Arabidopsis NPH3: A NPH1 Photoreceptor-Interacting Protein Essential for Phototropism

See allHide authors and affiliations

Science  29 Oct 1999:
Vol. 286, Issue 5441, pp. 961-964
DOI: 10.1126/science.286.5441.961


Phototropism of Arabidopsis thalianaseedlings in response to a blue light source is initiated by nonphototropic hypocotyl 1 (NPH1), a light-activated serine-threonine protein kinase. Mutations in three loci [NPH2, root phototropism 2 (RPT2), and NPH3] disrupt early signaling occurring downstream of the NPH1 photoreceptor. TheNPH3 gene, now cloned, encodes a NPH1-interacting protein. NPH3 is a member of a large protein family, apparently specific to higher plants, and may function as an adapter or scaffold protein to bring together the enzymatic components of a NPH1-activated phosphorelay.

Plants are able to sense and respond to changes in light quality, quantity, and direction through the action of a number of photoreceptors and associated signal-response systems. A variety of photoreceptor molecules, such as the red or far-red light–absorbing phytochromes (phy) and blue light–absorbing cryptochromes (cry) (1, 2) have been characterized. Components functioning downstream of such photoreceptors include two photoreceptor-interacting proteins PIF3 and PKS1, which are both phytochrome-interacting proteins (3). Potential postperception signaling components identified through mutational analyses (1, 2) include a set of genes inArabidopsis (NPH2, RPT2, andNPH3) that is required for phototropism, or the bending response of plant organs toward or away from directional light stimuli (4, 5).

Null mutations in the NPH3 locus abolish phototropic responses of etiolated seedlings to blue light at a low fluence rate (4) (Fig. 1A). A similar aphototropic phenotype is observed with seedlings carrying null mutations at the NPH1 locus (4) (Fig. 1A), which encodes NPH1, a primary photoreceptor for phototropism (6–8). However, mutations in neitherNPH1 nor NPH3 affect other light-dependent responses, such as blue light– and red light–dependent hypocotyl growth inhibition (Fig. 1, B and C). Moreover, despite efforts to identify additional alterations in development, the only clear phenotypic changes observed in nph1 and nph3mutants are those associated with phototropism (9). The phototropism-specific phenotypes of the nph1 andnph3 mutants imply that NPH1 and NPH3 act in the same genetic pathway and suggest that NPH3 may function biochemically close to the photoperception event mediated by NPH1.

Figure 1

Physiological characteristics ofnph1 and nph3 mutants. (A) Hypocotyl phototropism in 3-day-old etiolated wild-type (WT) and mutant seedlings exposed to 8 hours of unilateral blue light (25). (B) Blue light– and (C) red light–dependent hypocotyl growth inhibition (25) in wild-type and mutant seedlings. The cry1-101 (26) and phyB-9 (27) mutants are shown as negative controls for (B) and (C), respectively. Because symbols often overlap, some symbols and error bars (standard error) are not visible.

We have cloned the NPH3 gene by positional cloning (Fig. 2A). Several predicted genes mapping to the region containing the NPH3 locus (Fig. 2A) were sequenced in mutant backgrounds to identify the NPH3gene (Fig. 2B). A full-length NPH3 cDNA was isolated (10) and found to encode a protein of 745 amino acids (11) (Fig. 2C). NPH3 is, however, part of a family of proteins in Arabidopsis (12). AlthoughNPH3-related sequences have also been found in other plant species, no paralogous sequences have been found outside the plant kingdom (9).

Figure 2

NPH3 cloning, gene structure, and amino acid sequence. (A) Mapping of NPH3 on chromosome 5. Initial mapping placed NPH3 between simple sequence length polymorphism (SSLP) marker nga129 (28) and restriction fragment length polymorphism marker g2368 (29). Sequence of phagemid P1 clone, MSJ1 (AB008268), was used to generate two new SSLP markers, AM40 and AM80 (30), which were used to narrow the potential NPH3 coding region to three predicted genes: msj1.9, msj1.10, and msj1.11 (31). msj1.10 was identified as NPH3 by sequencing multiple nph3mutant alleles (24). T3J14 is a BAC clone (22). Numbers of recombinants for each marker are shown on the bottom. cM, centimorgan. (B) Structure of the NPH3 gene and positions of nph3 mutations. The locations of start (ATG) and stop (TGA) codons are indicated. Exon (boxes) and intron (lines) positions were determined by a comparison of the genomic and cDNA sequences. Position and identity of various nph3mutations are indicated. (C) Deduced amino acid sequence (13) of NPH3. Amino acid residues are numbered at the right. Bold-faced residues indicate regions of sequence conserved across the NPH3 family (12). Motifs of highest sequence conservation are boxed. The BTB/POZ and coiled-coil domains are underlined with single and double lines, respectively. Locations ofnph3 mutations are indicated by asterisks. In thenph3-1 allele, amino acids 272 to 295 are replaced with residues (LHIGYFSLKHFHKHLVYISQSSIH) before insertion of a stop codon at position 296.

