL+FR), suggesting that the PIF quartet is required for this phyA-induced response in the absence of phyB. The data also suggest that hypocotyl elongation in phyBpifq mutants in response to enriched FR (WL+FR) might be mediated by yet other phys, possibly by that previously reported phyD (Devlin et al., 1999).

To examine the potential role of the phyB–PIF molecular feedback loop (Figure 1E) in modulating the phenotypic responses to shade under our defined high and low R/FR-ratio conditions, we monitored endogenous PIF3 and phyB abundance by Western blot analysis. First, we measured the endogenous PIF3 protein levels in 2-day-old dark and WLc-grown WT and phyB mutant seedlings, or in seedlings grown for 2 d in WLc and then exposed to 1 h of simulated shade (WL+FR) (Figure 2C). The data show that WLc (2dWL) imposes a reduction in PIF3 levels in WT seedlings compared to dark (2dD) controls, as expected—an effect that is strongly compromised in phyB mutants, in which only a partial degradation of PIF3 is observed (Figure 2C). The phenotypic analysis of the phyBpifq mutant (Figure 2B) is consistent with these data, suggesting that the phyB hyposensitive phenotype in WLc is established by the absence or incomplete degradation of members of the PIF quartet transcription factors. The data also show that partial inactivation of phy in response to an imposed reduction of the R/FR ratio (WL+FR) induces a rapid (1-h) partial reaccumulation of the endogenous native PIF3 protein in WT seedlings (Figure 2C), as observed in a recent similar study focused on transcriptome analysis (Leivar et al., 2012). These data are consistent with the reported shade-induced reaccumulation of ectopically overexpressed PIF4 and PIF5 proteins (Lorrain et al., 2008) and suggest a similar role for PIF3 in inducing SAS responses.

To assess whether the PIF quartet feedback-modulates phyB levels under WLc and in response to simulated shade, we measured phyB levels in WT and pifq seedlings over a time course in 2-day-old WLc-grown seedlings exposed to WL+FR for 0, 6, 12, and 24 h (Figure 2D). Similarly to Rc conditions (Figure 1C), the data show that pifq mutants have threefold higher phyB levels than WT under WLc conditions before the FR supplementation (Figure 2D). These increased phyB levels might contribute to the slightly short-hypocotyl phenotype observed in the pifq mutant (Figure 2B) under these saturating WLc conditions discussed above (Supplemental Figure 2B). The data also show that phyB levels increase about threefold in WT seedlings exposed to 12–24 h of FRc supplementation (Figure 2D)—an effect that is already observed to a certain extent within 6 h after the initiation of the FRc treatment. Interestingly, the starting elevated levels of phyB in the pifq mutant seedlings at time-zero (WL0) remain relatively constant, becoming similar to those of the WT seedlings within 12 h of initiation of the FRc supplementation (Figure 2D). These data suggest that the PIFs induce the degradation of phyB under high R/FR conditions, but that exposure to simulated shade (low R/FR) reverses this effect, leading to a reaccumulation of phyB.

Collectively, the data indicate that the mutually negative phyB–PIF regulatory loop is dynamically modulated in response to changes in the R/FR ratio. In high R/FR, the phy photoequilibrium is shifted towards the active state, leading to a phy-induced reduction in the PIF transcription factors and a reciprocal PIF-induced reduction in phyB photoreceptor levels. Conversely, in response to a reduction in the R/FR ratio, the phyB–PIF loop is partially inactivated, leading to a reaccumulation of both phyB (predominantly in the inactive Pr form) and PIFs, the latter presumably promoting the SAS program though their intrinsic transcriptional activities (Lorrain et al., 2008, 2009; Leivar et al., 2012).

phyB Levels in Selected pif-Mutant Lines Suggest that PIF Feedback Regulation of phyB Abundance Is Not the Dominant Determinant of the Magnitude of SAS Responses under Prolonged Continuous Shade Conditions

As shown above, phyB levels are elevated in pifq mutants compared to WT in prolonged WLc (high R/FR) conditions (Figure 2D)—a differential that is eliminated upon exposure to simulated shade (low R/FR) for 6–12 h, as phyB levels rise in the WT. Considering that phyB overexpression reduces the responsiveness of seedlings to simulated shade (Roig-Villanova et al., 2006), it is possible that this limited period of elevated phyB levels observed here determines, at least partially, the responsiveness of the pif mutants to the supplemental FRc. In order to examine possible correlations between phyB levels and the phenotypic responses to simulated shade, we first measured the phyB levels in selected pif-mutant combinations that either did (pif4pif5, pif3pif4pif5, and pifq; Figure 2 and 3A; Lorrain et al., 2008) or did not (pif3 and pif1pif3; Figure 3A) display a hypocotyl growth phenotype in response to the simulated shade treatment. Hypocotyl growth data for these lines, determined here in parallel, are included for direct comparison (Figure 3A, top) and are presented as a growth difference with respect 2-day-old WLc-grown seedlings (Figure 3A, bottom). We have recently used similar conditions to examine the relative functional roles of the individual PIFs in regulating phenotypic development in parallel with selected marker-gene expression during deetiolation and shade-avoidance responses (Leivar et al., 2012). A PIF3–OX line was also included here as not showing a shade phenotypic response, despite the growth-enhancing effect in WLc (Supplemental Figure 3).

