LKP1 (LOV kelch protein 1): a factor involved in the regulation of flowering time in Arabidopsis

Authors

  • Tomohiro Kiyosue,

    Corresponding author
    1. 1 Gene Research Center, Kagawa University, Miki-cho, Kita-gun, Kagawa 761–0795, Japan, 2 Biological Regulation Division, Department of Regulation Biology, National Institute for Basic Biology (NIBB), Okazaki 444–8585, Japan, and 3 Department of Biology, Faculty of Science, Tokyo Metropolitan University, Minami-ôsawa 1–1, Hachioji, Tokyo 192–0397, Japan
    Search for more papers by this author
  • and 1,2, Masamitsu Wada 2,3

    1. 1 Gene Research Center, Kagawa University, Miki-cho, Kita-gun, Kagawa 761–0795, Japan, 2 Biological Regulation Division, Department of Regulation Biology, National Institute for Basic Biology (NIBB), Okazaki 444–8585, Japan, and 3 Department of Biology, Faculty of Science, Tokyo Metropolitan University, Minami-ôsawa 1–1, Hachioji, Tokyo 192–0397, Japan
    Search for more papers by this author

*For correspondence (fax +81 878 91 3021; e-mail tkiyosue@ag.kagawa-u.ac.jp).

Summary

In plants, light is not only an energy source but also a very important signal that modulates development and differentiation. Here, we report a putative photo-regulatory factor sequence in LKP1 (LOV kelch protein 1). LKP1 cDNA encodes a protein of 610 amino acids and with a molecular weight of 65 905 with an LOV domain and kelch repeats. LOV domains are present in a number of sensor proteins involved in the detection of light, oxygen or voltage. The LKP1 LOV is very similar to the LOV domains in NPH1, a plasma membrane-associated blue light receptor kinase that regulates phototropism (Huala, E., Oeller, P.W., Liscum, E., Han, I-S., Larsen, E. & Briggs, W.R. (1997) Science, 278, 2120–2123). LKP1 mRNA accumulates in roots, stems, flowers and siliques. It is most abundant in leaves, and least abundant in seeds. Transgenic plants with a β-glucuronidase (GUS) reporter gene driven by a 1.5 kb LKP1 promoter display strong GUS activity in leaves. Transgenic plants with a 35S::LKP1 cDNA gene overexpress LKP1 mRNA. These plants have elongated hypocotyls and petioles with elongated cells, and exhibit distinct cotyledon movement during the day. Expression of 35S::LKP1 in transgenic Arabidopsis promotes late flowering in plants grown under long-day, but not under short-day conditions. Vernalization does not affect the late flowering phenotype of the 35S::LKP1 plants. Transgenic plants possessing the 35S::GFP-LKP1 construct also have long hypocotyles and petioles, and a late flowering phenotype, suggesting that the GFP-LKP1 fusion protein is active. The GFP-associated fluorescence in 35S::GFP-LKP1 plants is observed in nuclei and cytosol, indicating that LKP1 is a new nucleo-cytoplasmic factor that influences flowering time in the long day pathway of Arabidopsis.

Introduction

Due to the fact that plants are immobile, they adapt with environmental stimuli and use them as signals for development and differentiation through physiological and biochemical changes and gene expression at both plant and cellular level. For plants, light is very important as a signal in addition to being an energy source. Unlike animals, plants mainly perceive red and blue light. These wavelengths regulate many plant responses, including the breaking of dormancy, germination, cell elongation, tropic response, chloroplast movement, stomata opening, flavonoid biosynthesis, and flowering.

Cryptochromes are the first sequence-identified blue light receptors ( Ahmad & Cashmore, 1993). Cryptochrome 1 (CRY1) regulates blue-light-dependent phenomena, such as hypocotyl, petiole and stem elongation, cotyledon and leaf expansion, anthocyanin accumulation, and flowering ( Ahmad et al. 1995 ; Jackson & Jenkins, 1995). Arabidopsis possesses another cryptochrome, CRY2, which has functions that overlap with those of CRY1 to a certain extent ( Ahmad et al. 1998 ). Both CRY1 and CRY2 play roles in hypocotyl elongation inhibition, anthocyanin formation, flowering and blue-light-induced suppression of stem growth ( Lin et al. 1996 ; Lin et al. 1998 ). CRY1 has been shown to function in perceiving light signals which effect the oscillator that drives circadian rhythms ( Somers et al. 1998 ; Zhong et al. 1997 ). Overexpression of CRY1 significantly shortened the free-running period at high fluences of blue or white light, whereas the loss of Cry1 resulted in period lengthening over both high and low fluence-rate ranges ( Somers et al. 1998 ).

