Large-scale screening of Arabidopsis circadian clock mutants by a high-throughput real-time bioluminescence monitoring system

Authors

  • Kiyoshi Onai,

    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    2. Bio-oriented Technology Research Advancement Institution (BRAIN), Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    Search for more papers by this author
  • Kazuhisa Okamoto,

    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    2. Aichi Science and Technology Foundation, 2-4-7 Marunouchi, Naka, Nagoya 460-0002, Japan
    Search for more papers by this author
  • Harumi Nishimoto,

    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    Search for more papers by this author
  • Chisato Morioka,

    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    Search for more papers by this author
  • Minako Hirano,

    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    Search for more papers by this author
  • Nobunori Kami-ike,

    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    2. Bio-oriented Technology Research Advancement Institution (BRAIN), Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    Search for more papers by this author
  • Masahiro Ishiura

    Corresponding author
    1. Center for Gene Research, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    2. Bio-oriented Technology Research Advancement Institution (BRAIN), Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
    3. Aichi Science and Technology Foundation, 2-4-7 Marunouchi, Naka, Nagoya 460-0002, Japan
    4. Division of Biological Science, Graduate School of Science, Nagoya University, Furo, Chikusa, Nagoya 464-8602, Japan
      (fax +81 52 789 4526; e-mail ishiura@gene.nagoya-u.ac.jp).
    Search for more papers by this author

(fax +81 52 789 4526; e-mail ishiura@gene.nagoya-u.ac.jp).

Summary

Using a high-throughput real-time bioluminescence monitoring system, we screened large numbers of Arabidopsis thaliana mutants for extensively altered circadian rhythms. We constructed reporter genes by fusing a promoter of an Arabidopsis flowering-time gene – either GIGANTEA (GI) or FLOWERING LOCUS T (FT) – to a modified firefly luciferase gene (LUC+), and we transferred the fusion gene (PGI::LUC+ or PFT::LUC+) into the Arabidopsis genome. After mutagenesis with ethyl methanesulfonate, 50 000 M2 seedlings carrying the PGI::LUC+ and 50 000 carrying PFT::LUC+ were screened their bioluminescence rhythms. We isolated six arrhythmic (AR) mutants and 29 other mutants that showed more than 3 h difference in the period length or phase of rhythms compared with the wild-type strains. The shortest period length was 16 h, the longest 27 h. Five of the six AR mutants carrying PGI::LUC+ showed arrhythmia in bioluminescence rhythms in both constant light and constant dark. These five AR mutants also showed arrhythmia in leaf movement rhythms in constant light. Genetic analysis revealed that each of the five AR mutants carried a recessive mutation in a nuclear gene and the mutations belonged to three complementation groups, and at least one of which was mapped on a novel locus. Our results suggest that the three loci identified here may contain central clock or clock-related genes, at least one of which may be a novel.

Introduction

The circadian clock regulates circadian rhythms – the endogenous daily fluctuations in physiological and biological activities that are observed in organisms from cyanobacteria to humans (Dunlap, 1999). In higher plants, the clock regulates many processes, including stomatal aperture, leaf movement, hypocotyl elongation, photosynthetic activity, and photoperiodic flowering induction (Lumsden and Millar, 1998; McClung, 2001). In Arabidopsis thaliana, the clock regulates about 6% of the genes that are expressed rhythmically in constant light (Harmer et al., 2000).

Central clock genes that generate circadian oscillations have been isolated in Drosophila melanogaster (Jackson et al., 1986), Neurospora crassa (McClung et al., 1989), cyanobacteria (Ishiura et al., 1998), and the mouse (Sun et al., 1997; Tei et al., 1997). The mechanism proposed for the generation of circadian oscillations common to all those organisms is negative and positive feedback control of clock genes (Dunlap, 1999; Ishiura et al., 1998; Stanewsky, 2003). In Arabidopsis, the CIRCADIAN CLOCK ASSOCIATED 1 (CCA1; Wang and Tobin, 1998), LATE ELONGATED HYPOCOTYL (LHY; Schaffer et al., 1998) and TIMING OF CAB EXPRESSION 1 (TOC1/APRR1; Makino et al., 2000; Strayer et al., 2000) genes may be involved in clock feedback (Hayama and Coupland, 2003; Millar, 1999; Yanovsky and Kay, 2003). Expression of CCA1 and LHY is out of phase with expression of TOC1. CCA1 and LHY expression peaks in the morning whereas TOC1 expression peaks in the evening. TOC1 maintains CCA1 and LHY mRNA levels whereas CCA1 and LHY repress TOC1 transcription (Alabadi et al., 2001). However, there are some contradictions: (i) cca1-1 and lhy-12, null mutations of the CCA1 and LHY genes (Green and Tobin, 1999; Mizoguchi et al., 2002), respectively, and toc1-2, a probable null mutation of the TOC1 gene (Alabadi et al., 2002; Strayer et al., 2000), do not cause the loss of circadian rhythms (arrhythmia), although they do shorten the period, (ii) cca1-1 lhy-12, a double null mutation of the CCA1 and LHY genes (Mizoguchi et al., 2002), and cca1-1 lhy-RNAi (Alabadi et al., 2002) also do not cause arrhythmia, although they cause rhythms to dampen within a few days, (iii) TOC1 overexpression does not cause arrhythmia nor increase CCA1 and LHY transcription (Makino et al., 2002), (iv) LHY overexpression does not eliminate rhythmic accumulation of EARLY FLOWERING 3 (ELF3) mRNA (Hicks et al., 2001). These observations are not consistent with the current Arabidopsis circadian feedback model.

