Primary Phototropic Signal-Response

(last updated: 28 Aug 2006 - PLEASE NOTE: This page is under MAJOR revision, hence the broken look - black text has been updated, grey has not and is woefully out of date)


Phototropism - the directional curvature of organs in response to lateral differences in light intensity and/or quality - represents one of the most rapid and visually obvious responses of plants to changes in their light environment (Liscum, 2002). Most studies, including our own, have focused on the positive phototropic response (bending towards the actinic light) observed in seedling stems. Figure 1 shows the phototropic response of etiolated Arabidopsis seedlings as an example.

Figure 1. Time-lapse movie of phototropism in dark-grown Arabidopsis seedlings.

Three-day-old etiolated seedlings were exposed to five 1 sec pulses of unidirectional blue light at a fluence rate of 0.002 mmol m-2 s-1 (obtained from one blue light emitting diode; lmax = 440 nm, 30 NM half-band) from the left, separated by 20 min intervals of darkness. Seedlings were allowed to develop phototropic curvatures n complete darkness. Photographs were taken with a digital camera under infra-red light every 10 min over a 300 minute period. Note that the slight backwards bend in the upper hypocotyls region is an artifact that results from the seedlings being grown along an agar surface such that the cotyledons constrain the positive phototropic response because they often lodge in the agar.

For other neat movies of "plants in motion" we direct the reader to Roger Hangarter's cool movie page - here

In nature stem phototropism likely provides plants with an effective means for maximizing photosynthetic light capture and thus may have appreciable adaptive significance (Iino, 1990; Liscum and Stowe-Evans, 2000; Iino, 2001, 2006). Any adaptive advantage provided by phototropic responses are likely to be particularly important during the early stages of growth and establishment of seedlings (Iino, 1990; Liscum, 2002) and during gap filling situations in dense canopy conditions (Ballare, 1999). In collaboration with Prof. Candi Galen in Biological Sciences, we are examining phototropism in Arabidopsis seedlings under greenhouse and field conditions and have already found that phototropism provides a dramatic fitness advantage during seedling establishment and early development (Galen et al., 2004), apparently through influences on root phototropism and subsequent drought tolerance (Galen et al., 2006). Ongoing physiological and molecular ecology studies are described elsewhere.

At the physiological level phototropism can be distinguished from other light-modulated directional growth responses such as nastic (Satter and Galston, 1981) and circadian-regulated (McClung, 2001) leaf movements by two major criteria: 1) directionality of response and 2) cellular basis of the growth response. With respect to directionality, phototropic curvatures are oriented relative to the direction of the incident light, while nastic and circadian-regulated movements are not. Changes in cell elongation rates across the bending organ drive the development of phototropic curvatures (Baskin et al., 1985; Briggs and Baskin, 1988; Orbovic and Poff, 1993), while reversible swelling/shrinking of pulvinar cells is prerequisite for many leaf movement responses (Hart, 1988; Koller, 1990).

As will be discussed later, the changes in elongation rates that mediate phototropism appear to be established through an integrated response as cells respond to a lateral gradient of auxin across the organ (lowest on the lit side, highest on the shaded side) formed as a result of the directional light stimulation (Went and Thimann, 1937; Iino, 1990; Liscum and Stowe-Evans, 2000; Liscum, 2002; Esmon et al., 2005, 2006; also see Figure 2).

Figure 2. Cellular model for role of auxin in phototropism.



[The downloadable PowerPoint Presentation above represents elements of models previously discussed in press (Liscum and Stowe-Evans, 2000; Tatematsu et al., 2004; Esmon et al., 2006), as well as unpublished works. As such this PowerPoint Presentation represents INTELLECTUAL PROPERTY of the Liscum Laboratory. Feel free to utilize this Presentation for classes or other presentations but PLEASE give proper credit to Mannie Liscum for its intellectual and physical development.]

The following is a description of the model (specific details are referred to later on this page)

Frame one of this animated model presents a stylized hypocotyl of Arabidopsis in longitudinal section through three rows of cells (middle, complete view; top and bottom, partial views) in a resting (dark) state, where auxin is present in a slight aplical to basal gradient (the deeper the tone of yellow the more auxin). The apoplast is shown in white.