NPH1 or phototropin, which is involved in phototropism, is the most recently sequenced blue light receptor ( Huala et al. 1997 ). NPH1 is a plasma membrane-associated serine/threonine protein kinase, the activity of which is induced by light ( Christie et al. 1998 ). NPH1 possesses two domains that have been found in light sensors, oxygen sensors, and the eag family of voltage-gated potassium channel proteins, designated as LOV ( Huala et al. 1997 ). Both the LOV domains in NPH1 bind the flavin mononucleotide (FMN) with a flavin/protein ratio of 1 which acts as a UV-A/blue light absorbing chromophore ( Christie et al. 1998 ). The crystal structure of the human HERG LOV domain and the Bradyrhizobium FixL LOV domain have been determined, where the LOV domain forms a β-sheet core flanked by α-helices making a hydrophobic pocket that can bind a ligand; the hydrophobic pocket is accessible for ligand binding via a hydrophilic entryway (Cabral et al. 1998; Gong et al. 1998 ). There is a reported NPH1-related sequence (NPL1: non-phototropic hypocotyl-like 1) that encodes a protein with two LOV domains in Arabidopsis, the function of which is unknown ( Jarillo et al. 1998 ).

Zeaxanthin is postulated as the photoreceptor chromophore for stomatal opening ( Niyogi et al. 1998 ; Zeiger & Zhu, 1998), but the apoprotein for the receptor remains as yet unknown.

Since there are numerous phenomena regulated by blue light, these four receptors (CRY1, CRY2, NPH1 and the blue-light photoreceptor for the stomatal response) may not be responsible for all of them. Indeed, the blue light receptor involved in chloroplast movement is speculated to be different from these four receptors ( Kagawa & Wada, 2000). We tried to identify new factors involved in blue light reception in Arabidopsis and isolated several cDNAs that encode LOV domain proteins. Here, we report on one of these clones, LKP1.

Results and Discussion

Isolation and characterization of LKP1 sequence

An Arabidopsis database was searched with sequences for NPH1 LOV domains in order to identify putative additional blue light photoreceptors. Several sequences were found. One of the sequences was located in a PI clone of MSF19, which has been entirely sequenced, in chromosome 5. The DNA sequence was analyzed by the GENSCAN program to identify the putative coding sequence that encodes the LOV domain containing protein. The cDNA for this predicted protein was screened from the cDNA library from 3-day-old hypocotyls ( Kieber et al. 1993 ). Among 4 × 300 000 phages, 18 clones were isolated and analyzed. End-sequencing of these clones revealed that nine cDNAs corresponded to the putative open reading frame (ORF). The entire nucleotide sequences of these nine clones were determined. The longest clone had an insert of 2297 bp with an ORF of 1830 bp encoding a protein of 610 amino acids with a molecular mass of 65905.17 Da and a putative nulear localization signal ( Fig. 1a). BLAST search analysis revealed that the protein had a LOV domain at the N-terminal part, and six internal repeats belonging to the kelch family ( Bork & Doolittle, 1994; Xue & Cooley, 1993) at the C-terminal half ( Fig. 1b,c). Therefore, we named this protein LKP1 (LOV kelch protein 1). The LOV domain of LKP1 shows striking sequence homology (48.8–58.1% identity, 74.4–86.0% similarity) with those of blue light signaling molecules, Arabidopsis NPH1 and Neurospora White Collar-1 ( Ballario et al. 1996 ), and of putative photoreceptors, Arabidopsis NPL1 ( Jarillo et al. 1998 ), Adiantum PHY3 ( Nozue et al. 1998 ), and Halobacterium bacterio-opsin acitivator ( Leong et al. 1988 ). The kelch repeats of LKP1 show sequence homology with those of yeast tip elongation aberrant protein 1 (TEA1) ( Mata & Nurse, 1997) (25–29% identities and 37–45% similarities), human and mouse host-cell factors C1 ( Wilson et al. 1993 ) (25–31% identities and 39–46% similarities), and Physarum actin-fragmin kinase ( Eichinger et al. 1996 ) (26–27% identities and 39–44% similarities). Kelch repeats were originally identified in the kelch gene product, a component of ring canals, which regulates the cytoplasmic flow between nurse cells and the oocyte during the maturation of the Drosophila oocyte ( Xue & Cooley, 1993), and are present in several proteins that are known, or are suspected, to bind to F-actin ( Eichinger et al. 1996 ; Varkey et al. 1995 ; Way et al. 1995 ; Xue & Cooley, 1993). From the analogy to the solved three-dimensional structure of galactose oxidase from the fungus Dactylium dendroides that contains seven kelch repeats, six kelch repeats are speculated to be organized as a one domain superbarrel structure consisting of six times four β-strands ( Bork & Doolittle, 1994; Eichinger et al. 1996 ).