Many Arabidopsis genes that affect circadian rhythms, such as PHYTOCHROME B (PHYB), SENSITIVITY TO RED LIGHT REDUCED 1 (SRR1), ELF3, ELF4, LHY, and GIGANTEA (GI), have been isolated (Doyle et al., 2002; Hayama and Coupland, 2003; Lumsden and Millar, 1998; Staiger et al., 2003; Yanovsky and Kay, 2003), and most of them were isolated as photoreceptor genes, photosignal transduction genes or flowering time genes. Millar et al. (1995) screened 8000 M2Arabidopsis plants for clock mutants and Saloméet al. (2002) screened 6500, and they isolated several interesting rhythm mutants, such as toc1 (Somers et al., 1998), ztl (Somers et al., 2000), tic (Hall et al., 2003), ctd (Saloméet al., 2002), and oop1 (Saloméet al., 2002). However, those screenings were inadequate for the isolation of all possible mutants. A more comprehensive, large-scale screening is required. To that end, we developed a high-throughput real-time bioluminescence monitoring system. The system (K. Okamoto, K. Onai, and M. Ishiura, unpublished results) consists of two high-throughput bioluminescence-monitoring apparatuses and an analyzing program. The system enables automatic measurement and analyzing of bioluminescence rhythms of 2880 plants in one assay under uniform light conditions.

Here we constructed two new bioluminescence reporter genes as follows. We fused the promoter of a clock-regulated Arabidopsis flowering-time gene – either GI (Fowler et al., 1999; Park et al., 1999) or FLOWERING LOCUS T (FT; Kardalisky et al., 1999; Kobayashi et al., 1999) – to a coding region of a modified firefly luciferase gene (LUC+) and transferred them into the Arabidopsis genome. Using our monitoring system, we screened 50 000 M2 plants carrying PGI::LUC+ and 50 000 carrying PFT::LUC+ for mutants whose circadian rhythm differed from the wild type by more than 3 h in period length or phase (the time of day at which the first bioluminescence peak appears). We isolated 29 such mutants and six arrhythmic (AR) mutants, and analyzed five of the AR mutants genetically.

Results

Measurement of the circadian bioluminescence rhythms of PGI::LUC+ and PFT::LUC+ reporter strains by a high-throughput bioluminescence monitoring system

We named the Arabidopsis reporter strain carrying PGI::LUC+ strain G-38 and the one carrying PFT::LUC+ strain F-1 (Figure 1a). We measured the bioluminescence from the seedlings of the two reporter strains grown under constant light from white fluorescent lamps at 50 μmol m−2 sec−1 (LL conditions) with our high-throughput monitoring system. Both strains showed robust circadian bioluminescence rhythms with different wave forms for at least 9 days (Figure 1b,c). The wave form in strain G-38 was symmetric but not a cosine curve, and the peaks were sharp (Figure 1b). The period length (mean ± SD) was 23.1 ± 0.5 h (range: 20.6–25.3 h) (Figure 1d) and the phase was 9.9 ± 0.5 h (n = 1,920). The wave form in strain F-1 was saw-toothed; it increased rapidly and decreased slowly (Figure 1c). The mean period length and its range was the same as for G-38 (23.1 ± 0.5 h; range: 21.6–25.3 h), but the phase was 3.4 ± 1.4 h (n = 552) and the phase distribution was wider than that of strain G-38 (Figure 1e,f). Similar results were obtained from other transgenic strains carrying the PGI::LUC+ or PFT::LUC+ reporter gene, although levels of bioluminescence differed. These results indicate that the PGI::LUC+ and PFT::LUC+ reporter genes were expressed rhythmically in a circadian manner when grown in LL. Furthermore, the bioluminescence rhythms from strains G-38 and F-1 satisfied other criteria for circadian rhythms as well: period lengths were compensated against changes in ambient temperature, and the phases were reset by light or temperature (data not shown).