Figure 3.

Examination of phyB Levels in Selected pif Mutant Lines Suggests that PIF-Regulation of phyB Levels Is Not a Major Determinant of the Magnitude of SAS Responses.

(A) Quantification of the hypocotyl elongation of selected WT and pif-mutant seedlings grown as in Figure 2. Top panel: Absolute hypocotyl lengths in 2dWL, 2dWL+5dWL, and 2dWL+5d[WL+FR]. Data represent the mean and the standard error of at least 20 seedlings. Bottom panel: Differential shade-responsiveness ([2dWL+5d[WL+FR]]–[2dWL]). For each genotype, mean hypocotyl length at 2dWL was subtracted from mean hypocotyl length at 2dWL+5d[WL+FR]. Differential responsiveness in light ([2dWL+5dWL]–[2dWL]) was also plotted as reference.

(B) PIF3 and PIF4/PIF5 differentially and dynamically regulate phyB levels in response to changes in the R:FR ratio. phyB protein levels were determined in wild-type Col-0 (WT), pif mutants, and PIF3-overexpressor (PIF3-OX) seedlings. Seedlings were grown under the same conditions described in Figure 2, either for 2 d in WL (2dWL, upper panel) or for 2 d in WL and then for 5 additional days in WL with (2dWL+5d[WL+FR], lower panel) or without (2dWL+5dWL, middle panel) supplemental FR. WT-1 and WT-2 refer to WT seedlings of two groups of seeds of different age used in the experiment. Samples were immunoblotted with antibodies against phyB and tubulin was used as a loading control.

(C) Quantification of phyB protein levels in WT and pif-mutant lines. phyB levels in each of the blots in (A) and in Supplemental Figure 4 were quantified, normalized to tubulin, and expressed relative to the WT value set at unity for each treatment. Mean phyB/tubulin relative to WT for each treatment was then calculated from the two biological replicates, except for pif1pif3 at 2dWL+5dWL in which only one replicate was available. Bars indicate standard deviation.

(D) PIL1 transcript levels in WT and pif-mutant seedlings grown for 2 d in WL (2dWL) or for 2 d in WL and then for 1 h in WL supplemented with FR (2dWL+1h[WL+FR]). PIL1 normalized to PP2A were determined by qPCR and the data are expressed relative to WT_2dWL set at unity. Data represent the mean and the standard deviation from two technical replicates.

The data indicate that PIF3, PIF4, and PIF5 act collectively to down-regulate phyB levels under high R/FR, both at 2 d (2dWL, Figure 3B, top) and 7 d (2dWL+5dWL, Figure 3B, middle), but that there are differential contributions of each of these PIFs to this process at the two tested developmental stages. First, after 2-d WLc, pif3 and pif1pif3 mutant seedlings show similarly higher levels of phyB than WT, whereas a marginal (if any) effect is observed in the pif4pif5 mutant under these conditions (Figure 3B, top). The data suggest a more prominent role for PIF3, compared to PIF4 and PIF5, in down-regulating phyB levels under these conditions. However, a more compelling effect of the pif4pif5 mutation was observed in the absence of PIF3, since both pif3pif4pif5 and pifq mutants showed higher levels of phyB than pif3 and pif1pif3 under these conditions (Figure 3B, top). Similar data were observed in 2dWL+5dWL seedlings, although, in this case, the pif4pif5 double mutant already showed significantly elevated phyB levels compared to WT (Figure 3B, middle). The data thus suggest that PIF3 plays a more prominent role than PIF4/5 in down-regulating phyB levels in 2-day-old WLc-grown seedlings, whereas the PIF4 and/or PIF5 contribution increases in 7-day-old WLc-grown seedlings compared to 2-day-old seedlings. Also consistent with this PIF3 activity, PIF3-OX seedlings show reduced phyB levels compared to WT both at 2 and 7 d in WLc (Figure 3B, top and middle).