Figure 1.

Structure of the LKP1 cDNA.

(a) Nucleotide sequence of LKP1 cDNA and its deduced amino acid sequence. A putative nuclear localization signal is shaded.

(b) Similarity between the LKP1 LOV domain and similar domains in other proteins. Amino acid residues identical in half or more of the sequences are coloured in green, and conserved residues are in red. Amino acid sequences shown are from the LOV1 and LOV2 domains of Arabidopsis NPH1 and NPL1, those of Adiantum PHY3, and the LOV domain of Neurospora White Collar 1 (WC1) plus that of Halobacterium probable bacterio-opsin activator (PBOA).

(c) Alignment of the kelch repeats. Identical residues on at least three repeats are shown in green, and conserved residues are shown in red.

Organ-specific expression and promoter activity of the LKP1 gene

The organ-specific LKP1 mRNA level was examined by means of RT–PCR. LKP1 mRNA was detected in every organ at a different level: dry seeds < stems < siliques < roots < flowers and leaves ( Fig. 2). Since the mRNA level of some blue light receptors is reported to change during the day, the LKP1 mRNA level of rosette leaves was examined every 4 h, but no clear rhythmic RNA expression was detected (data not shown). The promoter activity of the LKP1 gene was histochemically monitored by the GUS activity of transgenic Arabidopsis containing the LKP1P::GUS construct. Strong signals were detected in cotyledons, rosette leaves and cauline leaves ( Fig. 3a,b,d). The GUS activity was also detected in roots, sepals, siliques and siliques possessing flower axes, whereas little activity was detected in hypocotyls, basal parts of petioles, petals, young flower axes and main axes ( Fig. 3a–e). This promoter activity agrees with the LKP1 mRNA level shown in Fig. 2, suggesting that the 1.5 kb promoter region has an important role in regulating the LKP1 gene expression.

Figure 2.

Analysis of LKP1 mRNA accumulation by RT–PCR.

R, roots; St, stems; RL, rosette leaves; CL, cauline leaves; F, flowers; Si, siliques; DS, dry seeds. As a control, 18S rRNA was amplified. Two μg total RNA was used in each reaction.

Figure 3.

Histochemical localization of GUS activity in transgenic Arabidopsis.

Transgenic Arabidopsis (T2) plants containing LKP1 promoter/GUS fusion gene were grown on GM agar that contained 50 μg ml−1 kanamycin. GUS activity is shown in seedlings (a), rosette plants (b), flowers (c), cauline leaves (d) and siliques (e).

Gain of function of 35S::LKP1 transgenic Arabidopsis

1. Elongated hypocotyl and petiole. To elucidate the function of LKP1, knock-out lines for the LKP1 gene were PCR screened from a T-DNA pool of 50 000 lines organized by the Kazusa DNA Research Center. Unfortunately, no knock-out lines were found. As a next approach, transgenic Arabidopsis plants were generated in which the LKP1 cDNA was expressed under the control of the constitutive cauliflower mosaic virus 35S promoter ( Mitsuhara et al. 1996; Sanders et al. 1987 ) either in the sense or antisense orientation. Among 62 kanamycin-resistant T1 plants transformed with the sense LKP1 construct, a significant number (11) showed a distinct set of phenotypes. In contrast, 67 kanamycin-resistant putative antisense T1 plants did not show any phenotypes described below (data not shown), which suggests that a functionally redundant gene for LKP1 might exist in the Arabidopsis genome. The sense plants, the transgenic plants overexpressing the LKP1 transgene, displayed phenotypes with elongated hypocotyls and petioles ( Fig. 4a,c) under LD or SD. No ‘long hypocotyls and petioles’ phenotypes of 35S::LKP1 seedlings were observed when plants were grown under dark conditions (data not shown). The epidermal cell lengths of hypocotyls of 35S::LKP1 plants were three- to fivefold longer than those of the control or wild type ( Fig. 4d,e), while the total epidermal cell numbers of hypocotyls along with the axes were the same (data not shown). The petioles of 35S::LKP1 seedlings also possessed elongated epidermal cells (data not shown).