Four regions of sequence conservation have been identified within the NPH3 family (12) (Fig. 2C). Region IV exhibits the highest level of sequence identity within the family (≥52.5%, relative to NPH3), with two motifs being most prominent: LYRAID and HAAQNERLPL (13) (Fig. 2C). The functional importance of these conserved sequence motifs is currently unknown. However, Tyr545 within the LYRAID motif is part of a consensus phosphorylation site ([RK]-x(2,3)-[DE]-x(2,3)-Y; where x(2,3) can be any two or three amino acids) (14), in which the Arg, Asp, and Tyr are invariant across the entire NPH3 family (12), and deletion of this residue in the nph3-2mutant (Fig. 2B) results in a complete loss of phototropic responsiveness (4) (Fig. 1A). NPH3 also contains multiple potential Ser or Thr phosphorylation sites (9), raising the possibility that reversible phosphorylation may be important for its function.

Primary sequence and secondary structure analyses have also identified two potential protein-protein interaction domains in NPH3. First, a BTB (broad complex, tramtrack, andbric à brac)/POZ (pox virus and zinc finger) domain (15, 16) was found within the NH2-terminal portion of NPH3 (Fig. 2C), as well as within two additional NPH3 family members (12). Second, a coiled coil (15, 17) was identified in the COOH-terminus of NPH3 (Fig. 2C), and although sequence divergent, it was also found in a conserved position in eight additional members of the family (12). BTB/POZ domains and coiled coils have been found in a large number of proteins with quite disparate functions, and both are capable of mediating self-dimerization and self-multimerization or hetero-dimerization and hetero-multimerization with unrelated domains (16, 17).

Although NPH3 is largely hydrophilic and α helical in nature (18) and thus anticipated to be soluble, cell fractionation studies indicate that NPH3 is plasmalemma associated (Fig. 3A). Whereas the membrane localization does not change in response to in vivo irradiation, the mobility of NPH3 in SDS–polyacrylamide gel electrophoresis (SDS-PAGE) is enhanced in membrane fractions from seedlings that were irradiated with blue, but not red, light (Fig. 3B). A similar enhanced mobility of NPH3 was observed in membranes from nph1-5 mutant seedlings (Fig. 3B) that were grown in the dark and lack NPH1 (6,7). One interpretation of these observations is that NPH3 normally interacts with NPH1 in seedlings that were grown in the dark, either directly or indirectly through a protein complex, preventing its modification by other factors, whereas blue light–induced changes in NPH1, such as autophosphorylation (7), or simple removal of NPH1, expose sites on NPH3 for modification.

Figure 3

Plasma membrane localization and in vivo modification of NPH3. (A) Immunoblot analysis of NPH3 localization after cell fractionation (32) of 3-day-old seedlings that were grown in the dark. Samples containing 50 μg of soluble (Sol), 15 μg of microsomal membrane (MM), or 1.5 μg of plasma membrane (PM) protein were resolved by SDS-PAGE and blotted to polyvinylidene difluoride (PVDF) membrane (Bio-Rad) for subsequent immunodetection of NPH3 (left two panels) and NPH1 (right panel) proteins (21, 33). nph3-1 microsomal membranes represent a negative control for NPH3 antibody specificity, because any NPH3 protein in this mutant lacks the COOH-terminus (seeFig. 2B), which was used as antigen to generate the antisera (33). Molecular masses (in kilodaltons) are indicated at the left. Arrow and asterisk indicate the position of the NPH3 and NPH1 proteins, respectively. (B) In vivo modification of NPH3. Three-day-old seedlings that were grown in the dark were mock irradiated (D), or exposed to 103.5 μmol m–2of red (R) or blue (B) light (25), before the preparation of cell fractions (32). Fifty micrograms of soluble protein (B-Sol lane) or 15 μg of microsomal membrane protein (all other lanes) were resolved with SDS-PAGE and blotted to PVDF membrane; and NPH3 protein was visualized by immunodetection (23). Molecular masses (in kilodaltons) are indicated at the left. Arrow and asterisks (double asterisk indicates a change specific to thenph1-5 genotype) indicate the positions of the “unmodified” and “modified” NPH3, respectively, relative to NPH3 from seedlings that were grown in the dark. Although increased mobility of NPH3 in membranes of blue light–treated wild-type seedlings appears to result from dephosphorylation, limited proteolysis may also contribute to mobility changes observed innph1-5 membranes (9).