Figure 1.

Structure of PGI::LUC+ and PFT::LUC+ reporter constructs and bioluminescence rhythms of Arabidopsis PGI::LUC+ and PFT::LUC+ reporter strains.
(a) NPTII, a kanamycin-resistance gene cassette in pBI101; PGI, a promoter region of the Arabidopsis GI gene (−3720 to −1); PFT, a promoter region of the Arabidopsis FT gene (−4241 to −1), PCAB2, a promoter region of the Arabidopsis CAB2 gene (−529 to −1); LUC+, a coding region of a modified luciferase gene derived from Photinus pyralis; Rluc, a coding region of a luciferase gene derived from Renilla reinformis, TNOS, transcriptional terminator sequences of the NOS (nopaline synthase) gene from Agrobacterium; LB, left-border sequences of T-DNA; RB, right-border sequences of T-DNA. The transcriptional directions of gene cassettes are shown by the arrows.
(b, c) Circadian bioluminescence rhythms of wild-type reporter strains G-38 (b, PGI::LUC+ reporter) and F-1 (c, PFT::LUC+ reporter). Bioluminescence rhythms of each strain were examined in LL.
(d) Distribution of the period length of rhythms of PGI::LUC+ reporter strain in LL.
(e, f) Distributions of the phase of rhythms of strains G-38 (n = 1920) (e) and F-1 (n = 552) (f) in LL.

Large-scale screening of circadian clock mutants

We screened clock mutants whose bioluminescence rhythms in LL differed from the wild-type by at least 3 h in period length or phase as follows. We treated the seeds of the reporter strains with the mutagen ethylmethylsulfonate and then screened about 50 000 M2 seedlings from each strain ( totaling 100 000 M2 seedlings). We isolated a total of 147 candidate rhythm mutants and confirmed the rhythm phenotypes of homozygotes in the M3 or M4 generation. We could not obtain M3 seeds for 54 of 147 candidates because of abnormal growth or infertility, and 58 candidates did not meet our criteria (period length and phase) for rhythm mutants. The phenotypes of the remaining 35 rhythm mutants – 20 from G-38 and 15 from F-1 – are listed in Table 1. Six of 35 rhythm mutants show AR phenotype. The shortest and longest periods were 16.3 ± 0.8 and 27.0 ± 0.3 h, respectively, and some of the period mutants were also phase mutants. Bioluminescence rhythms of mutants with an altered period or phase were shown in Figure 2.

Table 1.  Homozygous rhythm mutants and their bioluminescence rhythms in LL
MutantReporterGenerationTypePeriod lengtha (h)Phasea (h)n
  1. LP, long period; SP, short period; P, phase alteration; AR, arrhythmic.

  2. aValues are shown as average ± SD.