Figure 4.

Long hypocotyl and petiole phenotypes of 35S::LKP1 transgenic Arabidopsis.

(a–d) Six-day-old T2 seedlings and 2-week-old rosette plants of 35S::LKP1 (a,c) and 35S-vector-only control (b,d) transformants grown under long-day conditions (∼90 μmol m−2 sec−1) on GM agar containing 50 μg ml−1 kanamycin. (e,f) Epidermal cells of the seedlings of 35S::LKP1 (e) and controls (f).

  • 2. Cotyledon movement. The other characteristic of 35S::LKP1 seedlings was distinct cotyledon movement. Cotyledons of 35S::LKP1 seedlings located vertically at midnight (ZT20–0) and horizontally in the afternoon (ZT8–16), while control or wild-type seedlings did not show such clear cotyledon movement under long-day conditions ( Fig. 5). This may suggest that cotyledon petioles grow with different cell expansion rates between inner epidermal cells and outer cells in the 35S::LKP1 seedlings. In Arabidopsis grown under relatively high intensity light (approximately 90 μmol m−2 sec −1), the promoter activity of the LKP1 gene was not detected in the petioles of cotyledons ( Fig. 3a). Thus, the distinct cotyledon movement might be due to the ectopically expressed LKP1 in petioles. Cotyledon and leaf movement of Arabidopsis is known to show circadian rhythm ( Hicks et al. 1996 ; Schaffer et al. 1998 ). Genes for GIGANTEA (GI) and late elongated hypocotyl (LHY) are known to be involved in that process ( Park et al. 1999 ). Therefore, LKP1 might undergo some interaction with these regulating molecules.
  • 3. Late flowering. Another striking finding is the late flowering phenotype of the 35S::LKP1 Arabidopsis. Under LD conditions, Columbia wild-type plants flowered around 4 weeks after sowing, whereas 35S::LKP1 plants did not ( Fig. 6a,b). The transgenic plants of 35S::LKP1 showed two types of late flowering phenotypes, Type 1 and Type 2. The Type 1 plants formed many rosette leaves (57.1 ± 6.6) before bolting. Bolting began 11 weeks after sowing ( Fig. 6c,d). The Type 2 plants started bolting 6 weeks after sowing when they possessed 17.4 ± 2.5 rosette leaves. The Type 2 plants possessed rosette-shaped cauline leaves, and resultant plants formed rosette leaf masses with an elongated primary inflorescence stem ( Fig. 6e–g). Type 1 and Type 2-resembling phenotypes were seen in wild plants grown under SD conditions with high (∼120 μmol m−2 sec−1) or low intensity (∼50 μmol m−2 sec−1 ) of light, respectively (data not shown). The expression level for the LKP1 gene was estimated by means of RT–PCR ( Fig. 7). Type 2 plants accumulate much higher levels of LKP1 mRNA than the wild type. Type 1 plants accumulate somewhat higher levels of LKP1 mRNA than the Type 2 plants. The phenotype of T3 plants of a Type 2 line were analyzed. The kanamycin-resistant T3 plants show both Type 1 and Type 2 phenotypes with a ratio of 1 : 2 (18 : 41, X2 = 0.22, P > 0.95). Thus, the different LKP-OX phenotypes are probably due to the difference in LKP1 expression level and/or copy number of the transgene.
Figure 5.

Cotyledon movement.

Five-day-old seedlings were grown under long-day (∼90 μmol m−2 sec−1) conditions on GM agar and photographed every 4 h. ZT (zeitgeber) is the number of hours after dawn (onset of illumination) ( Fowler et al. 1999 ; Zerr et al. 1990 ).

Figure 6.

Late flowering phenotypes of 35S::LKP1 transgenic Arabidopsis.

Wild type (Columbia) (a: 4.5-week-old) and 35S::LKP1 Arabidopsis (b: 4.5-week-old, c: 11-week-old, d: 12.5-week-old, e: 8.5-week-old, f: 11-week-old, g: 12-week-old) were grown under long-day conditions (∼60 μmol m−2 sec−1).

