Following light-induced nuclear translocation, the phytochromes induce changes in gene expression to regulate plant development. PIF3 and other PIFs (phytochrome-interacting factors), members of the bHLH (basic helix–loop–helix) family of transcriptional regulators, interact specifically with the active Pfr conformer of the phytochrome molecule, suggesting that the PIFs are key components of phytochrome signal transduction. The mechanism by which the PIFs transduce phytochrome signals is not understood. After initial studies that suggested that PIF3 was a positive regulator of phytochrome signalling, mutant studies indicated that the PIFs primarily act as negative regulators in the pathway. Furthermore, in some cases they accumulate in the dark and are degraded upon illumination by the ubiquitin–26S proteasome system. At least for PIF3, the protein degradation depends on direct interaction with the phytochrome molecule and is preceded by protein phosphorylation. In this review, the current understanding of the role of the PIFs in phytochrome-mediated photomorphogenesis will be summarized, and recent findings suggesting an unanticipated dual mechanism of action of the PIFs will be discussed.
Light is the most precious energy and informational resource for plants. As sessile organisms, plants need constantly to monitor their light environment to optimize their growth and development accordingly. Higher plants have at least four families of photoreceptors that sense the quantity, quality, direction, and duration of light: an elusive UV-B light receptor; the UV-A/blue light-absorbing cryptochromes; the UV-A/blue light-absorbing phototropins; and the red- (R) and far red- (FR) absorbing phytochromes (Schäfer and Nagy, 2006). These photoreceptors mediate different photomorphogenic responses throughout the life cycle of plants such as germination, seedling de-etiolation, shade avoidance, and flowering.
Phytochromes are soluble chromoproteins that have the capacity to interconvert reversibly between two conformers: the inactive R light-absorbing Pr form and the biologically active FR light-absorbing Pfr form (Tu and Lagarias, 2005). In Arabidopsis, the phytochromes are encoded by a small gene family of five members (PHYA–PHYE). Among them, phyA and phyB have the most prominent functions (Franklin and Whitelam, 2004), phyC has a complementary role to phyB in several responses (Franklin et al., 2003a; Monte et al., 2003) and has been recently shown to contribute to natural variation in flowering time (Balasubramanian et al., 2006), and phyD and phyE are often redundant to phyB but can have a more prominent role under specific environmental conditions (Franklin et al., 2003b; Halliday and Whitelam, 2003; Halliday et al., 2003).
Upon light activation, the phytochromes are translocated from the cytoplasm into the nucleus (Nagatani, 2004), where they induce rapid changes in gene expression (Jiao et al., 2007). Once in the nucleus, the active phytochromes can interact with PIF3 (phytochrome-interacting factor 3) and related PIF members of the basic helix–loop–helix (bHLH) family of transcription factors (Duek and Fankhauser, 2005). Based on the initial studies with phyB and PIF3, it was proposed that the phytochrome–PIF interaction may provide a direct transcriptional mechanism that could bring the light signal directly to the promoters of light-regulated genes (Quail, 2002a). However, no evidence has been obtained to support this model, and the mechanism by which the PIFs regulate photomorphogenesis is still under intense research. Here a review is presented of how our understanding of the role of the PIFs has progressed since the discovery of the founding member PIF3, and recent data that suggest that the mechanism of action of the PIFs in phytochrome signalling is more complex than originally anticipated are discussed.