22-6A9PGI::LUC+M3LP26.6 ± 0.812.2 ± 1.341
23-14H4PGI::LUC+M3SP20.1 ± 1.39.2 ± 0.940
23-15D9PGI::LUC+M3AR58
25-14C2PGI::LUC+M3AR77
25-16F5PGI::LUC+M3LP, P26.5 ± 1.014.2 ± 2.065
26-14E3PGI::LUC+M3AR70
26-18C12PGI::LUC+M3AR57
29-7E11PGI::LUC+M3LP26.2 ± 0.712.0 ± 1.247
29-10G10PGI::LUC+M3SP17.1 ± 0.58.6 ± 0.796
30-13E6PGI::LUC+M3SP18.5 ± 1.18.7 ± 0.948
31-16H6PGI::LUC+M3LP26.7 ± 1.711.7 ± 1.040
32-5E2PGI::LUC+M3AR80
33-14A4PGI::LUC+M3SP20.0 ± 0.89.0 ± 0.853
35-9E4PGI::LUC+M3SP18.4 ± 0.88.7 ± 0.644
35-9H2PGI::LUC+M3SP17.3 ± 1.08.1 ± 0.854
35-10B5PGI::LUC+M3SP17.5 ± 0.78.2 ± 0.654
38-11A8PGI::LUC+M3SP19.6 ± 0.78.8 ± 1.041
38-11H1gPGI::LUC+M3SP18.5 ± 1.57.5 ± 1.712
39-3H2PGI::LUC+M3LP26.7 ± 0.912.2 ± 0.621
43-5A1PGI::LUC+M3LP, P27.0 ± 1.413.4 ± 1.344
45-3H6PFT::LUC+M3LP26.4 ± 1.13.8 ± 1.038
45-7G6PFT::LUC+M3SP18.4 ± 1.05.3 ± 1.072
45-12A6PFT::LUC+M3AR32
45-14B7PFT::LUC+M3LP, P26.2 ± 1.28.1 ± 1.550
47-16H1PFT::LUC+M3SP19.0 ± 1.04.6 ± 0.778
50-5A9PFT::LUC+M3SP18.0 ± 0.95.2 ± 0.879
50-2G3PFT::LUC+M4SP17.8 ± 0.55.6 ± 1.430
51-5A4PFT::LUC+M3SP16.3 ± 0.85.0 ± 1.944
60-7A1PFT::LUC+M3LP, P26.6 ± 0.99.6 ± 2.834
60-7H5PFT::LUC+M3LP26.1 ± 0.84.0 ± 0.750
64-12D10PFT::LUC+M3P25.0 ± 0.67.4 ± 1.338
64-12F2PFT::LUC+M3P23.9 ± 0.66.5 ± 1.646
66-6C7PFT::LUC+M3P25.2 ± 1.110.6 ± 1.860
66-7B4PFT::LUC+M3LP27.0 ± 0.35.0 ± 1.45
69-12H11PFT::LUC+M3LP, P26.1 ± 1.47.1 ± 1.947
Figure 2.

Bioluminescence rhythms of mutants with an altered period or phase in LL. Bioluminescence rhythms of wild-type strain G-38 (a) and homozygous F3 generation of mutant 29-10G10 with an extremely short period (b), mutant 43-5A1 with an extremely long period and phase-alteration (c) and mutant 25-16F5 with a long period and large phase-alteration (d) were examined in LL. All strains carried the PGI::LUC+ reporter gene. The traces represent about 48 results in each strain, and similar traces were obtained from the results.

Bioluminescence rhythms of AR mutants in LL and DD

Because arrhythmia is the severest clock effect, we examined the bioluminescence of the five AR mutants carrying the PGI::LUC+ reporter gene (23-15D9, 25-14C2, 26-14E3, 26-18C12, and 32-5E2 in Table 1) in LL (Figure 3) and constant dark (DD) (Figure 4). While the bioluminescence from wild-type strain G-38 oscillated rhythmically under both conditions (Figures 3a and 4a) with period lengths of 23.1 ± 0.5 h in LL and 28.2 ± 1.4 h in DD, the five AR mutants showed arrhythmia in both LL and DD (Figures 3b–f and 4b–f). In LL, the levels of bioluminescence in the five AR mutants were constitutively higher than that in the parent strain, suggesting that their arrhythmia resulted from lack of repression of the reporter gene and that their arrhythmia was not affected by light.

Figure 3.

Arrhythmia (AR) of bioluminescence rhythms in AR mutants in LL. Bioluminescence rhythms of wild-type reporter strain G-38 (a) and homozygous F3 generation of AR mutants, 23-15D9 (b), 25-14C2 (c), 26-14E3 (d), 26-18C12 (e), and 32-5E2 (f), were examined in LL. All strains carried the PGI::LUC+ reporter gene. The traces represent about 48 results in each strain, and similar traces were obtained from the results.

Figure 4.

Arrhythmia (AR) of bioluminescence rhythms in AR mutants in DD. Strains and labels of each panel are as described in Figure 3. Seedlings were allowed to germinate for 4 days in LL and were then exposed to three cycles of LD at 22.0 ± 0.5°C to synchronize the circadian clock. After the third light period, plates containing seedlings were returned to DD and the bioluminescence from each seedling was measured in DD. The traces represent 48 results in each strain, and similar traces were obtained from the results.

Genetic analyses of AR mutants

To determine whether the rhythm mutations in the five AR mutants were recessive or dominant, we crossed each homozygous mutant with wild-type G-38 and, in LL, examined the bioluminescence rhythms of the F1 progeny (Table 2). Rhythms were consistently restored (although the period was slightly shortened), indicating that the AR mutations were recessive.