Animation results from subsequent individual key strokes as follows:

The first key stroke results in the appearance of nuclei, containing an inactive heterodimeric ARF-AUX/IAA complex bound to DNA (in this case the AuxRE region of a tropic stimulus-responsive gene, EXP8); the presence of inactive (color coded red!) EXP8 in the apoplast; and inactive H+-ATPase molecules (purple) in the plasmamembrane.

The second key stroke shows the first set of responses to unidirectional blue light given from the left: 1) auxin is relocalized such that highest levels accumulate (NOTE deepening of yellow shade) farthest from the incident stimulus such that the auxin gradient shift from weakly apical-basal to strongly lateral; 2) next the increased auxin promotes the proteosome-dependent (not shown) degradation of the AUX/IAA proteins in the right most cell, allowing the formation of ARF-ARF homodimers and an active complex (green); 3) which then helps recruit/stabilize interactions between the target gene (EXP8) promoter sequences and the core transcriptional machinery (light purple); and 4) thus resulting in transcription of EXP8 and accumulation of EXP8 mRNA (blue squiggly lines).

The third key stroke results in clearing of nuclear components (for simplification of subsequent animation) and translation of EXP8 mRNA into inactive apoplastic EXP8 protein (red) around the right most cell (where EXP8 mRNA was most highly produced).

The forth key stroke shows how the lateral auxin gradient would simultaneously activate increased H+-ATPas activity, thus "acidifying" the apoplast more on the side of the hypocotyl where auxin is highest (e.g., farthest from the incident stimulus). EXP proteins are inactive at neutral pH (resting state of the apoplast of this dark-grown seedling) but become active (turn green in the animation) in acidic conditions. Increased EXP8 activity is then predicted to allow cell wall expansion (creep) to allow turgor-driven cell elongation on the side farthest from the light.

We are currently testing several aspects of this working model - like all models it is only as good as it is testable!


What do we know? (aka, a historical tour of phototropism and the Liscum Lab involvement)

Perception of directional light cues

Phototropism has been the subject of study for more than 120 years (e.g., Darwin, 1880; Sachs, 1887), however the identity of the photoreceptor(s) mediating this UV-A/blue light response remained elusive for most of this period. In fact, heated debates about the identity of the chromophore (light absorbing co-factor) utilized by the phototropic receptor(s) – flavin versus carotenoid - have frequently flared up in the field over the past 40 or so years (Liscum, 2002). However, thanks to APOaG (Awesome Power of arabidopsis Genetics) this controversy finally seems to be over. Thanks to mutational analyses in Arabidopsis, a pair of related flavin-based photoreceptor proteins, phototropin 1 and phototropin 2, have been identified as the primary receptors of phototropic stimuli (Briggs et al., 2001a; Christie and Briggs, 2001; Briggs and Christie, 2002). While the phototropins have also been found to regulate blue light-induced stomatal and chloroplast movement responses (Jarillo et al., 2001; Kagawa et al., 2001; Sakai et al., 2001; Kinoshita et al., 2001, 2003; Doi et al., 2004), the following sub-sections describe the identification and biochemical characterization of the phototropins as they relate to phototropism.


Phototropin 1 (phot1)

Historically the first significant progress towards the biochemical identification of a phototropic receptor came in 1988 when Gallagher and colleagues reported a blue-light-activated phosphorylation of a plasma membrane-localized protein in etiolated pea seedlings. A number of subsequent studies examined the photophysiological properties of this light-dependent phosphorylation reaction in a variety of species (most notably maize, oat, pea, and Arabidopsis) and associated it with the phototropic response. For example, the phosphorylation reaction occurs in the most phototropically sensitive tissues, is strongest in the tissue closest to the light and decreases in strength moving away from the lit side, is fast enough to precede the development of curvature, its action spectrum matches that for phototropism, and shows similar dark-recovery kinetics as phototropism after a saturating irradiation (Short and Briggs, 1994; Briggs and Huala, 1999; Briggs and Christie, 2002).

