Arabidopsis NPH1: A Protein Kinase with a Putative Redox-Sensing Domain

See allHide authors and affiliations

Science  19 Dec 1997:
Vol. 278, Issue 5346, pp. 2120-2123
DOI: 10.1126/science.278.5346.2120


The NPH1 (nonphototropic hypocotyl 1) gene encodes an essential component acting very early in the signal-transduction chain for phototropism. Arabidopsis NPH1 contains a serine-threonine kinase domain and LOV1 and LOV2 repeats that share similarity (36 to 56 percent) with Halobacterium salinariumBat, Azotobacter vinelandii NIFL, Neurospora crassa White Collar–1, Escherichia coli Aer, and the Eag family of potassium-channel proteins from Drosophila and mammals. Sequence similarity with a known (NIFL) and a suspected (Aer) flavoprotein suggests that NPH1 LOV1 and LOV2 may be flavin-binding domains that regulate kinase activity in response to blue light–induced redox changes.

Plants irradiated with unilateral blue light grow toward the light source, a phenomenon known as phototropism. We studied a 120-kD plant plasma membrane–associated protein that becomes heavily phosphorylated on irradiation with blue light both in vivo and in vitro. Physiological and genetic evidence implicates this protein in an early step in the signal-transduction pathway for phototropism (1,2). We previously isolated phototropism mutants at four loci in Arabidopsis thaliana (L.) Heynh., designatedNPH1 through NPH4 for nonphototropic hypocotyl (2). Three nph1 alleles generated by fast neutron irradiation (nph1-1, nph1-4, andnph1-5) lack all known phototropic responses, demonstrating that the NPH1 protein is essential for all phototropic signal-transduction pathways in Arabidopsis. In addition, these mutants lack the 120-kD phosphoprotein, suggesting thatNPH1 might encode the phosphoprotein itself (2).

The NPH1 locus is located on chromosome III, within 26 centimorgans (cM) of GL1 (2). Using described methods (3), we employed amplified fragment length polymorphism (AFLP) (4) to identify DNA markers flanking theNPH1 gene at distances of 0.3 and 0.4 cM (Fig.1A). These markers were used to screen anArabidopsis yeast artificial chromosome (YAC) library (5) by polymerase chain reaction (PCR) (6). Of three YACs identified, two (yUP1C7 and yUP21F3) overlap each other and map to chromosome III near the NPH1 gene (2). Using sequences from yUP1C7 as a probe, we detected an altered restriction fragment pattern in nph1-5 and cloned an 18-kb Eco RI fragment that appeared in nph1-5 but not in the wild type. From this fragment, we subcloned a 7-kb Bgl II fragment that contains an end point of the DNA rearrangement in this mutant (Fig.1B). Because we had biochemical evidence that the NPH1 gene product was likely to be a 120-kD protein (2), we used a variety of probes from the region surrounding the breakpoint innph1-5 to look for an mRNA of the appropriate size (about 3.3 kb) on a northern blot of wild-type RNA. We detected a message of 3.35 kb (Fig. 1C) and isolated its wild-type cDNA and genomic clones. Complementation of the nph1-5 mutation with genomic clone p51 restored the mutant's phototropic response (Fig. 1, D to F).

Figure 1

Cloning of NPH1. (A) Autoradiogram of a typical AFLP gel. Lanes 1 and 2, wild-type Columbia (C) and Landsberg erecta (L) parental DNA; lanes 3 to 18, DNA from individual nph1-1 homozygous F2 plants (F2s) derived from a Columbia nph1-1/nph1-1 × Landsberg erecta NPH1/NPH1 cross. The arrow denotes a Landsberg-specific band that shows tight linkage to NPH1 (1 heterozygous recombinant in 258 homozygous nph1-1/nph1-1plants). The bands shown range in size from 120 to 220 base pairs. (B) Autoradiogram of a Southern blot of wild-type (WT) and nph1-5 genomic DNA digested with Eco RI (RI), Bgl II (B), and both enzymes (D) and probed with a 7-kb Bgl II fragment subcloned from an 18-kb Eco RI fragment present innph1-5 but not in the wild type. The 7-kb Bgl II fragment contains an end point of the deletion and rearrangement innph1-5. The positions of molecular size markers are shown in kilobases on the left. (C) Northern blot of wild-type mRNA probed with the 7-kb Bgl II fragment from nph1-5. The probe hybridized to a single band (arrow) corresponding to a message of ∼3350 nucleotides. The positions of molecular size markers are shown in kilobases on the right. (D to F) Wild type (D), nph1-5 (E), and nph1-5 (F) transformed with genomic clone p51. Plants were transformed by vacuum infiltration with p51, a 17-kb genomic clone containing the NPH1 gene. Seeds were vernalized, given 30 min of red light to stimulate germination, grown for 3 days in darkness, and treated with 16 hours of unilateral blue light from the right. The presence of both wild-type and mutant copies of NPH1 in transformed plants was verified by Southern blot analysis (25).