Figure 7.

LKP1 expression level in 35S::LKP1 transgenic and wild type Arabidopsis analyzed by RT–PCR.

Total RNA of 0.02–2 μg was used in each reaction.

In Arabidopsis, over 50 flowering time-controlling genes have been identified and characterized by genetic means using mutants, and several genetic pathways have been postulated to form part of the regulatory process ( Levy & Dean, 1998; Pineiro & Coupland, 1998). The autonomous pathway regulates flowering independently of photoperiod. Since mutants such as fca and ld flower later than the wild type in both LD and SD and show a decreased flowering time in response to vernalization treatments, the corresponding genes are thought to function in the autonomous pathway. The FCA gene encodes an RNA binding protein with a protein–protein interaction domain ( Macknight et al. 1997 ), and the LUMINIDEPENDENS gene encodes a glutamine-rich homeobox transcription factor ( Lee et al. 1994 ). The LD pathway regulates flowering specifically in response to LD conditions. The mutation or deletion of genes that act in the LD pathway, such as co, fha and gi, causes late flowering under LD conditions but not under SD conditions, and the mutants show little or no response to vernalization. CO encodes a putative transcription factor with two zinc fingers ( Putterill et al. 1995 ), FHA encodes cryptochrome 2 (CRY2) ( Guo et al. 1998 ), and GI encodes a novel membrane protein ( Fowler et al. 1999 ; Park et al. 1999 ). The SD pathway or GA pathway regulates flowering under SD conditions via gibberellic acid. Vernalization promotes flowering, possibly via reduction in DNA methylation.

In order to classify the LKP1 functioning pathway in 35S::LKP1 plants, the effects of light conditions and vernalization on the Type 1 late flowering phenotype of 35S::LKP1 plants were analyzed by rosette leaf number before bolting ( Fig. 8). Under LD conditions, late flowering of 35S::LKP1 plants was clearly observed, whereas late flowering did not take place under SD conditions. Since the fve mutant flowers later than wild-type plants under both LD and SD conditions and shows a decreased flowering time in response to vernalization, the fve gene is proposed to act within an autonomous pathway ( Martinez-Zapater et al. 1994 ; Pineiro & Coupland, 1998). Indeed, vernalization did affect late flowering phenotypes of the vernalization sensitive mutant of fve-2 ( Martinez-Zapater et al. 1994 ) under both LD and SD conditions. On the other hand, vernalization did not affect the late flowering phenotype of 35S::LKP1 under either LD or SD conditions. We conclude that the LKP1 functions in the long-day pathway of flowering in Arabidopsis ( Levy & Dean, 1998; Pineiro & Coupland, 1998).

Figure 8.

Flowering times of 35S::LKP1 transgenic Arabidopsis.

Flowering time was measured by the total leaf number of plants before bolting. Columbia wild type, 35S::LKP1 transgenic and fve-2 mutant (Landsberg background) plants were grown under LD conditions (18 h of light, 6 h of darkness) (∼60 μmol m−2 sec−1) or under SD conditions (9 h of light, 15 h of darkness) (∼120 μmol m−2 sec−1). To vernalize, imbibed seeds were incubated at 4°C for 30 days under SD conditions, and then incubated either under LD (VN LD) or SD (VN SD) conditions. Since the fve-2 plant did not bolt and possessed more than 60 leaves after it was incubated for 80 days under LD conditions, this datum is not plotted in the figure.

Intracellular localization of GFP-LKP1

To visualize LKP1 intracellular localization, the 35S promoter derived GFP-LKP1 fusion construct was introduced into Arabidopsis. Among 50 kanamycin-resistant T1 plants overexpressing a GFP-LKP1 fusion protein, seven plants showed a similar set of phenotypes under LD conditions ( Fig. 9b), suggesting that the GFP-LKP1 fusion protein is functionally active. GFP-associated fluorescence was found to be distributed in nuclei but not in nucleoli when the root cells of the 35S:GFP-LKP1 plants were visualized by fluorescence microscopy in the afternoon ( Fig. 9c,d). This nuclei localization of GFP-LKP1 was also detected in the trichome and epidermal cells of leaves (data not shown). In contrast, the signal was detected in the cytosol in the early morning under LD conditions ( Fig. 9f). This cytosolic localization of GFP-LKP1 did not alter during 1 h incubation under white light or darkness (data not shown).