The first genetic connection between the phosphorylation reaction and phototropism came when Reymond and colleagues (1992) showed that a phototropic mutant of Arabidopsis, strain JK224 (Khurana and Poff, 1989), exhibited little, if any, blue light-induced phosphorylation. Interestingly, strain JK224 was independently proposed, based solely upon the photophysiological properties of the mutant, to harbor a lesion in a low-fluence rate phototropic photoreceptor (Khurana and Poff, 1989; Konjevic et al., 1992). Liscum and Briggs found that null mutations in the NONPHOTOTROPIC HYPOCOTYL 1 (NPH1) locus, of which JK224 is an allele (nph1-2), lack both a hypocotyl phototropic response (Figure 3) and the target protein for the blue light-dependent phosphorylation reaction, and thus proposed that the NPH1 locus encodes the apoprotein for a phototropic receptor capable of blue light-induced autophosphorylation (Liscum and Briggs, 1995; see Figure 4a and b).


Figure 3. Phototropic response of wild-type and a nph1-null allele.

Seedlings were grown in darkness for 3 d and then exposed to 8 h unidirectional blue light (from left) from the left.

Figure 4a-b. Blue light-induced phosphorylation of the NPH1 protein in Arabidopsis.

Panel A, Autoradiograph of gel shown in B.

Panel B, Silver-stained gel of microsomal membranes isolated from various nph1 mutants (and their respective wild-type progenitors) and subjected to an in vitro phosphorylation assay with gamma-32P-ATP. Arrow indicats the position of the NPH1 protein, which is phosphorylated in response to blue light in wild-type (Col and Est) but is missing (or present at undetectable levels) in the mutants.

(Data are from Liscum and Briggs, 1995)


When the NPH1 gene was isolated by positional cloning it was found to encode a protein containing the eleven signature domains of Ser/Thr protein kinases (Figure 5a) and of the proper size to be the substrate for an autophosphorylation reaction (Huala et al., 1997). Although primary sequence analyses yield nothing obvious that indicated inherent photoreceptor activity for NPH1, a repeated amino-terminal sequence motif was identified that exhibits homology to a subfamily of PAS domains found in sensor-proteins regulated by light, oxygen, or voltage (Huala et al., 1997; Zhulin and Taylor, 1997; Taylor and Zhulin, 1999). These PAS-like domains of NPH1, designated LOV1 and LOV2 (for their relationship to light, oxygen, and voltage-regulated PAS domains), were each subsequently shown to bind one FMN molecule (Christie et al., 1998; see Figure 5a).


Figure 5a. Domain structure of phototropin 1 (phot1).
FMN-binding LOV1 and LOV2 domains are shown in yellow. The Ser/Thr protein kinase domain is shown in red.


Soluble nph1 holoprotein (NPH1 apoprotein with associated FMN cofactors) isolated from a heterologous baculovirus/insect cell expression system was shown to exhibit blue light-dependent autophosphorylation (Figure 6) with kinetic, fluence-response, and action spectrum characteristics essentially like those obtained with native Arabidopsis nph1, and for phototropism itself (Christie et al., 1998). Various solution spectroscopic methods, as well as x-ray crystallographic studies, have been used to demonstrate that the kinase domain of nph1 is likely activated in response to a conformational change(s) associated with a light-driven, but dark-reversible, covalent bond formation between a conserved Cys within the LOV domain and the FMN chromophore (Salomon et al., 2000, 2001; Crosson and Moffat, 2001, 2002; Swartz et al., 2001, 2002). The nph1 holoprotein has been given the trivial name, phototropin 1 (phot1), in order to reflect its physiological and biochemical properties (Christie et al., 1999; Briggs et al., 2001b; Table 1).


Figure 6. In vitro phosphorylation of recombinant phot1 (nph1) and phot2 (npl1) proteins.

Arabidopsis phototropins expressed in insect cells were purified and subjected to an in vitro phosphorylation assay. Arrows denote positions of the phototropin proteins.

(Data are from Sakai et al., 2001)


Table 1. Phototropin nomenclature.
Wild-type genes
Mutant genes
phot1, phot2
phot1, phot2

The nomenclature presented here has been described by Briggs et al. (2001) and is based upon that adopted for the phytochromes (Quail et al., 1994). As presented in the text, PHOT1 was originally designated NPH1 (Liscum and Briggs, 1995), while PHOT2 was designated NPL1 (for NPH1-Like) (Jarillo et al., 1998).