In agreement with the 24-h time-course data in WL+FR (Figure 2D), long-term exposure to shade (2dWL+5d[WL+FR]) not only increased the phyB levels of WT seedlings (Figure 3B, bottom) compared to WLc conditions (Figure 3B, top and middle), but it also attenuated the initial (at 2dWL) differences in phyB levels in the selected pif-mutant and PIF3–OX seedlings compared to WT (Figure 3B, bottom). The reproducibility of the observations in Figure 3B is demonstrated by a second biological replicate of the data (Supplemental Figure 4) and is additionally supported by the observed PIF3 regulation of phyB levels under dichromatic Rc and FRc conditions (Supplemental Figure 5). The phyB levels depicted in Figure 3B and Supplemental Figure 4 are quantified in Figure 3C. Collectively, the data suggest that shade relatively rapidly (within 12 h; Figure 2D) suppresses the capacity of PIF3, PIF4, and/or PIF5 to down-regulate phyB levels in the WT, resulting in the sustained absence of substantial differences in abundance between the WT and various pif-mutant seedlings over the majority (4.5 out of 5 d) of our standard experimental shade period.

Direct comparison of the phyB levels (Figure 3B and 3C) and corresponding phenotypic shade-avoidance responses of the pif-mutant lines grown under the same long-term light and shade conditions here (Figure 3A) reveals a lack of correlation. On the one hand, the robust, apparently cumulative effect of progressive genetic removal of the PIF quartet on increasing phyB levels in both 2- and 7-day-old WLc-grown seedlings does not result in substantial differences in hypocotyl elongation between lines under these saturating WLc conditions at either time point. Conversely, the differences in responsiveness of WT and the different mutant combinations to 5 d of simulated shade shown here (Figure 3A and in separate experiments elsewhere; Leivar et al., 2012) are displayed in the absence of detectable differences in phyB levels that appear to exist for over 90% of the shade-induced growth period (the last 4.5 d of this 5-d period) (Figures 2D, 3B, and 3C). Although, in principle, the differences in phyB levels between the WT and various pif-mutant genotypes observed over the first 6–12 h (≤10%) of the shade-treatment period (Figures 2D, 3B, and 3C) could conceivably contribute to the differences in hypocotyl lengths recorded at the end of that period, the magnitude of that contribution appears likely to be a small fraction of the total differential in growth accumulated over 5 d. In addition, the differences in phyB levels observed at the start of the initial 6–12 h of shade (Figure 3B and 3C) are variable between the pif genotypes in a manner that is not robustly correlated with the final differentials in shade-induced hypocotyl elongation (compare Figure 3C at 2dWL versus Figure 3A, bottom). Taken together, these data suggest that the initial but transiently higher levels of phyB observed in the pif mutants compared to the WT at the start of the shade period are not the dominant determinant of the magnitude of the morphogenically visible responsiveness of these seedlings to shade.

On the other hand, direct comparison of the phyB levels in the 2-d WLc-grown seedlings (Figure 3B and 3C) with the magnitude of the rapid (within 1 h) shade-induced increases in PIL1 transcript levels (an established shade-responsive marker gene; Salter et al., 2003) in these seedlings (Figure 3D) does appear to show a general tendency towards decreasing PIL1 expression with increasing phyB levels across the different genotypes (Supplemental Figure 6). However, the low correlation coefficient for this relationship (r² = 0.32) is statistically insignificant (P > 0.05), indicating the absence of a robust causal relationship (Supplemental Figure 6). Because the differences in phyB levels are not expected to change substantially over the first 1 h of shade treatment (Figure 2D), the data do not appear to suggest that the pre-existing differences in phyB levels between the genotypes tested are the only or dominant factor determining the magnitude of this expression response to shade. Nevertheless, the overall trend of the best-fit curve in Supplemental Figure 6 may be consistent with the possibility that these phyB differences contribute, at least partially, to modulating the early changes in shade-induced gene expression in these lines. Thus, the data seem to indicate that the PIFs likely contribute directly to inducing shade-triggered early gene-expression responses through their intrinsic transcriptional-regulatory activity, but that the phyB–PIF feedback loop may partially modulate the magnitude of that response.