Phytochrome signalling components: PIF3 as a dream positive regulator candidate
The original phytochrome signalling candidates were identified by genetic screens for mutants defective in seedling de-etiolation under continuous long-term (4–5 d) FR (FRc) or R (Rc) irradiation. These screenings looked for defects in hypocotyl length and cotyledon area characteristic of photomorphogenic mutants affected in phyA or phyB signalling (Fankhauser and Casal, 2004). PhyA is the only phytochrome mediating seedling de-etiolation in FRc, and phyB is the main phytochrome species mediating the de-etiolation response to Rc, in particular the inhibition of hypocotyl length (Franklin et al., 2003b; Tepperman et al., 2004). Collectively, genetic evidence gathered through the identification of photomorphogenic mutants affected in seedling de-etiolation in FRc and/or Rc suggested a complex network of phytochrome signalling pathways: while some of the components were specific, some others were shared by both phyA and phyB. Also, positive and negative signalling components on each pathway were identified (Hudson, 2000). Some of them have been cloned and characterized (Wang and Deng, 2002; Hiltbrunner et al., 2005; Hiltbrunner et al., 2006). Progress on the function of these phytochrome signalling components is beyond the scope of this review.
A breakthrough in the phytochrome field came with the use of yeast two-hybrid screens carried out with the phytochrome molecule. This approach led to the exciting identification of the phytochrome-interacting PIF3 (Ni et al., 1998), a member of the bHLH superfamily of transcription factors (Toledo-Ortiz et al., 2003). Although the original yeast two-hybrid screen was carried out with the non-photoactive C-terminal domain of phyB, PIF3 was also found to interact with the photoactive N-terminal domain of phyB (Shimizu-Sato et al., 2002) as well as with the non-photoactive C-terminal domain of phyA (Ni et al., 1998), indicating that sufficient determinants for PIF3 binding are present in both phytochrome domains. In vitro pull-down assays with the full-length phytochrome molecule showed that PIF3 interaction is specific to the biologically active Pfr form of the phytochrome, with higher apparent affinity for phyB than phyA (Ni et al., 1999; Zhu et al., 2000). The identification of PIF3, together with the finding that the phytochrome rapidly translocates from the cytoplasm into the nucleus upon light activation (Sakamoto and Nagatani, 1996; Kircher et al., 1999; Yamaguchi et al., 1999), revolutionized the field of phytochrome signalling and created a new paradigm of phytochrome action where the light perceived by the phytochrome could be directly transduced onto pre-existing transcription factors such as PIF3 to modulate the expression of light-regulated genes (Quail, 2002a).
Two initial lines of evidence indicated that PIF3 was a central player in phytochrome signalling: (i) the Pfr-dependent binding was reduced in phytochrome missense mutant derivatives that were previously shown to be defective in signalling in vivo (Ni et al., 1999), thus providing correlative evidence that PIF3–phytochrome binding was important in the plant cell; and (ii) Arabidopsis antisense lines with reduced levels of PIF3 transcript showed strongly reduced sensitivity to Rc and FRc light during seedling de-etiolation (Ni et al., 1998). A second set of evidence led to the postulation of an initial model of phytochrome action through PIF3. First, PIF3 localizes constitutively into the nucleus (Ni et al., 1998) and is able to bind in vitro specifically to the G-box DNA sequence CACGTG, a DNA motif enriched in the promoters of light-responsive genes (Martínez-García et al., 2000). In addition, in vitro analysis also showed that phyB Pfr can specifically and photoreversibly form a ternary complex with the DNA-bound PIF3 (phyB Pfr–PIF3–G-box complex) (Martínez-García et al., 2000). Based on all the above, a model was proposed in which PIF3 would act as a central positive regulator of phytochrome necessary to induce light-regulated genes. In this model, phytochrome molecules would function as integral, light-switchable components of transcription regulator complexes at target promoters following photoconversion-induced translocation into the nucleus. The photoreceptor could then modify transcriptional activity directly by acting as a co-regulator or indirectly by biochemically modifying either PIF3 or other components of the transcriptional machinery (Quail, 2000, 2002b). This attractive model provided an explanation of how the phytochrome would transmit the light signal directly into the promoters of target genes, thus transmitting changes of light input into rapid modulation of gene expression through interaction with positive acting transcription factors such as PIF3. However, later studies presented below challenged this initial working model.