We sequenced the cDNA and genomic clones of NPH1 and found that the gene encodes a protein of 996 amino acids (112 kD) (Fig.2A). The coding region consists of 20 exons extending for 5.4 kb (Fig. 2B). Analysis of the predicted amino acid sequence suggests that the COOH-terminal region of NPH1 encodes a serine-threonine protein kinase (Fig. 2A). The NPH1 kinase domain contains the 11 sequence motifs typical of protein kinases and falls into the PVPK1 family of serine-threonine protein kinases within the protein kinase C group (7). Previous experiments have shown that the 120-kD phosphoprotein from maize and pea is phosphorylated on serine and threonine residues and not on tyrosine residues (8) and is itself a kinase that mediates autophosphorylation (9).

Figure 2

The NPH1 coding region. (A) Predicted amino acid sequence of NPH1 (26). The sequence of amino acids 1 to 7 was obtained from a genomic clone, and amino acids 8 to 996 were obtained from a partial cDNA clone (GenBank accession number AFO30864). The genomic clone was isolated from a library made from partially Hind III–digestedArabidopsis (Columbia) genomic DNA cloned into the cosmid pBIC20 (27), and the cDNA clone was isolated from a cDNA library made from size-selected 3-day-old Arabidopsis(Columbia) seedling hypocotyls (28). Amino acid residues are numbered on the right. The LOV1 and LOV2 domains are doubly underlined, and the kinase domain is singly underlined. The last amino acid residue before the breakpoint in nph1-5 (position 566) is shaded in black. The amino acid residues affected innph1-1 and nph1-2 (loss of valine 774 innph1-1 and substitution of lysine for arginine 936 innph1-2) are also shaded in black. (B) Exon structure of NPH1. The restriction map of the region aroundNPH1 is shown above (E, Eco RI; Bg, Bgl II; S, Sal I; B, Bam HI; and X, Xba I), and the position of the NPH1 gene is shown below. Black boxes represent exons and lines represent introns.

We compared the deduced NPH1 protein sequence with GenBank sequences (10) and identified three sequences that are likely to be from COOH-terminal partial clones of NPH1 homologs from pea (PsPK5) (11), spinach (SK3) (12), and ice plant (MK10) (13). These three sequences are 60 to 76% identical at the protein level withArabidopsis NPH1 over their entire length, including sequences outside the kinase domain. Phy3 from the fern Adiantum capillus-veneris is also similar to NPH1 but, unlike NPH1, has a phytochrome-like domain at its NH2-terminal end (14).

We also sequenced three mutant nph1 alleles (Fig. 2A) (15). The ethylmethane sulfonate–induced nph1-2mutant (16) has a single base substitution near the end of the kinase domain, which results in the substitution of a lysine for an arginine at position 936 (17). This invariant arginine residue maintains the structure of the COOH-terminal lobe of the kinase domain by forming an ion pair with a glutamic acid residue at position 862 (7). Of the two fast neutron mutants sequenced, one (nph1-5) carries a large-scale rearrangement that disrupts the NPH1 gene within an intron following position 565 of the predicted protein sequence. The other, nph1-1(18), carries a three-nucleotide deletion that removes a valine residue from position 774 in the kinase domain, within a highly conserved motif that forms an α helical structure (7). Together with the complementation of the mutant with a genomic clone, the presence of a lesion in each of the three nph1 mutants we sequenced indicates that the gene we cloned is NPH1. The size of the NPH1 gene, the presence of a serine-threonine protein kinase domain in the NPH1-deduced protein sequence, and the absence of the 120-kD protein from the nph1 mutants all suggest that the NPH1 gene encodes the 120-kD phosphoprotein. Although the phosphoprotein is associated with the plasma membrane on isolation (1), a hydrophobicity analysis of NPH1 suggests that it is a soluble protein without membrane-spanning domains. The nature of its association with the plasma membrane remains to be determined.