PIF3 Regulates Shade and EOD–FR Responses in Conjunction with PIF4 and PIF5

The observed reaccumulation of PIF3 in response to low R/FR ratio (Figure 2C) suggests that PIF3 may be involved in promoting phyB-mediated shade-avoidance responses, as has been previously proposed for PIF4 and PIF5 (Lorrain et al., 2008). Recently, by comparing monogenic, triple, and quadruple pif-mutant combinations, we have examined the individual contributions of each member of the PIF quartet for their involvement in shade-avoidance responses (Leivar et al., 2012). Those studies showed marginal contributions of PIF3 in regulating long-term hypocotyl elongation, in the presence or absence of the other three members of the PIF quartet, in response to shade (see also Figure 3A). We have extended this analysis here by examining in more detail the role of PIF3 and its interactions with PIF4 and PIF5 by analyzing (1) pif3pif4 and pif3pif5 double mutants in response to shade (Figure 4A and 4B); (2) a phyBpif3 double mutant in WLc (Figure 4C); and (3) pif-mutant responses to End-Of-Day-Far-Red (EOD-FR) treatment (Figure 4D).

Figure 4.

PIF3 Regulates Shade and EOD-FR Responses in Conjunction with PIF4 and PIF5.

(A) Visible phenotypes of WT, pif3, pif4, and pif5 single, double, and triple mutant seedlings in high R/FR (Light, WL) or in low R/FR (Shade, WL+FR). Wild-type (WT) Col-0 and mutant seedlings were grown as in Figure 2, for 2 d in WLc and for 5 additional days in the same WL conditions with (2dWL+5d[WL+FR]) or without (2dWL+5dWL) supplemental FR light.

(B) (Left panel) Quantification of the hypocotyl elongation of WT and pif-mutant seedlings shown in (A). Data represent the mean and the standard error from at least 20 seedlings. (Right panel) Differential shade-responsiveness. Mean hypocotyl length at [2dWL+5dWL] was subtracted from mean hypocotyl length at [2dWL+5d[WL+FR]] for each genotype.

(C) Quantification of hypocotyl elongation of WT, phyB, and phyBpif3. Seedlings were grown for 7 d in WL (2dWL+5dWL) as in Figure 2.

(D) (Left panel) Quantification of hypocotyl elongation of WT and pif-mutant seedlings in End-of-Day-Far-Red (EOD-FR) conditions. Seedlings were grown for 2 d in WL (2dWL) and then transferred for 5 additional days to either WLc (+5dWL), 14 h of WL followed by 10 h of darkness (+5d[14hW–10hD]), or 14 h of WL followed by a saturating pulse of FR (FRp) prior to the 10-h dark treatment (+5d[EOD-FR]). Data represent the mean and the standard error from at least 20 seedlings. (Right panel) Differential EOD-FR-responsiveness. Mean hypocotyl length at [2dWL+5d[14hWL–10hD]] was subtracted from mean hypocotyl length at [2dWL+5d[EOD-FR]] for each genotype.

To examine the contributions of PIF3 in the absence of PIF4 or PIF5, we analyzed the responses of the pif3pif4 and pif3pif5 double mutants to shade (Figure 4A and 4B) in comparison to the previously analyzed pif single, pif4pif5 double, and pif3pif4pif5 triple mutants (Figure 3A; Lorrain et al., 2008; Leivar et al., 2012). These data show marginally, if at all, shorter hypocotyls of the single pif3, pif4, and pif5 mutants compared to WT in WL+FR (Figure 4A and 4B), reaffirming our previous observations (Leivar et al., 2012), whereas more robust shade-responsiveness phenotypes are observed in the double and triple mutants (Figure 4A and 4B). These results indicate that the function of each of the PIFs is only marginally observed in the presence of the other members in the single mutants, but the effect is more obvious when two or more of the transcription factors are absent. This suggests that a certain degree of redundancy exists among these PIFs in regulating shade-avoidance responses, reaffirming previous analysis (Lorrain et al., 2008; Leivar et al., 2012). Together, the data indicate that PIF3 works in combination with PIF4 and PIF5 to induce hypocotyl elongation in response to simulated shade, in a similar way to that reported for PIF4 and PIF5 alone (Lorrain et al., 2008).

Overall, the data suggest that reaccumulation of the PIF3 protein in response to a decrease in activated phy (Pfr) level (under low R/FR, Figure 2C) does contribute to hypocotyl elongation responses in light-grown seedlings (Figure 4A and 4B). Consistent with this view, the pif3 mutation significantly suppresses the phyB mutant elongated-hypocotyl phenotype (Figure 4C; compare phyB to phyBpif3). Although this effect of the pif3 mutation on the phyB mutant phenotype (Figure 4C) is more modest than the effect observed for the pif4pif5-double (Lorrain et al., 2008) and pifq (Figure 2B) mutations, the data suggest that the elevated PIF3 protein levels observed in the phyB mutant, due to reduced proteolytic degradation (Figure 2C), are in part responsible for the elongated phenotype of this mutant.