PIF3 and other related PIFs as negative regulators of phytochrome signalling
Contrary to the initial work on PIF3, subsequent evidence did not support the early model of PIF3 acting as a positive regulator of phytochrome signalling. Instead, analyses of mutants deficient in PIF3 and other related bHLH proteins indicated that they mainly act as negative regulators in the pathway.
In Arabidopsis, the bHLH family of transcription factors has >160 members (Bailey et al., 2003; Heim et al., 2003; Toledo-Ortiz et al., 2003). Based on the evolutionary identity of the bHLH domain, it was established that PIF3 belongs to a phylogenetic subfamily of 15 members (Toledo-Ortiz et al., 2003). Six of these PIF3-like factors (PIF1/PIL5, PIF4, PIF5/PIL6, PIL1, HFR1, and SPT) have also been shown to be involved in phytochrome signalling (Duek and Fankhauser, 2005; Penfield et al., 2005). Of them, HFR1 and PIL1 do not interact directly with the phytochromes (Fairchild et al., 2000; Fankhauser and Chory, 2000; Soh et al., 2000; Khanna et al., 2004), whereas PIF1 (also known as PIL5 for phytochrome-interacting factor-3-like 5) (Huq et al., 2004; Oh et al., 2004), PIF4 (Huq and Quail, 2002), and PIF5 (PIL6) (Yamashino et al., 2003; Fujimori et al., 2004; Khanna et al., 2004) also interact preferentially with the Pfr form of the phytochrome, like PIF3 (Khanna et al., 2004). All these PIFs carry in their N-terminal domain a short motif (13 residues long in PIF3) called APB (for active phyB binding) that mediates the interaction with phyB Pfr (Khanna et al., 2004). The APB motif is found exclusively in the PIF3-related subgroup and it appears to be sufficient for phyB Pfr interaction (Khanna et al., 2004). PIF1, as well as PIF3, also interacts with the Pfr form of phyA (Zhu et al., 2000; Huq et al., 2004). PIF3 has stronger affinity for phyB (Zhu et al., 2000), whereas PIF1 has similar affinity for phyA and phyB (Huq et al., 2004). In PIF3, the interaction with phyA Pfr is mediated through a short motif (18 residues long in PIF3) called APA (for active phyA binding) that is necessary for the interaction (Al-Sady et al., 2006).
De-etiolation studies of pif1, pif4, and pif5 mutants under Rc showed that they all participate in phyB signalling, consistent with their binding affinity. Surprisingly, pif4 and pif5 mutants displayed short hypocotyls under Rc, a phenotype opposite to that of phyB and indicative of a negative role in the phyB pathway (Huq and Quail, 2002; Fujimori et al., 2004). Likewise, pif3 mutants were also shown to be hypersensitive to Rc (Kim et al., 2003; Monte et al., 2004). This led to the conclusion that PIF3 acts negatively in the phyB pathway under Rc (Kim et al., 2003; Monte et al., 2004), in contradiction to the initial studies using antisense plants (Ni et al., 1998). For PIF1, although the pif1 mutant did not show any apparent phenotype in Rc, PIF1-overexpressor seedlings displayed a long hypocotyl in Rc suggestive of a negative role in phyB signalling (Huq et al., 2004; Oh et al., 2004). De-etiolation analyses under FRc of mutant seedlings deficient in the phyA-binding PIF1 and PIF3 factors suggest that they act as negative regulators in the phyA pathway. The pif1 mutant displayed a short-hypocotyl phenotype in FRc that is opposite to that of phyA (Oh et al., 2004), and although PIF3 does not appear to mediate hypocotyl elongation in FRc (Kim et al., 2003; Monte et al., 2004), pif3 and pif3-ox seedlings had larger and smaller cotyledons, respectively, under these conditions (Kim et al., 2003). All the above results showed that there is good correlation between the phytochrome binding properties of the PIFs and their role in phytochrome-regulated de-etiolation under long-term Rc or FRc irradiation. Furthermore, they led to the suggestion that the PIFs are negative regulators of phytochrome signalling.