Phototropin apoproteins and holoproteins differ only in that holoproteins have associated with them the two FMN chromophores, while apoproteins lack these co-factors. As such the holoproteins are the "photoactive" entity.


Phototropin 2 (phot2): a second phototropic receptor that functions redundantly with phot1 under high fluence rate conditions

While phot1 mutants clearly lack hypocotyl and root phototropism under low fluence rate blue light (e.g., < 1 mmol m-2 s-1), Sakai and colleagues (2000) reported that the phot1-101 allele retains a phototropic response at higher fluence rates - one essentially like wild-type at 100 mmol m-2 s-1. This finding, confirmed by studies of the phot1-5 null mutant exposed to directional light under greenhouse conditions (T. Campbell and E. Liscum, unpublished), indicated the function of a second phototropic receptor under high light conditions. The most obvious candidate for a second phototropic receptor apoprotein was the sole PHOT1 paralog PHOT2 (Jarillo et al., 1998; The Arabidopsis Genome Initiative, 2000; Briggs et al., 2001). Although the PHOT2 apoprotein is slightly smaller than PHOT1 (110 kD versus 124 kD), sequence and structural motifs are highly conserved. In particular, PHOT2, like PHOT1, contains a Ser/Thr protein kinase domain in its carboxyl terminal region and two LOV domains in its amino terminal half (Jarillo et al., 1998; see Figure 5b). Moreover, the basic photochemical properties of baculovirus/insect cell-expressed phot2 holoprotein are similar to those of phot1 (Sakai et al., 2001; see Figure 6 above), suggesting that phot2 likely functions as a blue-light receptor by a mechanism analogous to that of phot1.



Figure 5b. Aligned domain structures of phot1 and phot2.
FMN-binding LOV domains are shown in yellow. Kinase domains are shown in red.


While no alterations in phototropic responsiveness have been observed in phot2 single mutants (Sakai et al., 2000; Jarillo et al., 2001; Sakai et al., 2001), phot1 phot2 double mutants fail to exhibit seedling phototropic responses at both low and high fluence rates (Sakai et al., 2001). The fact that phot2 single mutants retain a phototropic response indistinguishable from wild-type under all fluence rates tested (Jarillo et al., 2001; Sakai et al., 2001) while the phot1 phot2 double mutant is essentially blind (Sakai et al., 2001), demonstrates that phot1 functions to some extent under all fluence rate conditions, while phot2 has redundant function for phot1 specifically under high fluence rate conditions.

How can we explain the overlapping, yet distinct, functions of the phot1 and phot2 receptors in the perception of phototropic stimuli? While there are currently no definitive answers to this question there are several pieces of data that suggest biochemical and molecular bases for this observation. First, while etiolated Arabidopsis seedlings do not exhibit acute light-dependent changes in PHOT1 mRNA abundance (Harmer et al., 2000; J. M. Christie, K. Sakamoto, and W. R. Briggs, personal communication), PHOT2 mRNA levels increase upon exposure to UV-A, blue, red, or white light (Jarillo et al., 2001; Sakai et al., 2001). Most interestingly, PHOT2 message levels increase more than two-fold at fluence rates of blue light (? 10 mmol m-2 s-1) where the redundant function of phot2 is most obvious (Sakai et al., 2001). Second, there are quite dramatic differences in the photocycling properties of the Cys-FMN bond formed within the LOV domains of phot1 and phot2 (Kasahara et al., 2002). In particular, there are differences between phot1 and phot2 in both the quantum efficiencies and dark-recovery kinetics for the formation of this covalent linkage suggesting that much higher fluences are required to drive phot2 to the same photoequilibrium established under lower fluences for phot1. Like the expression differences, these differing photochemical properties suggest that phot2 is likely to have similar physiological properties to phot1 (e.g., have redundant phototropic function) under relatively high light intensities.


What’s the role of phototropin kinase activity in phototropic signal-response?