Phytochrome inactivation induced by a terminal FR-pulse (FRp) at the end of the day period in seedlings growing under diurnal light–dark cycles (EOD-FR) has also been shown to promote higher reaccumulation of PIF3 during the subsequent night period than in non-FRp-treated seedlings (Monte et al., 2004). In order to test whether EOD-FR-induced growth responses (Smith and Whitelam, 1997; Franklin, 2008; Franklin and Quail, 2010; Kozuka et al., 2010) are regulated by PIF3, in combination with PIF4 and PIF5, we examined phenotypically the WT and selected pif mutants under EOD-FR conditions. These seedlings were grown for 2 d in WLc, and for 5 additional days in WLc (5dWL) or under a diurnal cycle of 14-h WLc and 10-h dark with (5d[EOD-FR]) or without (5d[14hWL–10hD]) a terminal FR pulse (FRp) at the end of the light period. The data show that introduction of a 10-h dark period (5d[14hWL–10hD]) induces a modest, if any, hypocotyl-growth response in all genotypes compared to seedlings kept in constant WLc (5dWL) (Figure 4D). By contrast, the EOD-FR treatment (5d[EOD-FR]) induces a pronounced growth effect in the WT compared to the diurnal dark periods alone (without the FRp) (5d[14hWL–10hD]), and this response is attenuated to varying degrees in the pif mutants tested (Figure 4D). In particular, although the pif3 and pif4 monogenic mutants do not display detectably shorter hypocotyls, the higher-order mutant combinations do display reductions in hypocotyl elongation in response to the terminal FR pulse. The data indicate that each pair of the three PIFs (PIF3, PIF4, and PIF5) can act together and that all three in combination act additively or synergistically to promote hypocotyl elongation. The evidence suggests, therefore, that PIF3, together with PIF4 and PIF5, acts over the night period to induce hypocotyl growth under diurnal cycles (14hWL–10hD), and that this effect is enhanced by removal of the phy Pfr conformer by the terminal FRp at the beginning of the dark period (EOD-FR). The abovementioned accelerated reaccumulation of the PIF3 protein during the dark period (Monte et al., 2004) is proposed to contribute to this response, and is considered also likely to occur for the PIF4 and PIF5 proteins, given their demonstrated rapid reaccumulation in response to shade (Lorrain et al., 2008).

Collectively, the data suggest that PIF3 plays a role in inducing shade-avoidance and EOD-FR growth responses—a function that is exerted in combination with PIF4 and PIF5. The PIF-dependence of the growth acceleration induced by EOD-FR removal of Pfr from the seedling is evidence that these PIFs are capable of additively promoting hypocotyl-cell elongation in light-grown plants, independently of the simultaneous presence of photoactivated phy, as suggested for PIF3, PIF4 and PIF5 from truncated light-dark diurnal cycling experiments (Nozue et al., 2007; Soy et al., 2012).

DISCUSSION

The data presented here address the question of whether, and to what extent, the PIF proteins retain the capacity to function autonomously in promoting hypocotyl elongation in light-grown plants, as opposed to indirectly, through feedback regulation of the level of photosensory sensitivity by negatively modulating phyB photoreceptor abundance. We have examined this question in three contexts: (1) deetiolated seedlings grown under prolonged Rc irradiation; (2) light-grown seedlings subjected to prolonged simulated vegetational shade (low R:FR ratio); and (3) seedlings grown under diurnal light–dark cycles, with or without a terminal FR pulse at the beginning of each dark period (EOD-FR treatment).

The evidence here that the phyB-null mutant displays negligible residual responsiveness to prolonged Rc, relative to dark-grown seedlings (Figure 1A and 1B), reaffirms the established view that phyB is the dominant phy-family member regulating hypocotyl cell elongation in response to this signal. These data also show that the maximum elongation rates achieved in the dark are sustained unabated in the light, in the absence of this specific photoreceptor, indicating that the cellular machinery that drives hypocotyl cell growth remains fully operative in this configuration. The observation that the pifq mutant has a short hypocotyl in the dark compared to WT reaffirms the established conclusion that the PIF quartet are, collectively, a major component of that machinery (Leivar et al., 2008a, 2009; Shin et al., 2009). Moreover, the absence of any residual responsiveness to light in the phyBpifq double mutant, relative either to the dark-grown pifq and phyBpifq seedlings or to the light-grown phyB mutant across all Rc fluence-rates tested, further supports the conclusion that the PIF quartet collectively continues to exert maximal growth-promoting capacity in the light in the phyB mutant. Interestingly, while the Rc-induced reduction in PIF3 protein levels in the phyB mutant relative to the dark control (Figure 1C) is not unexpected, given that photoactivated phyA is known to also induce PIF degradation (Bauer et al., 2004; Al-Sady et al., 2008), the absence of an effect of this reduction on hypocotyl elongation (Figure 1B) suggests that the high levels of the PIF proteins in the dark are collectively saturating for maximal growth promotion. Supporting this view, neither monogenic mutation nor overexpression of individual PIFs significantly reduces or increases hypocotyl elongation, respectively, in dark-grown seedlings (Huq and Quail, 2002; Kim et al., 2003; Fujimori et al., 2004; Huq et al., 2004; Monte et al., 2004; Oh et al., 2004; Shen et al., 2005).