The negative regulatory model
A pivotal study by Bauer et al. (2004) showed that the PIF3 protein is stable in the dark and it is rapidly degraded upon exposure to light. This light-induced degradation is promoted by phyA, phyB, and phyD (Bauer et al., 2004; Al-Sady et al., 2006) and is mediated by the ubiquitin–26S proteasome pathway (Park et al., 2004; Al-Sady et al., 2006). Similarly, it has been shown that PIF1 is also stable in the dark and rapidly degraded upon light treatment through the ubiquitin–26S proteasome machinery (Shen et al., 2005). Together, these results suggest that light-induced degradation of the PIFs might be a general mechanism to regulate their activity. Phytochrome-induced degradation of the PIFs would thus promote photomorphogenesis by removing repressors of the pathway. This negative regulatory model could be considered analogous to the stimulus-induced degradation of repressors in the auxin and gibberellin pathways (Huq, 2006). Protein degradation is a common regulatory system that equips the cells with a fast response mechanism to inactivate proteins rapidly (Moon et al., 2004). In particular, it has been shown that the degradation of transcription factors is important in the regulation of light signalling (Hoecker, 2005; Yang et al., 2005). Regulated proteolysis in light signalling also takes place in the dark to remove positive regulators such as HY5, LAF1, and HFR1, and keep photomorphogenesis repressed (Osterlund et al., 2000; Seo et al., 2003; Duek et al., 2004; Jang et al., 2005; Yang et al., 2005). Interestingly, these factors are stabilized upon transfer to light, in contrast to the PIFs (Osterlund et al., 2000; Seo et al., 2003; Duek et al., 2004), providing a balance of negative- and positive-acting regulators to optimize photomorphogenesis (Huq, 2006).
Recently, it has been shown that phyA and phyB induce rapid phosphorylation of PIF3 (Al-Sady et al., 2006). This occurs redundantly by direct interaction prior to the ubiquitination and degradation of PIF3. The requirement for direct binding of phyA and phyB to PIF3 was shown in transgenic plants expressing a PIF3 variant doubly mutated in the APA and APB motifs in which phosphorylation and degradation were not detectable (Al-Sady et al., 2006). These results confirmed the direct involvement of the phytochromes in the removal of PIF3 upon light exposure and provided further evidence supporting a negative regulatory model for PIF action. The phytochrome-induced phosphorylation of proteins such as PIF3 might be the primary biochemical modification following phytochrome–PIF interaction. It has been proposed that this phosphorylation following phytochrome–PIF interaction could represent a tagging system for the proteasome pathway to inactivate a previously active PIF3 (Al-Sady et al., 2006). Alternatively, phytochrome–PIF3 interaction, phosphorylation, and/or ubiquitination could be necessary to activate PIF3 transiently prior to degradation (Al-Sady et al., 2006), an attractive regulatory strategy common in other systems (Lipford and Deshaies, 2003).
However, although for most responses the PIFs act as negative regulators, for some responses they have a positive role in phy signalling. For example, PIF3 has been shown to be a positive regulator of chlorophyll and anthocyanin accumulation (Kim et al., 2003; Monte et al., 2004). These results suggest that the PIFs might act in different branches of phytochrome signalling that might be regulated through distinct mechanisms.