Table 2.  Bioluminescence rhythms in LL in the F1 progeny of crosses between arrhythmic mutants and wild-type strain G-38
Pollen donorPollen recipientMean bioluminescence rhythm ± SDn
Period (h)Phase (h)
  1. Period length and phase in wild-type reporter strain G-38 (n = 48) were 23.2 ± 0.6 and 9.4 ± 0.7 h, respectively.

23-15D9G-3822.1 ± 0.68.8 ± 0.750
25-14C2G-3821.1 ± 1.09.7 ± 0.957
26-14E3G-3822.7 ± 0.79.1 ± 1.38
26-18C12G-3822.0 ± 1.29.8 ± 1.057
32-5E2G-3822.3 ± 0.49.3 ± 0.487

To determine whether the AR phenotypes were the result of single-gene mutations, we self-pollinated the heterozygous F1 generation and, in LL, examined the bioluminescence rhythms of the F2 progeny. We scored the segregation patterns of AR versus rhythmic phenotypes and analyzed them by the chi-square test for goodness-of-fit (Table 3). The AR phenotypes in the F2 progeny of all five AR mutants segregated with a 1:3 ratio, suggesting that the phenotypes were the result of single nuclear gene mutations.

Table 3.  Segregation of arrhythmic phenotypes in the F2 progenies from crosses between arrhythmic mutants and wild-type strain G-38
Pollen donorPollen recipientNumber of plantsArrhythmic:rhythmic = 1:3 [χ2 (P-value)]
ArrhythmicRhythmic
  1. Bioluminescence rhythms were examined in LL.

23-15D9G-38431230.03 (>0.85)
25-14C2G-3827690.35 (>0.55)
26-14E3G-3819761.19 (>0.27)
26-18C12G-3817580.18 (>0.66)
32-5E2G-38391431.17 (>0.27)

To examine the complementarity of the five AR mutations, we crossed the homozygous AR mutants (F3 generation) in all possible combinations and, in LL, examined the bioluminescence rhythms of the progeny (Table 4). Crosses between 23-15D9 and 32-5E2 and between 26-14E3 and 26-18C12 did not restore rhythm in the F1, whereas the other crosses did, indicating that the mutations belonged to three complementation groups.

Table 4.  Complementation analysis of arrhythmic mutants (n = 48)
Pollen donorPollen recipient
23-15D925-14C226-14E326-18C1232-5E2
  1. Bioluminescence rhythms were examined in LL.

  2. +, complemented (rhythmic); −, not complemented (arrhythmic).

23-15D9 +++
25-14C2+ +++
26-14E3++ +
26-18C12++ +
32-5E2+++ 

Leaf movement rhythms of AR mutants

To determine whether mutations at three AR loci affected other circadian outputs, we examined the leaf movement rhythms of the homozygous F3 progeny of mutants 23-15D9, 25-14C2, and 26-18C12. We measured leaf movement rhythms by tracing the vertical position of cotyledon tips by CCD video imaging (Figure 5), as described previously (Dowson-Day and Millar, 1999).

Figure 5.

Imaging system for the measurement of the leaf movements in Arabidopsis seedlings.
(a) Schematic diagram of an imaging stand. The imaging stand was constructed by a camera holder (Ch), a XZ-stage with a dish holder (XZs), a shade (Sh), an optical bench (Ob) rigidly holding a CCD video camera (Cam) with a macro lens (Ml), and a square Petri dish (Spd). Two imaging stands were set on a growth chamber (Gc) with illumination from three fluorescent white lamps (Fl).
(b) Overview of a imaging stand.
(c) A video captured image of a seedling at initiation time point of recording in LL and traces of positions of cotyledon tips (open circles).

Wild-type G-38 seedlings showed rhythmic leaf movement, and the period length was 1 h longer (24.1 ± 0.2 h, n = 6) than that of their bioluminescence rhythms (Figure 6a). We obtained similar results from wild-type strain Col-0 without the LUC+ reporter genes (data not shown). The three AR mutants, however, did not show rhythmic leaf movement (Figure 6b–d), indicating that the two circadian outputs, bioluminescence and leaf movement, were regulated by the three AR loci identified here.

Figure 6.

Arrhythmia (AR) of leaf movement rhythms in AR mutants in LL. Leaf movement rhythms of wild-type reporter strain G-38 (a) and homozygous F3 generation of AR mutants, 23-15D9 (b), 25-14C2 (c), and 26-18C12 (d) were examined in LL. The vertical positions of cotyledon tips at the initial time point in each strain are plotted as zero. The traces represented zero results in each strain, and similar traces were obtained from the results.

