PKS1, a Substrate Phosphorylated by Phytochrome That Modulates Light Signaling in Arabidopsis

See allHide authors and affiliations

Science  28 May 1999:
Vol. 284, Issue 5419, pp. 1539-1541
DOI: 10.1126/science.284.5419.1539


Plants constantly monitor their light environment in order to grow and develop optimally, in part through use of the phytochromes, which sense red/far-red light. A phytochrome binding protein, PKS1 (phytochrome kinase substrate 1), was identified that is a substrate for light-regulated phytochrome kinase activity in vitro. In vivo experiments suggest that PKS1 is phosphorylated in a phytochrome-dependent manner and negatively regulates phytochrome signaling. The data suggest that phytochromes signal by serine-threonine phosphorylation.

Phytochromes are red/far-red photoreceptors that regulate many aspects of plant development (1). The prototypical phytochrome is a homodimer consisting of 125-kD subunits. The protein is composed of two major domains: an NH2-terminal chromophore-binding domain that is sufficient for normal photochemistry, and a COOH-terminal domain required for dimerization and thought to be a signal output domain (2). Phytochromes are localized in the cytoplasm or the nucleus, depending on the light conditions and the type of phytochrome (3). Phytochromes are synthesized in a red light–absorbing form (Pr) in the dark; upon absorption of red light, they convert to the far-red light–absorbing form (Pfr). Pfr is thought to be the active form because for many phytochrome-mediated responses, far-red light can dampen the response by converting Pfr back to Pr.

In Arabidopsis, each of the five phytochromes (phyA through phyE) has both unique and redundant functions (4). Analysis of mutants has demonstrated that the COOH-terminus is essential for phytochrome function in vivo, but the biochemical mechanism of phytochrome signaling remains unclear (2). Although cyanobacteria has a phytochrome with light-regulated histidine kinase activity (5), the COOH-terminal domain of plant phytochromes is only weakly related to histidine kinases (6,7). Biochemical data point toward serine-threonine kinase activity associated with phyA, and autophosphorylation of recombinant oat phyA is light-regulated (7). Phytochromes might therefore belong to a growing class of eukaryotic serine-threonine kinases with histidine kinase ancestry (7,8).

To investigate the mechanism of phytochrome signaling, we conducted a yeast two-hybrid screen to identify proteins that interact with the COOH-terminal domain of phyA (9). This screen yielded a clone encoding the COOH-terminus of PHYA (10), which confirmed the ability of this portion of phyA to dimerize. We also identified three other genes capable of specific interaction with phytochromes; one of these, PKS1, is the subject of this report. Yeast two-hybrid experiments showed that the PKS1 protein interacts with the COOH-termini of both PHYA and PHYB (Fig. 1A). These interactions were confirmed using full-length phyA or phyB purified from yeast and in vitro transcribed and translated PKS1 (11). Although Pr and Pfr, the two spectral forms of phytochrome, have different protein conformations, PKS1 appears to bind both forms (Fig. 1B).

Figure 1

PKS1 interacts with phytochrome. Yeast was grown on plates without leucine (−LEV) to check for leucine auxotrophy and on plates containing 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (XGAL) to check for LacZ activity. (A) Two-hybrid interaction of PKS1 with the COOH-termini of PHYA and PHYB. (B) In vitro interaction of PKS1 with full-length phyA and phyB in Pr form (PrA, PrB) or in Pfr form (PfrA, PfrB). apoA refers to PHYA protein without chromophore. phyA and phyB were expressed as GST fusion proteins (phyAGST, phyBGST). Coomassie blue staining and autoradiography of the same gel are shown.

PKS1 encodes a basic soluble protein of 439 amino acids with no recognizable sequence motifs or signals for subcellular targeting (12) (Fig. 2A). Fluorescence microscopy showed that PKS1-GFP (green fluorescent protein) fusions are cytoplasmic (10). PKS1 mRNA is expressed in young seedlings in both darkness and light; later in development,PKS1 mRNA amounts decrease, with moderate expression in rosette leaves and very low expression in roots and flowers (Fig. 2B). This distribution pattern overlaps with the mRNA distribution ofPHYA-PHYE in Arabidopsis (13).