Figure 9.

Late flowering phenotype of 35S::GFP-LKP1 transgenic Arabidopsis and subcellular localization of GFP-LKP1 fusion proteins.

T2 plant of 35S-vector-only control (a) and T1 plant of 35S::GFP-LKP1 (b) transformant were grown under long-day conditions (∼90 μmol m−2 sec−1) on GM agar containing 50 μg ml−1 kanamycin. Roots of the 35S::GFP-LKP1 plants were observed by optical microscopy (c) or fluorescent microscopy (d–f) in zeber time 9 (ZT9) (c–e) or zeber time 21 (ZT21) (f).

Possible LKP1 function

Our data taken together with a number of facts suggest two possible functions of LKP1. One is the blue light signaling function. LKP1 possesses a LOV domain that is very highly homologous (84% similarity) to those present in NPH1, which is a blue light receptor involved in phototropic responses of etiolated seedlings to blue light at a low fluence rate in Arabidopsis ( Huala et al. 1997 ; Liscum & Briggs, 1995; Liscum & Briggs, 1996). In NPH1, the LOV domain non-covalently binds FMN with a flavin/protein ratio of 1 ( Christie et al. 1998 ). The LOV domains are postulated to be light-sensing chromophores for this blue light receptor kinase ( Christie et al. 1998 ). It is interesting to note that photo-signaling molecules such as phytochromes A and B and cryptochromes CRY1 and CRY2 have been shown to contribute to the inhibition of hypocotyl cell elongation ( Lin et al. 1996 ; Lin et al. 1998 ; Whitelam & Devlin, 1998), and that mutations in the PHYA, CRY1 and CRY2 affect the long-day pathway of flowering time in Arabidopsis ( Levy & Dean, 1998; Pineiro & Coupland, 1998); and also that nucleo-cytoplasmic localization is seen in some phytochromes and cryptochromes ( Inaizumi et al. 2000 ; Kircher et al. 1999 ; Yamaguchi et al. 1999 ).

The other possible function is the circadian clock-related function. The LOV domain also functions in protein–protein interactions. The NPH1 LOV1 domain was shown to interact with NPH3, a signaling component working downstream of NPH1 ( Motchoulski & Liscum, 1999). LOV domains are related to PAS domains ( Zhulin et al. 1997 ), which are found in many proteins that are related not only to light perception and signaling, such as phytochromes, but also to clock-related proteins, such as WC-2, PER and CLOCLK ( Ballario & Macino, 1997; Huang et al. 1993 ; King et al. 1997 ; Linden & Macino, 1997). The PAS domains of these proteins also function in protein–protein interactions. A motif search of LKP1 picked up a PAS domain motif in the LKP1 LOV domain with an E-value score of 0.0045.

The long-hypocotyl phenotype is also seen in LHY and CCA1 mutants, where myb-related genes are activated by 35S promoters ( Schaffer et al. 1998 ; Wang & Tobin, 1998). The LHY and CCA1 are postulated to be closely associated with the central oscillator of the circadian clock in Arabidopsis.

The 35S::LKP1 transgenic seedlings showed distinct cotyledon movement, while wild-type seedlings did not. Leaf movement is known to show a circadian rhythm ( Park et al. 1999 ; Schaffer et al. 1998 ).

The late flowering phenotype in the LD pathway was also seen in both the LHY and CCA1 dominant mutants, where circadian clock regulation of gene expression and leaf movements were disrupted ( Schaffer et al. 1998 ; Wang & Tobin, 1998).

Finally, the nucleocytoplasmic localization of GFP-LKP1 is not simply regulated by light because the GFP signals did not transfer to the nuclei when GFP-LKP1 expressing plant tissues incubating in the dark were moved to under white light.