Discussion

Arrhythmia is the severest phenotypic clock effect and is caused by recessive mutations in central clock genes such as per01, tim01, Clkar, and cyc01 in D. melanogaster (Stanewsky, 2003) and frq9, wc-1, and wc-2 in N. crassa (Loros and Dunlap, 2001). In the cyanobacterium Synechococcus sp. strain PCC 7942, the loss of any one of the central clock genes –kaiA, kaiB, or kaiC– results in the AR phenotype (Ishiura et al., 1998). Here, we isolated six new AR mutants from 100 000 M2 plants, and five of them with the PGI::LUC+ reporter gene showed no bioluminescence rhythms when grown in either LL or DD (Figures 3 and 4). The AR mutations were recessive and located at three different nuclear loci (Tables 2–4). The mutations also caused the loss of leaf movement rhythms (Figure 5). Therefore, the three AR loci identified here are central clock or closely related genes and play very important roles in the progression of the circadian clock in Arabidopsis.

Several elf3 mutants of Arabidopsis did not show rhythmic gene expression and leaf movement rhythms in LL, but they did in DD (Covington et al., 2001; Hicks et al., 1996). AR phenotype of the five AR mutants described here, in contrast, was not affected by light (Figures 3 and 4), suggesting that the three AR loci may be different from the ELF3 locus. In fact, at least one of the AR loci was mapped to a different locus than those of known clock-affecting gene including ELF3 (data not shown).

Many Arabidopsis rhythm mutants that show alterations in period length or phase have been reported (Hall et al., 2003; Lumsden and Millar, 1998; McClung et al., 2002; Staiger et al., 2003), but the extent to which they differ from wild-type strain is small. For example, cca1-1 and lhy-12, null alleles of the CCA1 and LHY genes, show at most a 3-h period shortening (Green and Tobin, 1999; Mizoguchi et al., 2002). Here, we isolated 29 mutants that showed extreme alterations in period length or phase (Table 1 and Figure 2), including mutant 29-10G10, whose nuclear, semi-dominant mutation caused a 6-h period shortening under LL conditions (Table 1 and Figure 2b). Mutant phenotypes of extremely short or extremely long period length have been reported in the semi-dominant alleles involved in the central clock in many organisms, including Drosophila and Neurospora (Loros and Dunlap, 2001; Stanewsky, 2003). Extremely altered period length or phase, such as observed in mutant 29-10G10, may be caused by a defect in one of the central clock genes. One of the many mutants isolated here, 45-14B7, showed a novel clock phenotype with a long period and an altered phase. Its clock was reset at random times of subjective day and night by temperature (data not shown). This mutation may provide a clue to the molecular mechanisms involved in temperature-resetting of the circadian clock. Further genetic and physiological analyses are needed to classify the many circadian rhythm mutants isolated here and to determine whether they are new genes or just new alleles at known loci.

We used the PGI::LUC+ and PFT::LUC+ reporter genes in this study. GI and FT are key genes that promote flowering (Fowler et al., 1999; Kardalisky et al., 1999; Kobayashi et al., 1999). GI acts upstream and FT acts downstream of the flowering regulator gene CONSTANS (Huq et al., 2000;Suárez-López et al., 2001). We are now analyzing the effects of the mutations isolated here on flowering time.

We were able to achieve the comprehensive screening of clock mutants described here in a relatively short amount of time because of our high-throughput real-time bioluminescence monitoring system. We were able to screen 100 000 M2 plants in 25 weeks and this scale should be enough to isolate clock mutants comprehensively in our experimental conditions as evidenced by the allelic mutations we identified from the five AR mutants. The screening, however, could have been biased by our experimental conditions, such as the character of the reporter genes and the plants’ ecotype, age, and growth conditions. For that reason, we plan additional screenings to isolate other types of clock mutants under different experimental conditions.

The positional cloning of several mutated genes identified here, including AR genes, are in progress. We expect the cloning and characterization of the genes to extend our understanding of the molecular mechanisms for the circadian clock in higher plants and reveal unknown aspects of the mechanisms.