The residual Rc-light responsiveness of the pifq mutant relative to the phyBpifq mutant indicates that phyB acts through an additional pathway, independently of the PIF quartet, to suppress hypocotyl elongation. This pathway is represented schematically as factor X in Figure 1E, indicating that such a factor(s) could act either positively or negatively on hypocotyl elongation and, conversely, that phyB could act either negatively or positively on this factor’s activity. Other PIF transcription factors, such as PIF6, 7, and PIF8 (Leivar and Quail, 2011), are potential candidates for this function. Regardless of the mechanism underlying this additional pathway, the enhanced levels of phyB in the pif mutants documented here (Figure 1C) and elsewhere (Khanna et al., 2007; Al-Sady et al., 2008; Leivar et al., 2008b, 2009; Jang et al., 2010) suggest that PIF-induced feedback reductions in phyB photosensory capacity in the WT are likely to ameliorate the extent of reductions in PIF levels, thereby accentuating the effectiveness of the residual growth promotive activity of the PIFs in light-grown seedlings. Taken together, the above data suggest that the residual, albeit low, steady-state levels of the PIFs in light-grown WT seedlings continue to act autonomously to promote hypocotyl elongation in these seedlings in antagonism of phyB-imposed suppression of elongation, and that the visible phenotype is the net result of the balance between these two antagonistic activities. The PIF-imposed negative-feedback attenuation of phyB activity inherent to this interaction may reflect a receptor desensitization mechanism responsible for signal termination, as proposed by Jang et al. (2010).

The pif-mutant analysis presented here shows that members of the PIF quartet act to promote hypocotyl elongation in light-grown seedlings, in response to prolonged simulated vegetational shade, and that PIF3 contributes to this promotive activity in addition to, and in conjunction with, the established involvement of PIF4 and PIF5 (Figure 4A and 4B and Lorrain et al. (2008)). The shade-induced increase in the endogenous PIF3 abundance in WT seedlings (Figure 2C), similar to that reported for PIF4 and PIF5 in overexpressing lines (Lorrain et al., 2008), is consistent with higher rates of cell elongation in response to the shade signal as are promoted in dark-grown seedlings by high PIF levels. However, although the data indicate that the phyB–PIF negative feedback loop is operative in WT seedlings and results in rapid (within 6–12 h) adjustments in phyB levels in response to the shade signal (Figure 2D), overall, there is a lack of a robust correlation between phyB levels and the extent of the phenotypic and early marker-gene responsiveness to this signal (Figure 3 and Supplemental Figure 6). Despite a clear reduction in hypocotyl elongation in the pifq mutant compared to WT in response to prolonged shade, no elevation in phyB levels is detected in the mutant compared to the wild-type across the majority (90%) of the shade-induced growth period (Figures 2D and 3C). Although differences observed between pif-mutant and WT seedlings in early (within 45 min) shade-induced increases in hypocotyl elongation rates (Cole et al., 2011) could reflect pre-existing elevated phyB levels in the mutants at the initiation of the shade treatment, the contribution of such differences to the final cumulative growth response after 4.5 d of shade (Figure 3A) seems likely to be negligible, given the transient nature of this phyB differential (Figures 2D and 3C). Similarly, comparison of the phyB levels with the shade-induced growth phenotype across multiple pif-mutant combinations displaying varying degrees of shade-responsiveness does not show a robust correlation (Figure 3). These results suggest, therefore, that PIF feedback modulation of phyB abundance does not play a dominant role in modulating the magnitude of the PIF-promoted, shade-responsive phenotype under these conditions. Taken together, the data indicate that the PIF quartet contribute collectively to the accelerated hypocotyl elongation rate induced by simulated shade and suggest that a major fraction of that contribution results from the intrinsic, autonomous promotive activities of the PIF proteins, rather than indirectly via modulation of phyB levels. Similarly, it appears that the intrinsic transcriptional promotive activities of the PIF quartet collectively contribute substantially to the observed rapid induction of the marker gene PIL1 to the shade signal (Figure 3D and Supplemental Figure 6), in accordance with the proposed action of PIF4 and PIF5 (Lorrain et al., 2008; Hornitschek et al., 2009), although there is also evidence that the phyB–PIF negative-feedback loop may modulate the magnitude of this contribution through short-term regulation of phyB abundance (Figure 3 and Supplemental Figure 6; Roig-Villanova et al., 2006).