Kay's group has recently characterized a long period mutant, Zeitlupe 1. This mutant has a dramatically long period in dim blue light, implying that blue light signaling to the clock is altered. After chromosome walking, they found that Zeitlupe 1 mutants have mutations in the kelch repeat coding region of the LKP1 gene (S.A. Kay, personal communication). Their data support our views on LKP1 function. LKP1/ZTL1 appears to be a new blue light signal transduction molecule involved in controls of flowering time and circadian rhythm. Interestingly, Bartel's team also recently walked to a gene defective in a late Arabidopsis flowering mutant (fkf1) that they had isolated as part of a deletion on the bottom of chromosome 1 (B. Bartel, personal communication). This mutant also has hypocotyl elongation defects and the FKF1 transcript is regulated in circadian fashion. FKF1 is homologous to LKP1/ZTL1 (65.9% in terms of amino acids), and LKP1/ZTL1 and FKF1 belong to the same small gene family. However, in our RT–PCR experiment, the LKP1 transcript was not clearly regulated in circadian fashion (data not shown). It is surprising that there is no perfect complementary function between these genes on flowering since each mutant shows the late flowering phenotype. The targets for these proteins may be diverse. Since mutation and overexpression of LKP1/ZTL1 affects flowering time, LKP1/ZTL1 has potential as a useful tool applicable to the control of flowering time in higher plants.

Experimental procedures

Plant materials

Seeds of Arabidopsis thaliana (Columbia ecotype) were sown axenically on GM (germination medium containing 0.09 m sucrose) ( Nakashima et al. 1997 ) containing 0.8% agar and incubated at 4°C in the dark for 3 days to break dormancy, and then grown under LD conditions of 16 h light (∼90 μmol sec−1 m−2) and 8 h of dark at 22°C. Rosette seedlings were transferred to and grown on soil under the same conditions. For flowering tests, seeds were sown on soil and incubated at 4°C in the dark for 3 days, and then grown under either LD (16 h dark/8 h light, ∼60 μmol sec−1 m−2) or SD (8 h light/16 h dark, light ∼120 μmol sec−1 m−2) at 22°C. To vernalize, seeds were sown on soil and incubated for 30 days at 4°C under the SD regime. A vernalization sensitive late flowering mutant, fve-2 (Landsberg background) ( Koornneef et al. 1991 ; Martinez-Zapater et al. 1994 ), was used as the control.

CDNA library screening

Screening of cDNA libraries [CD4–13 (0.5–1 kb), CD4–14 (1–2 kb), CD4–15 (2–3 kb), and CD4–16 (3–6 kb)] ( Kieber et al. 1993 ) was performed by plaque hybridization as described by Maniatis et al. (1982) using a 4.15 kb PCR fragment generated from Arabidopsis genomic DNA, which corresponds to the 6771–10920 region of the MSF19 P1 clone, as the probe. The primers used were GAGTTTTATGGTTTTATCTACTTGACCCGA and ACTAGCGATTA-CGCTCGGCCGGTAGAGGAC. Each cDNA library was plated as 6 × 5000 plaques and screened.

DNA sequencing and analysis

Plasmid DNA templates for sequencing were prepared by using an automatic plasmid isolation system (model PI-100, Kurabo, Osaka). DNA sequences were determined by the BigDye-terminator sequencing method using a DNA sequencer (model 317; Applied Biosystems, Foster City, CA, USA). The GENETYX (Software Development, Tokyo, Japan) and DNASIS (HITACHI Software Engineering, Tokyo, Japan) software systems were used for the analysis of DNA and amino acid sequences. A BLAST search was performed via the NCBI web site ( http://www.ncbi.nlm.nih.gov/). GENSCAN was performed via the web site at Stanford ( http://genomic.stanford.edu/GENSCANW.html). Motif search was performed via the http://www.motif.genome.ad.jp/web site.

RNA methods

Total RNA was isolated by means of the phenol/SDS method as described previously ( Kiyosue et al. 1992 ). One step RT–PCR was performed with 0.02–2 μg of RNA as starting material by using Ready-To-Go RT–PCR beads (Amersham Pharmacia Biotech). Primers used to amplify LKP1 RNA were TCATGGAGTGGGACA-GTGGT corresponding to the first exon and CAACAACTCGCCTT-GAAATT corresponding to the second exon. The 18S ribosomal RNA was amplified by using Quantum RNA 18S Internal Standards primers for plants (Ambion, Austin, TX, USA). The reaction volume of each reaction was 50 μl, and 10 μl of each reaction was loaded per lane.