Given that both phototropins are light-activated Ser/Thr protein kinases (Christie et al., 1998; Sakai et al., 2001) it is quite reasonable to hypothesize that phototropic signal transduction might involve phot1 and phot2-activated phosphorelays (Fankhauser and Chory, 1998; Briggs and Huala, 1999; Liscum and Stowe-Evans, 2000; Briggs and Olney, 2001; Christie and Briggs, 2001; Liscum, 2002). Yet to date, no protein kinases or phosphatases have been shown to associate with the phototropins. The only known kinase substrates for the phototropins are the phototropins themselves (Christie et al., 1998; Sakai et al., 2001). While the role of phototropin autophosphorylation is not currently understood it appears not to be required for the induction of phototropism since phototropism is 2-3 orders more sensitive to blue light than the autophosphorylation response (Palmer et al., 1993; Salomon et al., 1997; Christie et al., 1998).

If phototropin autophosphorylation is not an inductive signal, what is its function in phototropism? Given that the “dark-state”, or light-sensitive state, of the phototropins is the unphosphorylated form (Christie et al., 1998; Sakai et al., 2001), autophosphorylation may be a cue for desensitization of the receptors (Liscum, 2002). Sensor adaptation, where the receptor is desensitized after an initial perception event and then is “reset” so that future signals can be utilized as a signal, is a common, if not essential, feature of photosensory biology (Galland, 1989, 1991). In Arabidopsis phototropic desensitization occurs (Janoudi and Poff, 1991, 1993) but without reducing the sensitivity (i.e., increasing the fluence threshold) of the response (Janoudi and Poff, 1991). Thus, sensor adaptation, as classically defined (Galland, 1989, 1991), is apparently not operating in Arabidopsis phototropism. However, it is plausible that the phototropic desensitization observed in Arabidopsis reflects desensitization of a “receptor complex” rather than the receptor per se (Liscum and Stowe-Evans, 2000; Liscum, 2002). As discussed in the following section, blue light-dependent autophosphorylation of phot1 may lead to dissociation of a complex containing phot1 and NPH3, a putative scaffold/adapter protein (Motchoulski and Liscum, 1999).


Signal-response components acting downstream of the phototropins


NPH3 and RPT2: phot1- and phot2-interacting proteins that scaffold phototropin signaling complexes?

Mutations in two Arabidopsis genes have been described that apparently disrupt the function of proteins acting early in phot1 and phot2 phototropic signaling pathways. First, loss-of-function nph3 mutations appear to disrupt phot1 signaling since they decrease phototropic responses of etiolated seedlings specifically under low fluence rate conditions (Liscum and Briggs, 1996; Motchoulski and Liscum, 1999; see Figure 7), without altering blue light-induced autophosphorylation of phot1 (Liscum and Briggs, 1995; see Figure 4c and d). In contrast, the loss-of-function rpt2 (mutant designation, root phototropism) mutant retains nearly normal phototropism in response to unilateral low fluence rate blue light, but exhibits decreasing response with increasing fluence rates (Okada and Shimura, 1992; Sakai et al., 2000). While the phototropic phenotypes of rpt2 are essentially complementary to those of the nph3 mutants, they exhibit a fluence rate dependence that mirrors the apparent range of fluence rates under which phot2 functions (Sakai et al., 2000, 2001). The congruence between RPT2 and phot2 is also observed at the level of transcript abundance. Specifically, the RPT2 and PHOT2 transcripts are barely detectable in etiolated seedlings but increase in abundance in response to light with a similar fluence and wavelength dependencies (Sakai et al., 2000, 2001). Together with the mutant phenotype, this result suggests that RPT2 functions in a phot2-specific signaling pathway (Liscum, 2002).


Figure 7. Phototropic response of a nph3 mutant.
Seedlings were handled as described in Figure 3.


Figure 4c-d. Blue light-induced phosphorylation of the NPH1 protein in Arabidopsis.

Panel A, Autoradiograph of gel shown in B.

Panel B, Silver-stained gel of microsomal membranes isolated from various nph mutants (and their respective wild-type progenitors) and subjected to an in vitro phosphorylation assay with g32P-ATP. Arrow indicats the position of the NPH1 protein, which is phosphorylated in response to blue light in wild-type (Col and WS). Note that in contrast to what is observed in the phot1 mutants (see Figure 4a and b), nph2, nph3, and nph4 mutants all retain blue light-induced phosphorylation.