The long second look: diversity of function early following light exposure
Old perceptions of phytochrome signalling in photomorphogenesis harboured two essential concepts: first, that the photomorphogenic phenotypes observed in different phytochrome mutants following long-term irradiation represent the sum of the actions of the individual phytochrome photoreceptor over time; and, secondly, that the extent of the phenotype observed is directly linked to the differences in gene expression dependent on that photoreceptor. A recent series of papers examining gene expression profiles in dark-grown seedlings exposed to different durations of R or FR light in different phy backgrounds have shed doubts on these central beliefs and also greatly broadened understanding of the roles of different phytochromes, phyA and phyB, in particular, in establishing the photomorphogenic programme. First, while both R and FR light exposure of dark-grown seedlings affects expression of over ∼10% of the Arabidopsis genome (Tepperman et al., 2001, 2004, 2006), the phytochromes responsible for particular expression pattern changes, and the temporal nature of the gene expression profile were surprising: both R and FR light affect different kinds of genes early (within 1–3 h) and late (3–24 h) following light exposure. Signalling and transcription factor genes tend to be induced/repressed early, while genes in energy metabolism and photosynthesis were affected later. More interestingly, the effect of phyA and phyB on the gene expression profile did not in fact correlate with their previously postulated roles: given their visible phenotypes, particularly in hypocotyl and cotyledon growth inhibition, phyA and phyB had been postulated to be the only or major photoreceptor for FR or R light, respectively. The gene expression studies showed, however, that phyA dominates the gene expression programme early under both FR and R, and that phyB has a relatively minor role early in R (Tepperman et al., 2006). PhyB controls a somewhat larger gene set later during photomorphogenesis, but that still only amounts to 5% of the genes induced or repressed >2-fold under long-term Rc (Tepperman et al., 2004). This is consistent with the protein lability profile of phyA in R (Clough and Vierstra, 1997). The dominant role of phyA in control of early response genes also became apparent in a gene profiling study on PIF3 in dark-grown seedlings exposed for 1 h to Rc, similar to the above studies (Monte et al., 2004). Interestingly, PIF3-dependent genes at 1 h Rc are co-regulated by phyA, but not phyB, despite pif3 mutants being hypersensitive in hypocotyl growth in Rc (but not FRc), a pathway dominated by phyB. Importantly, rather than affecting a broad range of the genes induced/repressed by phytochromes in R or FR, as previously predicted (Martínez-García et al., 2000), PIF3 affects a very specific and small cluster of genes at 1 h Rc (chloroplast function-related genes).
In summary, when analysing phytochrome signalling from a gene expression perspective, one is led to the conclusion that first, most of the gene expression changes following initial exposure to FR or R light are under phyA control, and, second, that phytochrome-interacting transcription factors such as PIF3 affect specialized subsets of these early response genes that control a particular early photomorphogenic phenotype. This contrasts with the older views about PIF3 that placed it as a general regulator of photomorphogenesis in the phyB pathway (Martínez-García et al., 2000). These studies also emphasize the possibility that phytochrome signalling, early following initial irradiation (meaning in the initial hours following light exposure of the etiolated seedling) and in the long term (meaning growth in Rc or FRc for 3–5 d), may be implemented in different ways: early gene expression changes as executed mainly by phyA, and carried out by PIFs, may in fact not correlate with the ultimate outcome as measured by changes in gross morphology after 4–5 d in Rc; and conversely, phyB-dependent effects on hypocotyl and cotyledon morphology may not be the outcome of the observed early changes in the gene expression profile.
In fact, the notion that the classical long-term morphological phenotype may not represent the early action of the phytochromes has prompted a search for morphological markers closer to the initial light signal. One direct way to assess changes early is to examine de-etiolation phenotypes at earlier stages, a path previously hampered by technical difficulty. Some early studies have addressed early morphological markers such as the apical hook opening, cotyledon unfolding, and chlorophyll accumulation following initial light exposure, and the dependence of these responses on the phytochromes (Liscum and Hangarter, 1993; Reed et al., 1994). Recent advances in live imaging techniques have allowed real-time measurements of the hypocotyl growth rate as well as the growth direction of etiolated seedlings following light exposure (Parks and Spalding, 1999). Such studies have revealed that phyA and phyB, as consistent with the gene expression studies, impact hypocotyl growth in different ways: phyA acts initially to achieve transient growth inhibition, and phyB takes over later for the duration of the irradiation period. Studies focusing on early morphological responses such as hook opening and cotyledon separation have shown that both phyA and phyB are involved in these responses and that they are most active in different temporal windows following initial light exposure (E Monte and P Quail, unpublished results).