Transgenic plants

A 1.5 kb 5′ non-coding region with 21-b overlap of LKP1 coding sequence from the initiation codon of ATG was PCR amplified by Pfu Turbo DNA polymerase (Stratagene) from genomic DNA of Columbia ecotype Arabidopsis. The primers used were GAGATCTAGATAGTTGCCCCATATCGGTAA and TCTCCCCGGG-GAACCACTGTCCCACTCCAT, which introduced an XbaI site and a SmaI site at each end, respectively. The PCR fragment was subcloned into pCR-Blunt II-TOPO (Invitrogen), sequenced to verify the sequence, and then ligated into a promoterless pC-glucronidase (GUS) expression vector pBI101 (Clontech, Palo Alto, CA, USA). The resultant construct was sequenced to verify in-frame fusion of the seven amino acid regions of the LKP1 to GUS protein region.

The expression vector of pBE2113 that contained cauliflower mosaic virus 35S promoter ( Mitsuhara et al. 1996) was used for overexpression of LKP1 cDNA either in the sense or antisense orientation. The LKP1 coding sequence was PCR amplified by Pfu Turbo or LA-Taq DNA polymerase (Takara, Kyoto, Japan) from the cDNA with the primers of GAGAGGATCCTCATGGAGTGGGACA-GTGGT and TCTCGGATCCGGTTACGTGAGATAGCTCGC, which introduced a BamHI site at each end. The PCR fragment was subcloned into pCR-BluntII-TOPO or pCR2.1-TOPO (Invitrogen), sequenced to verify the sequence, and then ligated into pBE2113.

The S65T modification version of GFP (smRS-GFP) ( Davis & Vierstra, 1998) was used to make GFP-LKP1 construct. The GFP coding region was PCR amplified from psmRS-GFP from the Arabidopsis Biological Resource Center, with the primers of GAGATCTAGACAATGAGTAAAGGAGAAGAA and TCTCAGGCC-TTTGTATAGTTCATCCATGCC, which introduced a XbaI site and a StuI site at each end, respectively. The LKP1 coding region was also PCR amplified from the cDNA with the primers of GAGAAGGCCTCATGGAGTGGGACAGTGGT and TCTCGGATCC-GGTTACGTGAGATAGCTCGC, which introduced a StuI site and a BamHI site at each end, respectively. Both PCR fragments were subcloned, sequenced to verify the sequence, ligated at the StuI site, and then the GFP-LKP1 fragment was introduced into pBE2113.

Agrobacterium-mediated transformation of Arabidopsis plants was performed by a simplified in planta infiltration method ( Clough & Bent, 1998). Transgenic lines were selected on GM agar that contained 50 μg ml−1 kanamycin.

GUS staining and GFP observation

Histochemical localization of GUS activities in the transgenic plants was performed as described previously ( Nakashima et al. 1997 ). Roots or leaves were detached from the 35S::GFP-LKP1 plants, incubated for 30 min in the dark and submerged in water, after which the cells were observed using a Zeiss Axioplan microscope (Carl Zeiss, Oberkochen, Germany).

Acknowledgements

We thank Drs B. Bartel (Rice University, Houston, TX, USA), W.R. Briggs and J.M. Christie (Carnegie Institution of Washington, Stanford, CA, USA), S. Kay (The Scripps Research Institute, La Jolla, CA) and Y. Komeda (Hokkaido University, Hokkaido, Japan) for valuable discussions; Drs Y. Ohashi (National Institute of Agrobiological Research, Tsukuba, Japan) and K. Shinozaki (RIKEN, Tsukuba, Japan) for providing the vector pBE2113; Drs H. Abe (JIRCAS, Tsukuba, Japan), T. Nanjo (RIKEN, Tsukuba, Japan), Y. Oono (Atomic Power Research, Tsukuba, Japan) and N. Ueno (NIBB, Okazaki, Japan) for technical advice and assistance; Drs S. Tabata and T. Kato (KAZUSA DNA Research Institute, Kisarazu, Japan) for providing T-DNA pools for PCR screening, and the Arabidopsis Biological Resource Center (Columbus, OH, USA) for providing psmRS-GFP clone, seed stocks and cDNA libraries. We also thank Dr R.L. Fischer (UC-Berkeley, Berkeley, CA, USA) for his linguistic suggestions. This work was supported in part by the Grant-in-Aid for Encouragement of Young Scientists (grant number 11740450) to T.K. from the Ministry of Education, Science, Sports, and Culture of Japan, and by the NISSAN Science Foundation to T.K., and was also supported in part by PROBRAIN (Program for Promotion of Basic Research Activities for Innovative Biosciences) and the Grant-in-Aids for Scientific Research (B) (grant number 10044214) to M.W.

Footnotes

  1. GenBank accession number AB038796.

Ancillary