The results of the EOD-FR experiments reported here support the conclusions from the shade-avoidance experiments and add an additional dimension. The evidence shows that the well-known enhancement of hypocotyl elongation rates induced by the essentially complete, FR-light-driven removal of photoactivated phy at the initiation of the dark period, in seedlings grown under diurnal light–dark cycles (Wester et al., 1994; Smith and Whitelam, 1997; Franklin, 2008; Kozuka et al., 2010), depends on the collective promotive activities of PIF3, PIF4, and PIF5 during the dark period. This conclusion is based both on the observed rapidly induced accumulation of high levels of PIF3 protein in darkness following a FR pulse (Monte et al., 2004) and on the reduced magnitude of FRp-induced hypocotyl elongation in the multiple pif-mutant combinations tested here (Figure 4D). These data support the interpretation that, in WT seedlings, the residual Pfr present at the light-to-dark transition persists and acts to suppress PIF levels through degradation for some period, thus prolonging the light-imposed inhibition of hypocotyl elongation into the dark period and, conversely, that early Pfr removal strongly enhances PIF growth-promotive activity over the diurnal dark cycle. Importantly, this interpretation supports the conclusion that the PIFs can act autonomously to promote growth in light-grown seedlings over a period when Pfr is essentially absent, and so any indirect effects of feedback-imposed differences in phyB abundance should be negligible or absent.

Taken together, the present data indicate that the hypocotyl-elongation phenotype of light-grown WT seedlings is largely the net result of the dynamic mutual antagonism between the sustained, intrinsic, growth-promotive activities of the PIF transcription factors and the phyB-induced reduction of PIF abundance, but that the extent of this reduction in PIF levels can be partially ameliorated by PIF-induced feedback reduction of phyB abundance, thus accentuating the PIF promotive activity (Figure 5). In addition, the residual responsiveness of the pifq mutant to shade suggests the existence of a pathway or factor(s) (Factor X in Figure 5), in addition to that of the PIF quartet, that promotes growth in a manner similar to that proposed for the deetiolation response under continuous light (Figure 1E).

Figure 5.

Model Depicting a Central Role of the phyB–PIFs Negative Regulatory Loop in Regulating Growth Responses to Changes in R/FR Ratio.

In seedlings growing under continuous red light (Rc) or under high R/FR conditions (upper panel), the phy photoequilibrium is shifted towards the active Pfr state. Photoactive phy interacts with and induces the proteolytic degradation of members of the PIF quartet, and this reduction of the transcription factors inhibits the growth of the seedling. In turn, the PIFs induce the proteolytic degradation of the phyB photoreceptor, thus reducing the overall sensitivity of the seedling to light to optimize growth. In seedlings growing under low R/FR conditions (lower panel), phy photoequilibrium is shifted towards the inactive Pr form, causing a partial reversion to the etiolated state (Leivar et al., 2012). Under these conditions, the phyB–PIF negative regulatory loop is abrogated, resulting in an increase in phyB and PIF abundance. Accumulation of these PIFs induces a downstream gene-expression regulatory network that results in the induction of SAS responses (Leivar et al., 2012). A significant part of the phy-mediated growth responses to changes in R/FR is mediated by factors other than the PIF quartet, here depicted as X.

METHODS

Plant Materials

Mutant combinations used for this study were derived from single mutant alleles phyB-9 (Reed et al., 1993), pif1-1 (Huq et al., 2004), pif3-3 (Monte et al., 2004), pif4-2 (Leivar et al., 2008b), and pif5-3 (Khanna et al., 2007). Some of the higher-order mutant combinations used were described elsewhere, such as pif3phyB (Monte et al., 2004), pif3pif4 (Leivar et al., 2008b), pif4pif5 (Leivar and Quail, 2012), and pif1pif3, pif3pif4pif5, and pif1pif3pif4pif5 (pifq; Leivar et al., 2008a). pif3pif5 was obtained by crossing pif1pif3, pif4-2, and pif5-3 as described (Leivar et al., 2008a). phyBpifq was obtained by crossing pifq and pif3pif4phyB (previously obtained by crossing pif3pif4 and pif3phyB). PIF3 overexpression (PIF3-OX) line was described elsewhere (HA:WT-PIF3; Al-Sady et al., 2008).