Collectively, the above studies suggest that key aspects of primary phytochrome signalling are difficult to uncover by analysis of the end-point phenotype after a 4–5 d light treatment; and cannot capture the dynamic interplay between the different receptors. The same can be said for pif mutants. While initially most of the PIF genes, such as PIF3, PIF4, PIF5/PIL6, and PIF1/PIL5, were analysed for their effect on long-term growth phenotypes, and concluded to be negative regulators of phytochrome-mediated morphological responses, recent studies discussed below have focused on how PIFs affect the initial transition of the dark-grown seedling to light. While much remains to be analysed, early indications are that PIFs in fact affect a large diversity of early photomorphogenic responses, and appear to have occupied niches of specialization. An above-mentioned gene profiling study on PIF3 [examing gene expression changes after 1 h Rc treatment of etiolated seedlings (Monte et al., 2004)] revealed that most PIF3-dependent genes fall into one broad category, nuclear-encoded genes for chloroplast components. This finding is consistent with the observation that pif3 mutants are defective or delayed in the greening process of etiolated seedlings when first exposed to light.
PIF3 plays a critical positive role in the expression of these genes in Rc (Monte et al., 2004), as well as for the expression of anthocyanin pathway genes in FRc (Kim et al., 2003; Shin et al., 2007). Recent studies further indicate that PIF5/PIL6 may affect early phytochrome responses mediated by ethylene (R Khanna, Y Shen, and P Quail, unpublished results), specifically affecting opening of the apical hook, and that other PIFs including PIF3 and PIF4 may differentially affect early morphological responses such as apical hook opening and cotyledon separation (P Leivar, E Monte, and P Quail, unpublished results). A more highly specialized role is filled by PIF1, which does not appear to play a role during light signalling per se but even before initial light exposure as a regulator of both germination (Oh et al., 2004) and protochlorophyllide synthesis in the dark (Huq et al., 2004). PIF1 acts to decrease active gibberellic acid levels in the seed by both down-regulating gibberellic acid biosynthesis enzymes and up-regulating gibberellic acid-catabolizing enzymes (Oh et al., 2006). Further, PIF1 negatively regulates protochlorophyllide synthesis in the etiolated seedling to prevent accumulation of free protochlorophyllide which can result in oxidative damage and bleaching, as evident in pif1 mutants exposed to light following prolonged growth in darkness (Huq et al., 2004). PIF1 therefore keeps both the seed and etiolated seedling poised for light signals, removal of PIF1 by phyA and phyB allows gibberellic acid accumulation and germination, and, later, PIF1 removal from the etiolated seedling relieves the block on chlorophyll synthesis (Shen et al., 2005). It therefore appears that first, several of the complexities and dynamics of phytochrome signalling can be better captured by focusing on early morphological phenotypes rather than growth phenotypes under prolonged continuous irradiation and, second, PIFs have a specific role in assisting the phytochrome photosensory system in efficiently and rapidly implementing the photomorphogenic programme following initial light exposure.
Positive and negative signalling roles of PIFs: dichotomy between early and long-term signalling
While PIFs have a variety of different roles early, carried out in conjunction with phyA, which dominantly implements the early light-induced gene expression profile, their long-term growth phenotypes and strong interaction with phyB (Khanna et al., 2004) has placed them as negative regulators of phyB-controlled growth responses (see above) (Duek and Fankhauser, 2005).