(Data are from Liscum and Briggs, 1995)


When the NPH3 and RPT2 genes were isolated by map-based cloning approaches they were found to encode members of the same family of novel plant-specific proteins (Motchoulski and Liscum, 1999; Sakai et al., 2000). Primary sequence conservation between members of the NPH3/RPT2 family is found in five discrete, positionally conserved, regions designated DIa, DIb, DII, DIII, and DIV (Figure 8). In addition to displaying modular sequence conservation members of the NPH3/RPT2 family exhibit modular structural features that are positionally conserved (although sequence diverged) (Figure 8). Two of these “modules” are represented by known protein-protein interaction motifs; a BTB (broad complex, tramtrack, bric à brac)/POZ (pox virus and zinc finger) domain (Albagli et al., 1995; Aravind and Koonin, 1999) in the amino-terminal region, and a coiled-coil domain (Cohen and Parry, 1990; Lupas, 1996) in the carboxyl-terminal region (Motchoulski and Liscum, 1999; Sakai et al., 2000). While no particular function can be inferred from the mere presence of a BTB/POZ or a coiled-coil domain, the fact that most of the NPH3/RPT2 family members contain one or both - 21 of 32 members, including NPH3 and RPT2, contain both, while only two contain neither (see our 2010 project site) - suggests that protein-protein interactions are an important feature of the biochemical function of this family.


Figure 8. "Consensus" features of the NPH3/RPT2 protein family.

Top, positions of conserved (white boxes, DIa-DIV) and variable (green areas) sequence. Middle, position of motifs identified by homology to known protein-protein interaction domains. Bottom, positions of strongly hydrophobic (white boxes) and predicted surface exposed (grey boxes) regions.


Sequence conserved domain DIV contains a consensus Tyr phosphorylation site ([RK]-x(2,3)-[DE]-x(2,3)-Y; Patchinsky et al., 1982) that is conserved in 29 of the 32 members of the family, including NPH3 and RPT2 (Motchoulski and Liscum, 1999; see Figure 9). Although it is not known whether phosphorylation can occur on the conserved Tyr of NPH3 or RPT2, it’s interesting to note that the nph3-2 allele, which carries in an in-frame deletion of this Tyr residue, is phenotypically indistinguishable from a null mutant (nph3-6) that contains a stop codon at the Trp2 position (Motchoulski and Liscum, 1999). NPH3 and RPT2 also contain multiple potential Ser and Thr phosphorylation sites, further raising the possibility that reversible phosphorylation may be important for their function (Motchoulski and Liscum, 1999; Sakai et al., 2000; also see our 2010 project site).


Figure 9. Amino acid alignment of the putative Tyr phosphorylation site and "DGLYRAID" motif within DIV of the NPH3/RPT2 family.
The consensus Tyr phosphorylation site sequence, as well as the majority sequence across the motif are shown above the alignments. Sequences are given in the order presented in the phylogenetic tree (see our 2010 project page). One member, At3g49900, is not shown since it lacks DIV. Amino acid identities are shaded yellow, while conservative substitutions are shaded grey.

Go to our Arabidopsis 2010 Project page for more details about the NPH3/RPT2 family.


NPH3, like phot1 (Briggs and Huala, 1999), has been shown to be associated with the plasmalemma (Motchoulski and Liscum, 1999; see Figure 10a). This property, along with fact that NPH3 contains multiple potential phosphorylation sites, suggested that NPH3 might be a substrate for phot1’s kinase activity. NPH3 does appear to be a phosphoprotein, however it is dephosphorylated, rather than phosphorylated, in response to blue light irradiation (Motchoulski and Liscum, 1999; see Figure 10b and c). Moreover, NPH3 seems to be dephosphorylated in etiolated phot1 null mutant seedlings indicating that the protein phosphatase that dephosphorylates NPH3 in wild-type seedlings is not directly regulated by light or phot1 (Motchoulski and Liscum, 1999; Figure 10b and c). Thus it appears that NPH3 is not a substrate for phot1’s kinase domain.


Figure 10a-c. Plasma membrane localization and in vivo modification of NPH3.