Experimental procedures

Plant materials and growth conditions

Arabidopsis thaliana (L.) Heynh (ecotype Col-0) seeds were a gift from Dr M. Ohto (University of California, Davis, CA, USA) and used for construction of transgenic lines. Surface sterilized seeds of Arabidopsis were incubated at 4°C in DD for 2 days and grown on Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) that contained 1.5% (w/v) sucrose and 0.3% (w/v) Gelrite (hereafter called MS-1.5/0.3) at 22.0 ± 0.5°C in LL. To obtain seeds, we grew Arabidopsis plants on MS-1.5/0.3 for 14–18 days and then on vermiculite at 22.0 ± 0.5°C in LL from white fluorescent lamps at a light intensity of 100 μmol m−2 sec−1.

PGI::LUC+ and PFT::LUC+ reporter constructs

Using standard DNA manipulation and sequencing techniques (Ausubel et al., 1987; Sambrook et al., 1989), we confirmed the nucleotide sequences of all constructs. PGI::LUC+ and PFT::LUC+ reporter gene cassettes were inserted just upstream of a PCAB2::Rluc gene cassette of binary vector pBIC2RL (K. Onai and M. Ishiura, unpublished results), giving pBIGLC2RL and pBIFLC2RL, respectively (Figure 1a). pBIC2RL carried the pBI101 (Jefferson et al., 1987) backbone and the PCAB2::Rluc gene cassette in place of both the β-glucuronidase coding region (gus) and the transcriptional terminator sequence of the Agrobacterium nopaline synthase gene (TNOS). The PCAB2::Rluc gene cassette consisted of the Arabidopsis CAB2 promoter (PCAB2; nucleotides −529 to −1), the coding region of the Renilla reinformis luciferase gene (Rluc; Promega, Tokyo, Japan), and TNOS. The GI::LUC+ and FT::LUC+ reporter gene cassettes consisted of an Arabidopsis GI promoter (PGI; −3720 to −1) or an Arabidopsis FT promoter (PFT; −4241 to −1), the coding region of a modified firefly luciferase gene (LUC+; Promega), and TNOS. Firefly luciferase (LUC+) and Renilla luciferase (RLUC) require different substrates d-luciferin and coelenterazine, respectively. In this study, we used only d-luciferin as a substrate.

Construction of Arabidopsis reporter strains and mutagenesis

We transferred the T-DNA regions of pBIGLC2RL and pBIFLC2RL into the Arabidopsis genome by the Agrobacterium tumefaciens-mediated floral dip method (Clough and Bent, 1998). We used A. tumefaciens strain GV3101::pMP90 because its high transformation potency in Arabidopsis (Koncz and Schell, 1986; Weigel and Glazebrook, 2002). Kanamycin-resistant (KmR) T1 plants were selected and grown by standard techniques (Weigel and Glazebrook, 2002). We selected T2 plants that showed a 3:1 segregation ratio in both KmR:KmS and bioluminescence:non-bioluminescence, and therefore contained a single T-DNA in the genome, and we obtained T3 plants from each T2 plant selected. We judged T3 plants carrying a single T-DNA as homozygous by the non-segregation of both KmR and bioluminescence phenotypes and confirmed them by Southern blot analysis. T4 seeds were generated on a large scale from each T3 plant selected. The PGI::LUC+ and PFT::LUC+ reporter plants were selected as wild-type reporter strains G-38 and F-1, respectively, and used for mutagenesis. About 10 000 T4 seeds from each strain were mutagenized by treatment with 0.3% (v/v) ethyl methanesulfonate (Sigma, Tokyo, Japan) for 14 h at 25°C. M2 seeds were collected and grouped into 100 pools: each pool contained seeds from about 100 M1 plants. About 500 seedlings in each M2 pool were screened for mutants that showed altered circadian bioluminescence rhythms.