Collectively, this indicates that PIFs mediate two distinct types of phenotypic responses. A problem in trying to understand how PIFs signal early and late is the phytochrome-induced degradation to lower steady-state levels of several of them: PIF1 (Shen et al., 2005; Oh et al., 2006), PIF3 (Monte et al., 2004; Park et al., 2004; Al-Sady et al., 2006), and PIF4 (B Al-Sady and P Quail, unpublished results). This degradation happens in a time frame where PIFs are also most active in gene regulation (Monte et al., 2004). It is not clear what role this degradation may play towards the signalling activity of the PIFs, given that they can act, under different circumstances, as positive regulators (PIF3 in dark to light transition) or negative regulators of growth responses under long-term Rc. It is possible that the PIFs use similar mechanisms in early and long-term signalling, and that the mechanism involved in degradation may either activate some of their functions or adjust their protein level to the specific response. However, recent evidence (B Al-Sady and P Quail; E Monte, P Leivar, and P Quail; R Khanna and P Quail, unpublished results) raises the question of whether the processes of early and long-term signalling through the phytochromes, as mediated by the PIFs, could be fundamentally different in nature. Whereas the dark-grown seedling has adapted to implement the photomorphogenic programme rapidly and precisely, the fully de-etiolated seedling needs to optimize its long-term growth responses to the prevalent light quality and quantity rather than to initiate new gene expression programmes. The lack of correlation between these two phases of photomorphogenesis has further become evident from a functional profiling study (Khanna et al., 2006). The effect on growth morphology under long-term irradiation of reverse-genetic disruption of 32 genes, previously shown to be strongly induced or repressed by light in the first hours, revealed that these genes mostly do not affect the long-term phenotype. Further, a time-course analysis of the pif3 hypocotyl phenotype indicates that it appears very late, only after 2–3 d of Rc (Monte et al., 2004), far removed from the pif3 effect on greening following initial light exposure.
What are the mechanisms of PIF action early and in the long-term phase of de-detiolation? The mechanism of action of PIFs early does seem to involve their action as transcription factors. The lines of evidence supporting this conclusion are (i) that PIF3 is found at target promoters in vivo (in this case at anthocyanin biosynthesis pathway genes) (Shin et al., 2007); and (ii) that the DNA binding capacity of PIF3 is required for the light induction of PIF3-dependent genes (B Al-Sady and P Quail, unpublished results). These studies, as well as studies on phyA and pif3 early mutant phenotypes (Monte et al., 2004; Tepperman et al., 2006), and their respective gene expression profiles, have suggested a general framework model where PIFs implement subsets of primarily phyA-dependent gene expression programmes to achieve certain early cellular and morphological light responses. However, little is understood about the mechanism of PIF-mediated phytochrome signalling, and phyB signalling in particular, under long-term irradiation. Importantly, this function of the PIFs is apparently facilitated by the lower steady-state level established by the phytochromes during long-term irradiation. While it is possible that the long-term phenotype is a residual manifestation of early PIF action, prior to its degradation, it appears more likely that PIFs act more immediately on phytochrome signalling while at this low level, especially since the pif mutants have different early phenotypes, and almost identical phenotypes under long-term Rc. What then could be a mechanism by which PIFs act late, during long-term irradiation? One clue comes from the well-established observation that the hypocotyl and cotyledon growth responses are highly dependent on the absolute levels of the active phyB photoreceptor (Koornneef et al., 1980; Wester et al., 1994; Wagner et al., 1996). This raised the possibility that PIFs might act to inhibit active phyB in long-term irradiation. Indications are that this is indeed the case and that inhibition may be occurring directly as a result of the interaction of the PIFs and the phyB photoreceptor. This appears to be true at least for PIF4, which requires its ability to interact specifically with phyB to implement its negative regulatory role on hypocotyl growth (Khanna et al., 2004). In this context, degradation of the PIF proteins could also be critical in the regulation of the long-term phenotype. PIF3 is reduced to ∼15% of its dark levels during Rc (Monte et al., 2004), and the same appears true for PIF1/PIL5 (Shen et al., 2005; Oh et al., 2006). This lower level may be optimal for the regulation of phyB signalling output, possibly via the regulation of the absolute active Pfr phyB pool. However, the mechanism by which PIFs could act on Pfr phyB to reduce its activity remains to be defined. One possibility is that PIFs simply titrate Pfr phyB into an inactive pool in an interaction-dependent and reversible manner, or, alternatively, that PIFs act to inactivate a fraction of the phyB molecules irreversibly, by modification or degradation.