Figure 2

Amino acid sequence and expression pattern of PKS1. (A) Amino acid sequence of the PKS1 gene (12). This sequence has been deposited in GenBank (accession number AF149053). (B) Northern blot analysis ofPKS1 expression with 10 μg of total RNA for each lane.

PKS1 interacts with the COOH-termini of phyA and phyB in the region of highest homology with histidine kinases. Because oat phyA autophosphorylates on Ser/Thr (7) and because yeast two-hybrid screens have identified substrates for known protein kinases (14), we tested whether PKS1 is a substrate for the kinase activity of phytochrome in vitro. Recombinant oat phyA phosphorylated a glutathione S-transferase (GST)–PKS1 fusion protein (15). These experiments show that the extent of GST-PKS1 phosphorylation by phyA was light-regulated, with GST-PKS1 being a better substrate for Pfr's kinase activity than for Pr's kinase activity (Fig. 3A). Phytochrome autophosphorylation and PKS1 phosphorylation were stimulated by red light by a factor of 2 to 2.5. The residues phosphorylated on PKS1 were Ser and to a low extent Thr (10). Ser599 of oat phyA is phosphorylated preferentially as Pfr in vivo (16). When a mutant form of phyA (Ser599 → Lys, or S599K) was used for in vitro kinase assays, both autophosphorylation and phosphotransfer to PKS1 were no longer light-stimulated (Fig. 3A).

Figure 3

Light-regulated PKS1 phosphorylation by phyA in vitro. Each kinase assay used 1 μg of oat phyA and 1 μg of GST-PKS1 or PKS1 alone (15). (A) GST-PKS1 is phosphorylated in a light-dependent fashion by oat phyA in vitro. Amounts of phyA phosphorylation are expressed relative to lane 1 [dark (D)]; amounts of phyA S599K phosphorylation are expressed relative to lane 5 [far-red (FR) light, without PKS1]; and amounts of PKS1 phosphorylation are expressed relative to lane 1 (D, wild-type phyA). (B) PKS1 is phosphorylated in vivo. Immunoprecipitated PKS1 treated with or without CIP in the presence or absence of phosphatase inhibitors was subjected to protein immunoblotting and detected with antibodies to PKS1. (C) PKS1 is modified by red light, and this modification is enhanced in lines overexpressing phyB. Protein immunoblots of wild-type (Col) and phyB-overexpressing (B-OX) lines grown in the dark (D) or in red light (R) were probed with antibodies to PKS1. The same lines grown in the dark were also probed with an antibody to phyB.

Thus, phyA exhibits both light-regulated autophosphorylation and PKS1 phosphotransferase activities. In vitro, phytochrome phosphorylates substrates such as histone H1, cry1, and cry2 (17), but not in a light-regulated manner. Our observation of light-modulated phosphorylation in vitro suggests that PKS1 is a substrate of phytochrome's kinase activity in vivo. Analysis of PKS1 immunoprecipitates of light-grown seedlings with or without phosphatase treatment indicated that PKS1 is a phosphoprotein in vivo (18) (Fig. 3B). We have been unable to detect phyA or phyB coimmunoprecipitation under our assay conditions (10).

To assess whether PKS1 is phosphorylated in a light-dependent fashion, we performed PKS1 protein immunoblot analysis on dark-grown versus red light–grown seedlings (19). In the dark, PKS1 is mainly present as a faster migrating form, whereas in red light both slow- and fast-migrating forms of PKS1 can be seen. In red light–grown Arabidopsis plants that are overexpressing phyB, PKS1 is mainly present as a slower migrating form. Thus, PKS1 is phosphorylated in vivo in a phytochrome-dependent manner (Fig. 3C).

PKS1 maps to the top of chromosome II on bacterial artificial chromosome T17M13. No known photomorphogenic mutant maps to this region of the genome, and screening various T-DNA insertion collections did not yield any mutants disrupting the gene (10). Of 50 PKS1 antisense lines analyzed by protein immunoblotting, nine had reduced amounts of PKS1 protein without affecting the plants' phenotypes (10), perhaps as a result of redundancy with PKS1-related sequences such as EST (T13904), which codes for an open reading frame 55% identical to PKS1 over the entire protein (10).