Given the apparent genetic coupling of their protein activities (4) (Fig. 1), colocalization to the plasmalemma (6, 7) (Fig. 3A), and potential interaction suggested by in vivo irradiation studies (Fig. 3B), we used a yeast two-hybrid assay (19) to directly test whether NPH1 and NPH3 physically interact. Although all combinations of NPH1 and NPH3 fragments shown in Fig. 4A were tested, significant interactions were only observed between coiled-coil–containing NPH3 polypeptide fragments and the NH2-terminal two-thirds of NPH1 (Fig. 4B), which contains the chromophore-binding LOV (light, oxygen, or voltage) domains (6,7). Two hybrid results were confirmed by in vitro interaction studies, in which we assayed the ability of a radiolabeled NH2-terminal NPH1 fragment to be immunoprecipitated by antibodies against a cellulose binding domain (CBD), in the absence or presence of nonlabeled CBD-NPH3 fusion proteins (Fig. 4C). Although CBD antibodies were incapable of precipitating NPH1 itself, a NPH1-containing immunoprecipitate was obtained when coiled-coil–containing NPH3-CBD fusion protein (NPH3C2-CBD) was included in the precipitation reaction (Fig. 4C). The NPH3C2-CBD–dependent immunoprecipitation of NPH1 was not the result of nonspecific interaction with CBD, as a BTB/POZ-containing NPH3-CBD fusion protein did not promote a similar immunoprecipitation response (Fig. 4C). The flavin mononucleotide (FMN)–dependent interaction of NPH1 and NPH3 (Fig. 4C) likely reflects a conformational stabilization of the NPH1 holoprotein in relation to the apoprotein alone, a feature common to many cofactor-containing proteins, including flavin-binding proteins (20).

Figure 4

NPH1-NPH3 interaction. (A) Diagrammatic representations of NPH1 and NPH3. The NPH1 protein kinase and LOV domains (6) are shown as solid and cross-hatched blocks, respectively. The four conserved sequence domains (I to IV) of the NPH3 family are shown on NPH3 by cross-hatched blocks, and the BTB/POZ and coiled-coil domains are shown as solid blocks. The shaded cross-hatched region of NPH3 indicates where the BTB/POZ domain and domain I overlap. Portions of NPH1 and NPH3 used for interaction studies are indicated below each block diagram. Amino acid residues (positions within full-length proteins) used for each construct are shown in parentheses. (B) Yeast two-hybrid assay of NPH1-NPH3 interaction. All combinations of NPH1 and NPH3 polypeptides shown in (A) were initially tested for β-galactosidase (β-Gal) activity by colony filter-lift assay (34); however, only NPH1LOV-NPH3C1, NPH1LOV-NPH3C2, and NPH1LOV-NPH3N combinations gave a positive reaction (9). Hence, quantitative solution assays of β-Gal activity (34) are presented (right) only for those cotransformants. GBD-NPH1 and GAD-NPH3 constructs (34) are shown at the left. The GBD-NPH1LOV/GAD cotransformant is shown as a negative control. All yeast growth occurred in darkness. One unit of β-Gal activity is defined as the amount of enzyme that converts 1 nmol ofo-nitropheynl-β-d-galactopyranoside too-nitrophenol in 1 min at 30°C (pH 7.0); 1 nmol/ml ofo-nitrophenol has an optical density at 420 nm of 4.5 × 10–3 (35). Protein contents of assay solutions were determined with the Bio-Rad microassay. Significant interactions (0.1 > *P > 0.05, Students t test) and number of replicate cotransformants assayed are indicated; error bars indicate SD. (C) In vitro assay of NPH1-NPH3 interaction. CBD-NPH3C2 and CBD-NPH3N fusion proteins were generated as described (33). Sulfur-35–labeled NPH1LOV was generated by in vitro transcription and translation with the pT7Blue-2Ek/LIC system (Novagen). An equal volume (15 μl) of 35S-labeled NPH1LOV from the same in vitro translation (60 μl total volume) was used in all immunoprecipitation reactions. Where indicated with a plus, 2 μg of NPH3C2-CBD protein, 2 μg of NPH3N-CBD protein, or 1 μM (final concentration) FMN were added to reactions before the addition of antibody. Polyclonal antibodies against CBD (Novagen) were added to each reaction, immunoprecipitates were resolved with SDS-PAGE, and gels were dried and exposed to x-ray film. All steps up to SDS-PAGE were done in dim red light (4). The position of NPH1LOV is indicated by an arrow. Molecular masses (in kilodaltons) are indicated at the left. Although not shown, similar results were obtained in each of three replicate experiments.

Together, all of the results presented here indicate that NPH3 interacts with the chromophore-binding portion of NPH1, probably through its coiled-coil region, and that this protein complex is necessary for early phototropic signaling. The biochemical function of NPH3, however, remains unknown. In animal and fungal systems, multimolecular signaling complexes using protein kinases and phosphatases are often assembled around a class of proteins known as adapter or scaffold proteins, which contain multiple protein-protein interaction domains (21). NPH3 contains at least two protein-protein interaction domains, one of which promotes interaction with NPH1, and early phototropic signaling probably occurs through a NPH1-activated phosphorelay (2, 4). Thus, NPH3 may function as an adapter or scaffold protein in plants. Given their homologies to NPH3, other members of the NPH3 family may also function as adapter or scaffold proteins. Protein complexes assembled by the NPH3 family could provide a means of optimizing speed, specificity, and selectivity of early signaling events associated with a variety of physiological responses, while minimizing undesired cross talk with other response pathways. The identification of such a class of proteins could have broad implications for studies of plant signal transduction.

  • * To whom correspondence should be addressed. E-mail: liscumm{at}


View Abstract

Navigate This Article