Our understanding of phytochrome signalling has changed considerably over the last few years, away from the old monolithic idea of one diagnostic phenotype representing one mechanism, to more complex models. The enticing molecular properties of the PIF proteins initially seemed to support, erroneously, previous phytochrome signalling models, but have since served as excellent tools to dissect the apparent intricacies of phytochrome signalling. First, close study of the role of PIFs has uncovered the complex nature of early phytochrome signalling, with different PIF factors controlling separate facets of photomorphogenesis, such as control of germination, protochlorophyllide synthesis, chloroplast biogenesis, and co-ordination of rapid morphological responses (such as hook and cotyledon unfolding). Second, scrutiny of early and long-term irradiation phenotypes has indicated that the mechanisms of PIF action may be different in these two time frames. PIF3 is a good example. It acts positively as a regulator of chloroplast-related genes, but acts negatively on phyB-controlled hypocotyl inhibition, with the pif3 mutant effect on the latter only appearing after 2–3 d of long-term irradiation. This early versus long-term mechanistic dichotomy has yet to be fully explained, but there are already some exciting hints from ongoing research. Most of the PIFs studied negatively impact phyB long-term signalling, in an unknown manner. It is known that growth responses are highly sensitive to the absolute amounts of available phyB–Pfr. Cytologically, this may correlate with the finding that phyB accumulates in large nuclear speckles, specifically during long-term irradiation (Nagy et al., 2001; Kircher et al., 2002; Chen et al., 2003). While the significance of the late speckles has been doubted for long-term growth control (Matsushita et al., 2003), it has been hypothesized that the speckles may serve as storage vehicles for available phyB. It is not clear whether PIFs may participate in controlling phyB storage in late speckles. If PIFs do not participate in such a mechanism, what could be an alternative model? It is becoming more evident that PIFs in fact regulate phyB protein levels. Evidence is pointing in this direction for at least four PIFs, PIF3, PIF4, PIF5/PIL6, and the previously unpublished PIF7 protein (Monte et al., 2004; B Al-Sady and P Quail; P Leivar, E Monte, and P Quail; R Khanna, Y Shen, and P Quail, unpublished results). The mechanism appears to be at the post-transcriptional level, and requires direct interaction with the phyB photoreceptor, and, for the case of PIF3 where it has been examined, does not require its ability to target DNA. Therefore, rather than participating directly in the phytochrome signalling pathway, PIFs appear to be homeostatic regulators of the phytochrome output under long-term irradiation by controlling photoreceptor abundance. What about mechanisms of signalling through PIFs early after initial light treatment? Mounting evidence, as discussed above, suggests that PIFs (especially PIF1 and PIF3) do act as transcription factors controlling the light response of differing sets of genes at this stage of the de-etiolation process. Recently, PIF3 has been shown to bind to anthocyanin biosynthetic gene promoters in vivo in FR light and also in the dark (Shin et al., 2007), suggesting that PIF3 can bind to the promoters of early light-regulated genes in a phytochrome-independent fashion, presumably allowing for the rapid light induction upon the initial illumination.
These studies have helped sharpen the notion of a mechanistic dichotomy in phytochrome signalling early, following initial light exposure, and later during long-term irradiation. This dichotomy may be implemented via the PIFs, acting in both phases by radically different mechanisms: acting alternatively as transcription factors of sets of genes critical to certain aspects of photomorphogenesis early, and as regulators of the phyB photoreceptor itself, to fine-tune phytochrome output during long-term irradiation. Future studies will have to address by which mechanism PIFs regulate phyB levels, and how they function early with phytochromes to act on photo-responsive genes.