Measurement of circadian bioluminescence rhythms

We assayed the circadian bioluminescence rhythms from seedlings as follows. Surface-sterilized seeds were incubated at 4°C in DD for 2 days, sown on MS-1.5/0.3, and germinated in LL. We used an automated apparatus (S-Feeder Type B; MK Research Co. Ltd., Tokyo, Japan) that enabled us to sow seeds on medium at 5 sec per seed. Four days after sowing, we used a glass tube (6 mm internal diameter) to punch out agar plugs containing a single seedling and transferred each into a well of a 96-well microplate (CulturePlate-96; Perkin–Elmer, Tokyo, Japan) containing 20 μl liquid MS medium supplemented with 1.5% (w/v) sucrose. We added 20 μl d-luciferin-K (Biosynth, Naperville, IL, USA) dissolved in sterilized Milli-Q water (Millipore, Tokyo, Japan) to a final concentration of 100 μm, and sealed the plates with a plate-seal (Plate Seal T, Sanplatec Corp., Osaka, Japan). To synchronize the circadian clock, we incubated the plates at 22.0 ± 0.5°C for three cycles of 12 h of light (50 μmol m−2 sec−1) alternating with 12 h of darkness (LD). After the third dark period, the plates were returned to LL and the bioluminescence from each well was measured continuously by two kinds of automated bioluminescence monitoring apparatuses (K. Okamoto, K. Onai, and M. Ishiura, unpublished results). One apparatus consisted of TopCount NXT (Perkin–Elmer) carrying six photomultiplier tubes as photon detectors, a plate platform, and a plate-conveying robot. Another apparatus integrated the three units, a photon detecting unit consisting of eight photomultiplier tubes and an eight-channel counter, a plate platform, and a plate-conveying unit. In each measurement, bioluminescence was measured for 3 sec in the dark after an exposure to a 240-sec dark period for decreasing the delayed light emission of chlorophyll, and the plates were returned to LL. This measuring sequence was repeated automatically every 60–130 min. Although it is possible that an exposure to the short dark period could have minor effects on circadian rhythms (Broda et al., 1986), we could not detect any interferences on the bioluminescence rhythms described here. Bioluminescence data were stored in a microcomputer and analyzed by the cosinor method based on the Fourier transform (Nelson et al., 1979) or the visual-inspection method with linear regression (Onai et al., 2004; Pittendrigh and Daan, 1976), using a rhythm analyzing program rap (K. Okamoto, K. Onai, and M. Ishiura, unpublished results). Phase was defined as the time of day at which the first peak of bioluminescence appeared. After bioluminescence rhythms were measured, we transferred the seedlings from the plates to MS-1.5/0.3 and grew them in LL for 7 days, then transferred them to soil and grew them at 22.0 ± 0.5°C in LL from white fluorescent lamps at a light intensity of 100 μmol m−2 sec−1.

Measurement of circadian leaf movement rhythms

We assayed the circadian leaf movement rhythms of seedlings as follows. Surface-sterilized seeds were incubated in DD for 2 days at 4°C, sown on MS-1.5/0.3, and allowed to germinate for four LD cycles. Using a surgical knife, we cut out six agar plugs containing a single-seedling each and transferred them to individual compartments of 25-compartment square Petri dishes (Bibby Sterilin Ltd., Stone, Staffordshine, UK). The two dishes were placed vertically in LL in a modified growth chamber (MIR-153; Sanyo, Osaka, Japan) supplemented with a light control unit. The images of the 12 seedlings were converted to video signals by two monochrome CCD video cameras (B05-3M; Vixen, Tokorozawa, Japan) and recorded automatically as an 8-bit grayscale TIFF file (1280 × 480 pixels) at 30 min intervals for 4 days by a video capture board (LG-3; Scion Corp., Frederick, MD, USA) and the nkcap program based on Scion image (beta 4.0.2; Scion Corp.). The vertical positions of cotyledon tips in each TIFF file were detected automatically by the nktrace program based on igor pro (release 4.0.7.0; WaveMetrics, Lake Oswego, OR, USA) and stored as a CVS file. The vertical positions of cotyledon tips were analyzed by the cosinor method with the rap program.

Acknowledgements

We thank Masa-aki Ohto (University of California, Davis) for the kind gift of Arabidopsis Col-0 seeds and binary vector pBI101 and for many helpful suggestions; Atsushi Morikami (Chubu University), Ken Matsuoka (RIKEN), and Kenzo Nakamura (Nagoya University) for kind gifts of Agrobacterium strains and several vectors and for technical advice on the construction of reporter genes; Chieko Namba (NIBB) for technical assistance in breeding transgenic lines; Kazunari Suzuki, Haruhito Takaoka, and Takumi Chiba (MK Research) for technical support and customization of S-Feeder; and Miriam Bloom (SciWrite Biomedical Writing and Editing Services) for professional editing. This study was supported by grants to MI from the following sources: The Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT), ‘The Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN)’ promoted by BRAIN, ‘Research for the Future Novel Gene Function Involved in Higher-Order Regulation of Nutrition-Storage in Plants’ promoted by the Japan Society for the Promotion of Science, ‘Ground-based Research for Space Utilization’ promoted by the Japan Space Forum, and ‘The Promoting Cooperative Research Project’ sponsored by the Aichi Science and Technology Foundation. The Division of Biological Science, Graduate School of Science, Nagoya University is supported by a 21st COE grant from MEXT.

Ancillary

Advertisement