Seedling Growth and Measurements

Seeds were sterilized and plated in a GM medium without sucrose as described (Monte et al., 2003; Leivar et al., 2009) and were generally stratified for 5 d in darkness at 4°C. Dark and Rc-grown seedlings were induced to germinate with 3 h of WL, followed by a terminal 5-min saturating FR pulse to avoid pseudo-dark effects as reported (Leivar et al., 2008a, 2009). Seedlings were then incubated in the dark or in the indicated Rc fluence rate at 21°C for the specified time.

Unless otherwise indicated, studies under simulated shade conditions were performed by first transferring the seeds to continuous WL (19 μmol m⁻² s⁻¹, R/FR ratio of 6.48) for 2 d at 21°C and then by growing the seedlings for 5 additional days under the same WL fluence rate with (WL-FR, R/FR ratio of 0.006) or without (WL) supplemental FRc.

End-of-day FR experiments were performed as described before (Wester et al., 1994) with modifications. Seedlings were grown for 2 d in WLc (20 μmol m⁻² s⁻¹) and then transferred for 4 additional days to light–dark cycles (14-h WL–10-h dark) with or without a 10-min saturating FR pulse at the end of the light period. Control seedlings were grown for 6 d in WLc.

For experiments performed mixing monochromatic Rc+FRc light, seedlings were generally grown directly for 3 d in Rc (10 μmol m⁻² s⁻¹) and then grown under the same Rc fluence rate supplemented (low R/FR ratio of 0.02) or not (Rc) with FRc for the indicated time. As an exception, a Rc fluence rate of 11 μmol m⁻² s⁻¹ and a low R/FR ratio of 0.03 was used in Supplemental Figure 2A (top).

Hypocotyl elongation was measured typically from at least 20 seedlings as described (Monte et al., 2004; Leivar et al., 2008a). Measurement of light fluence rates was as described before (Monte et al., 2004).

Protein Extraction and Immunoblotting

Protein extracts were prepared from Arabidopsis seedlings as described (Leivar et al., 2008b). The Protein DC kit (Biorad, Hercules, CA) was used to quantify total protein and β-mercaptoethanol was added to the samples just before loading. Equal amounts of protein per sample were then subjected to polyacrylamide gel electrophoresis. PIF3 was immunodetected by using affinity-purified anti-PIF3 antisera (Al-Sady et al., 2006), whereas mouse monoclonal anti-phyB (B1 and B7) antibodies (Somers et al., 1991) were used to immunodetect phyB. A mouse monoclonal antibody against alpha-tubulin (Sigma) was used as a control for loading. Anti-rabbit-HRP and anti-mouse-HRP were used as secondary antibodies (Promega), and ECL or ECL plus chemiluminescence kits (Amersham) were used for detection. PIF3 and phyB normalized to tubulin was quantified from the blots using Image J software as described (Leivar et al., 2008b).

PIL1 Expression Analysis by q–PCR

Q–PCR analysis was performed as described (Leivar et al., 2009). Briefly, 1 μg of total RNA was extracted using the RNeasy Plus plant mini kit (Qiagen) and then treated with DNAse I (Invitrogen) according to the manufacturer’s instructions, to further eliminate genomic DNA. First-strand cDNA synthesis was performed with the Super-Script First Strand cDNA synthesis for RT–PCR kit (Invitrogen) and oligo (dT20) as a primer. 10 μl of 30× diluted in water cDNA samples were used for real-time PCR as described (Khanna et al., 2007; Leivar et al., 2009). Primers to detect PIL1 (AT2G46970) and the reference gene PP2A (AT1G13320) were as described (Shin et al., 2007; Leivar et al., 2009). Each PCR was repeated twice and normalized gene expression was represented relative to the WL-grown WT set as a unity.

Accession Numbers

Arabidopsis Genome Initiative database accession numbers are: AT2G20180 (PIF1/PIL5), AT1G09530 (PIF3), AT2G43010 (PIF4), AT3G59060 (PIF5/PIL6), and AT2G18790 (phyB).

SUPPLEMENTAL DATA

Supplementary Data are available at Molecular Plant Online.

FUNDING

This work was supported by Comissionat per a Universitats i Recerca del Departament d’Innovació, Universitats i Empresa fellowship of the Generalitat de Catalunya (Beatriu de Pinós program) and Marie Curie International Reintegration Grant PIRG06-GA-2009–256420 to P.L., by grants Marie Curie IRG-046568, Spanish “Ministerio de Ciencia e Innovación” BIO2006-09254 and BIO2009-07675, and Generalitat de Catalunya 2009-SGR-206 to E.M. and by National Institutes of Health Grant GM-47475, Department of Energy Grant DE-FG03-87ER13742, and USDA Agricultural Research Service Current Research Information System Grant 5335–21000–030–00D to P.H.Q.