A, Immunoblot analyses of NPH3 (left two panels) and phot1 (right panel) proteins in different cellular fractions isolated from dark-grown seedlings. Sol, soluble proteins; MM, microsomal membranes; PM, plasmamembrane subfraction. Arrow indicates the position of the NPH3 protein. Star indicates the position of the phot1 protein. Although the NPH3 and phot1 signals in the MM and PM fractions look equivalent only 1/10th the amount of protein loaded for the MM fraction is loaded for the PM sample.

B, In vivo modification of NPH3. Three-day-old dark-grown seedlings were mock irradiated (D) or exposed to red (R) or blue light (B) prior to cell fractionation. With the exception of the B-Sol lane (Soluble proteins from blue light-treated seedlings) all lanes are microsomal fractions. Arrow indicates the position of the "unmodified" NPH3 protein, while the asterisks indicate "modified" NPH3.

C, In vivo modification of NPh3 appears to at least in part reversible phosphorylation. Seedlings were grown and light treated as in B, with the exception of two samples (lanes 3 and 4) where seedlings were treated with the phosphatase inhibitor okadiac acid (OKA) prior to cell fractionation. Note that the higher mobility NPH3 species observed in blue light (B)-treated seedlings is replaced by the lower mobility species upon OKA treatment.

(Data in panels A and B are from Motchoulski and Liscum (1999), while those in panel C are unpublished)

One interpretation of the aforementioned results is that phot1 normally interacts with NPH3 in etiolated seedlings, thus “protecting” NPH3 from protein phosphatase action. Blue light-dependent changes in phot1, such as autophosphorylation (Christie et al., 1998, 1999), might disrupt this interaction, exposing sites on NPH3 for dephosphorylation. Yeast two-hybrid and in vitro co-immunoprecipitation studies have demonstrated that carboxyl-terminal coiled-coil-containing portions of NPH3 interact with amino-terminal LOV-domain-containing portions of phot1, consistent with the proposal that phot1 and NPH3 are components of a signaling complex (Motchoulski and Liscum, 1999). It is worth mentioning in the context of the model outlined above that the phot1-NPH3 interaction in yeast is much stronger in darkness than in blue light (A. Motchoulski and E. Liscum, unpublished). These findings suggest, as introduced earlier, that phototropic “sensitivity” in low fluence rate blue light conditions may be regulated in part by the integrity of an NPH3-phot1 complex. In other words phototropic desensitization may occur through complex dissociation upon blue light-induced phot1 autophosphorylation.

NPH3 has been proposed to act as a modular scaffold bringing phot1 together with early acting proteins (e.g., protein kinases or phosphatases) whose activities are modulated by phot1 to transduce the low fluence rate phototropic signals (Motchoulski and Liscum, 1999; Liscum and Stowe-Evans, 2000). Based on its mutant phenotypes and homology to NPH3, RPT2 is proposed to scaffold phot2 and associated early signaling proteins in a situation analogous to that proposed for NPH3 and phot1 (Liscum, 2002). The assembly of multimolecular complexes on protein scaffolds has emerged as a common mechanism of optimizing speed, specificity, and selectivity of signaling in fungi and animals (Elion, 1998; Faux and Scott, 1996; Tsunoda et al., 1998; Sim and Scott, 1999; Fisher et al., 1999). One of the “pioneer” scaffold proteins is Ste5 (Sterile5) of S. cerevisae that is involved in mating-pheromone sensing and response. Ste5 scaffolds a functional signaling complex that includes the G-protein pheromone sensor (via the Gb subunit, Ste4), a sensor-activated protein kinase (Ste20), and a signal amplifying MAP kinase cascade (MAPKKK, Ste11; MAPKK, Ste7; MAPK, Fus3) (Faux and Scott, 1996; Pawson and Scott, 1997). Despite being divergent in both sequence and structure, scaffold proteins are able to assemble signaling complexes via presence of multiple protein-protein interaction domains, or modules (Faux and Scott, 1996; Newton, 1996; Pawson and Scott, 1997; Tsunoda et al., 1998).
NPH3 and RPT2 are certainly modular in nature and contain two known protein-protein interaction domains: a BTB/POZ domain and a coiled-coil (Motchoulski and Liscum, 1999; Sakai et al., 2000). Scaffold proteins have also most frequently been found associated with complexes that utilize protein kinases and phosphatases as signal carriers (Faux and Scott, 1996; Pawson and Scott, 1997; Sim and Scott, 1999). The experimental demonstration that NPH3 interacts with phot1 (Motchoulski and Liscum, 1999) and likelihood that RPT2 interacts with phot2 (Liscum, 2002), are compelling since the phototropins are both sensors and protein kinases (Christie et al., 1998, 1999). Correct or incorrect, the scaffold model provides the impetus for several experimentally addressable questions.