In the NH2-terminal portion of the predicted NPH1 protein, we found two 107–amino acid sequences that share 43% identity (61% similarity) (Fig. 2A). The motif of these repeats (LOV1 and LOV2) also appears as a single copy in proteins from archaea, eubacteria, and eukaryotes (Fig. 3A) that otherwise share no similarity with NPH1 and have diverse functions (Fig. 3B). However, those for which a function is known are regulated by environmental factors that could change their redox status: light, oxygen, or voltage (hence, LOV). For both NIFL (19) and Aer (20), the mechanism for redox sensing appears to be oxidation or reduction of a flavin prosthetic group. For voltage-gated potassium channels such as ELK, Eag, and ERG, voltage sensing is mediated by charged residues in the fourth transmembrane segment S4 (21). The NH2-terminal part of the channel protein, including the LOV domain, participates in the formation of homotetrameric structures (22). Although no function in voltage sensing is currently attributed to the LOV domain in potassium channels, its participation in regulating channel status has not been ruled out. For the remaining proteins that contain the LOV domain, the redox-sensing mechanism is still unknown.

Figure 3

(A) Similarity between NPH1 LOV1 and LOV2 domains and similar domains in other proteins. Amino acid residues identical in half or more of the sequences are shaded in black, and conservative substitutions are shaded in gray. Amino acid sequences shown are from Arabidopsis thaliana (At) NPH1 LOV1 and LOV2 domains, spinach (So) LOV2 domain, ice plant (Mc) LOV2 domain, Bacillus subtilis (Bs) YtvA (29),Neurospora crassa (Nc) White Collar–1 (WC-1) (30), Halobacterium salinarium (Hs) Bat (31), Drosophila melanogaster (Dm) ELK (32), rat Eag (R-Eag) (33), humanEag-related gene product H-ERG (32),Synechocystis hypothetical proteins slr0359 and slr1305 (34), Klebsiella pneumoniae (Kp) NIFL (35), Azotobacter vinelandii (Av) NIFL (36), Enterobacter agglomerans (Ea) NIFL (37), Synechocystis hypothetical protein slr1759 (34), Escherichia coli (Ec) Aer (20), and transposon Tn1721 open reading frame 1 (ORF1) hypothetical protein (38). The alignment was done with the program PIMA 1.4 (39). The percentage of similarity to NPH1 LOV1 ranges from 56 for YtvA to 34 for Tn1721 ORF1. (B) Structural features of proteins with LOV domains. Shown are Arabidopsis NPH1;D. melanogaster ELK, a voltage-sensitive potassium-channel subunit (32); K. pneumoniae NIFL, a flavoprotein regulator of nitrogenase transcription in response to oxygen (19); H. salinarium Bat, a regulator of bacterio-opsin in response to oxygen or light (or both) (40); N. crassa WC-1, a regulator of blue light responses (30); and E. coli Aer, a putative flavoprotein required for aerotaxis signaling (20). LOV domains are shown as black boxes, and other features are shown as white boxes, with abbreviations as follows: kinase, serine-threonine protein kinase domain; S1-S6, six membrane-spanning regions (32); cNBD, cyclic nucleotide-binding domain (32); NIFA-binding, region that interacts with NIFA (41); Q-rich, glutamine-rich region (30); PAS, PAS domain (42); Zn finger, zinc finger (30); H, hydrophobic segment (20); MCP-like, resembling the signaling domain of methyl-accepting chemotaxis proteins (20).

One possible function for the LOV domain in this diverse group of proteins is as a flavin-binding site with the bound flavin acting as a redox sensor. Although NIFL is a flavoprotein (19) and Aer is a putative flavoprotein (20), we do not yet know whether the FAD attachment site is located within the LOV domain. If the LOV domain is a flavin attachment site, it shares little or no sequence similarity with known flavin-binding sites, such as those of the CRY1/CRY2 blue light photoreceptors (23). The LOV domain may also mediate protein-protein interactions, because it shares similarity with PAS domains, which participate in protein-protein interactions (24). Regardless of its function, the LOV domain has been conserved during evolution over a wide range of taxa from archaea through higher plants and animals.

Elucidation of the properties and functions of the LOV domain, including the role of two such domains in NPH1, will help unravel the details of the signal-transduction pathway for phototropism.

  • * Present address: DNA Plant Technology, 6701 San Pablo Avenue, Oakland, CA 94508, USA.

  • Present address: Biology Department, University of Missouri, Columbia, MO 65211, USA.

  • Present address: Department of Biology, University of Ulsan, Kyungnam, Ulsan 680-749, Korea.

  • § Present address: Department of Botany and Plant Science, University of California, Riverside, CA 92521, USA.

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


View Abstract

Navigate This Article