To assess whether PKS1 might be sufficient (if not essential) for an altered light response, we generated lines overexpressing PKS1 (20) (Fig. 4C). These lines had an elongated hypocotyl when grown in white light (Fig. 4, A and B); this phenotype is reminiscent of phytochrome-deficient mutants and is indicative of reduced sensitivity to light. Growth in different monochromatic lights indicated that PKS1-overexpressing lines were less sensitive to red light and retained normal sensitivity to blue and far-red light (Table 1). Because phyB is the major photoreceptor for red light, these results are consistent with PKS1 acting as an inhibitor of phyB signaling. In accordance with this interpretation, the PKS1 overexpression phenotype is only 28% as effective in a phyB mutant background, relative to the effect in the wild type (10). Interestingly, when PKS1 is overexpressed in phyA mutants, the phenotype is more pronounced than in a wild-type background (Fig. 4, A and B). In white light, the role of phyA in hypocotyl elongation is only uncovered inphyB mutants; thus, the PKS1 overexpression phenotype inphyA mutants is consistent with PKS1 being an inhibitor of phyB signaling.

Figure 4

Seedlings overexpressing PKS1 have reduced sensitivity to white light. This phenotype is more pronounced in aphyA-205 background. We obtained 12 lines with increased amounts of PKS1 in a wild-type Col background and 4 lines inphyA-205. All lines have a longer hypocotyl in white light. Two representative lines in each background are shown. C, control lines transformed with an empty vector. (A) Hypocotyl elongation phenotype. (B) Quantification of hypocotyl length; average length (n = 16) and SE are indicated. (C) Protein immunoblot analysis of seedlings overexpressing PKS1 probed with a PKS1 antibody.

Table 1

Arabidopsis seedlings overexpressing PKS1 have a reduced sensitivity to red light. Seedlings were grown and measured as described (21). Mean hypocotyl lengths ± SE are indicated.

View this table:

Our results suggest that plant phytochromes signal through light-regulated protein kinase activity. Although the amino acid sequence of PKS1 does not suggest a biochemical function, these results confirm that oat phyA is an atypical serine-threonine kinase (7). Other prokaryotic-like histidine kinases found in eukaryotes phosphorylate Ser and Thr rather than His or Asp (8, 21). However, phytochrome phosphorylation may pass through a less stable phospho-His or phospho-Asp intermediate, as seen with mitochondrial α-ketoacid dehydrogenase kinase phosphorylation (8).

PKS1 appears not to associate differently with the two spectral forms of phytochrome; however, phosphorylation might be an important regulatory step of this interaction. Although the exact function of PKS1 remains unknown, the phenotype of PKS1-overexpressing lines suggests that PKS1 is a negative regulator of phyB signaling. PKS1 might modulate phyB kinase activity or subcellular localization. Alternatively, it may coordinate phyA and phyB function or inhibit phytochrome signaling by titrating out a positively acting factor.

There are precedents for negative regulators of a protein kinase being substrates of that same protein kinase. For example, Far1 and Sic1 are inhibitors of Cdc28 protein kinase activity that are degraded after phosphorylation by Cdc28 (22). Alternatively, PKS1 may promote segregation of PfrB to the nucleus, much as I-κB retains NF-κB in the cytoplasm until appropriate signals dissociate this interaction and expose NF-κB's nuclear localization signals (23). Light-regulated cytoplasmic sequestration of phytochromes by PKS1 represents an attractive model because of phytochrome's newly postulated role in the nucleus (3,24).

The mechanism by which phytochromes affect numerous developmental responses has long eluded us. Phytochromes could phosphorylate multiple substrates present in various subcellular locations, in tissues, and during different developmental stages. Alternatively, these different responses could result from branching later in the signaling pathways. Determining the number of substrates phosphorylated by phytochromes is one way of answering this important question.

  • * Present address: Monsanto Company, 700 Chesterfield Parkway, North St. Louis, MO 63198, USA.

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


View Abstract

Navigate This Article