Differential auxin localization

Most mechanistic models of tropic growth responses, including phototropism, have auxin as a linchpin that holds pieces of the model together (Estelle, 1996; Chen et al., 1999; Liscum and Stowe-Evans, 2000). Nearly all such models are based in large part on the Cholodny-Went theory which holds that tropic stimuli induce differential lateral auxin transport that leads to the unequal distribution of auxin, and hence growth, in the two sides of a curving organ (Went and Thimann, 1937). While lateral auxin transport, or accumulation, has proven remarkably difficult to demonstrate in many systems (Trewavas et al., 1992), studies of a number of Arabidopsis mutants have clearly demonstrated that the transport of, and response to, auxin is prerequisite for the development of tropic curvatures (Estelle, 1996; Leyser, 1998; Chen et al., 1999; Palme and Gälweiler, 1999; Rosen et al., 1999; Liscum and Stowe-Evans, 2000; Liscum, 2002; Friml and Palme, 2002). Most recently Friml and colleagues (2002) have shown that a member of the PIN family of putative polar auxin efflux carriers, PIN3, is localized to the lateral face of endodermal cells in hypocotyls and non-functional alleles carrying transposon insertions reduce the phototropic response of the seedling stem.

Several studies have shown that auxin efflux processes, which are necessary for the establishment of a lateral gradient of auxin (Lomax et al., 1995; Friml et al., 2002), may be regulated via reversible protein phosphorylation (Bernasconi, 1996; Garbers et al., 1996; Delbarre et al., 1998; Christensen et al., 2000; Rashotte et al., 2001; Delong et al., 2002). These findings provide a compelling potential connection between the phototropins, which are light-activated protein kinases (Christie et al., 1998; Sakai et al., 2001), and the differential auxin gradients that have been observed in seedlings irradiated with unilateral blue light (Iino, 1990). A number of possible biochemical pathways from phototropin activation to changes in auxin transport can be postulated, including direct phosphorylation of an auxin efflux carrier by phototropins, and phototropin-dependent modulation of auxin transporter localization (Liscum, 2002). Determining exactly how phototropin activation leads to changes in auxin transportor activity/localization is likely to be one of the most active areas of phototropic research over the next few years.

Differential auxin response: changes in gene expression

What happens after a lateral auxin gradient is formed across the phototropically stimulated stem? At least one of the consequences appears to be a change in gene expression (Harper et al., 2000; Liscum and Stowe-Evans, 2001; Liscum and Reed, 2002). This conclusion is based mainly on analyses of the NPH4/MSG1/TIR5 locus (hereafter referred to as NPH4) of Arabidopsis. First, mutations in the NPH4 locus cause reduced phototropic and gravitropic responses in seedlings stems (Liscum and Briggs, 1996; Watahiki and Yamamoto, 1997; Watahiki et al., 1999), as well as severely impaired auxin-induced stem bending (Watahiki and Yamamoto, 1997), stem growth inhibition (Watahiki and Yamamoto, 1997; Stowe-Evans et al., 1998), and gene expression responses (Stowe-Evans et al., 1998). Second, map-based cloning of the NPH4 gene has revealed that it encodes the auxin-responsive transcriptional activator ARF7 (Harper et al., 2000), consistent with impaired auxin-induced gene expression profiles observed in the nph4/arf7 mutants (Stowe-Evans et al., 1998). Unfortunately, to date, specific targets of NPH4/ARF7 that are necessary for the establishment of phototropic responses have not been identified. However, with the functional genomics tools now at our disposal targets for NPH4/ARF7 regulation should not remain elusive for long.