SEARCH

SEARCH BY CITATION

Keywords:

  • Arabidopsis;
  • flowering time;
  • GIGANTEA;
  • long day;
  • photoperiod

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References

Flowering of Arabidopsis is promoted by long days and delayed by short days. Mutations in the GIGANTEA (GI) gene delay flowering under long days but have little or no effect under short days. We have now isolated the GI gene and show that it encodes a novel, putative membrane protein. By comparing the sequence of the Arabidopsis gene with that of a likely rice orthologue and by sequencing mutant alleles, we identify regions of the GI protein that are likely to be important for its function. We show that GI expression is regulated by the circadian clock with a peak in transcript levels 8–10 h after dawn. The timing, height and duration of this peak are influenced by daylength. We analysed the interactions between GI and the LHY, CCA1 and ELF3 genes, previously shown to affect daylength responses; we show that the rhythmic pattern of GI expression is altered in the elf3, CCA1-OX and lhy genotypes, and that CCA1 and LHY expression are reduced by gi mutations. Our results are consistent with the idea that GI plays an important role in regulating the expression of flowering time genes during the promotion of flowering by photoperiod.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References

The reproduction and behaviour of many organisms are synchronized to favourable seasons of the year (Saunders, 1977). Many species recognize particular seasons by measuring daylength (photoperiod). Photoperiod controls a wide range of processes ranging from migratory restlessness and growth of sexual organs in birds to flowering in plants (Saunders, 1977). Arabidopsis thaliana L. Heynh (Arabidopsis) provides a model system in which to identify genes involved in photoperiodic flowering. Arabidopsis flowers rapidly in long days (LD, 16 h light, 8 h dark), but has delayed flowering in short days (SD, 8 h light, 16 h dark). Genetic screens have identified mutations that cause Arabidopsis to flower with reduced sensitivity to daylength. In genetic models of flowering in Arabidopsis, these genes have been placed in a single genetic pathway, often called the long-day flowering pathway, because it was first defined by a class of mutations that delayed flowering under long but not short days (Redei, 1962; Koornneef et al., 1991). Several other Arabidopsis flowering time genes are placed in a second genetic pathway, called the autonomous or constitutive flowering pathway, because they affect flowering time in both photoperiods (see recent reviews by Koornneef et al., 1998b; Levy and Dean, 1998; Piñeiro and Coupland, 1998).

Interactions between photoreceptors and the circadian clock are thought to allow plants to distinguish between different daylengths. The circadian clock is an endogenous timekeeper that controls many rhythmic processes (circadian rhythms) in organisms as they experience the 24 h cycle of day and night (Dunlap, 1999). Several genes in the long-day flowering pathway affect both circadian rhythms and flowering time in Arabidopsis. Plants carrying mutations in the EARLY FLOWERING 3 (ELF3) gene have no detectable circadian rhythms in continuous light conditions, but do in darkness (Hicks et al., 1996). This suggests that ELF3 may be responsible for light input signals to the clock. The elf3 mutant also shows early, daylength-insensitive flowering indicating that ELF3 functions to measure daylength and to repress flowering under short days (Zagotta et al., 1996). Late flowering plants resulted from overexpression of either of two related genes encoding proteins with a single Myb repeat (LATE ELONGATED HYPOCOTYL, LHY, Schaffer et al., 1998; and CIRCADIAN CLOCK ASSOCIATED 1, CCA1, Wang and Tobin, 1998). In both genotypes, circadian clock-controlled expression of several genes was disrupted. The LHY and CCA1 genes were proposed to function either within the circadian oscillator or in output pathways from it. Finally, mutations in the TIMING OF CAB 1 gene (TOC1, Somers et al., 1998), which is thought to be part of the clock, accelerate circadian rhythms in Arabidopsis and can lead to flowering with reduced sensitivity to photoperiod in the Landsberg erecta (Ler) ecotype of Arabidopsis by causing early flowering under short days.

Other genes that act within the long-day pathway and affect flowering time apparently do not affect circadian clock function. The CONSTANS flowering time gene (CO, Redei, 1962; Koornneef et al., 1991) has been placed in the long-day flowering pathway, because co mutations lead to late flowering in long days, but have little effect on flowering in short days. Hence the CO gene is thought to promote flowering in long days. The CO gene has been isolated and encodes a putative zinc finger transcription factor (Putterill et al., 1995). CO appears to promote flowering via up-regulating the activity of the Arabidopsis floral meristem identity gene LEAFY (Weigel et al., 1992; Simon et al., 1996; Nilsson et al., 1998). Six other mutants (gi, fd, fe, fha, ft and fwa) are late flowering, have little or no response to daylength and fall into the same epistatic group as co in genetic tests (Koornneef et al., 1998a). These genes are thought to function mainly to promote flowering in long days (Koornneef et al., 1998a,b). One of the genes, FHA, encodes a flavin-containing blue light receptor involved in floral promotion (Koornneef et al., 1991; Lin et al., 1996; Guo et al., 1998).

The Arabidopsis GIGANTEA gene (GI) appears to act upstream of CO, as the late flowering phenotype of gi mutants is corrected by ectopic overexpression of the CO gene (I.Igeno, F.Robson and G.Coupland, unpublished data). The gi mutation has pleiotropic effects in addition to its effect on flowering time. For example, gi mutants have elongated hypocotyls (Araki and Komeda, 1993) and are resistant to the herbicide paraquat (Kurepa et al., 1998). Furthermore, gi mutant plants accumulate starch under some conditions, but the high starch levels are not the cause of the late flowering phenotype of gi mutants (Eimert et al., 1995; Hollis, 1999).

To analyse the molecular role of GI in Arabidopsis flowering, we have isolated and characterized the gene. Here we report that GI encodes a novel, putative membrane protein. We show that GI transcript levels are controlled by the circadian clock. The ELF3, CCA1 and LHY genes are involved in its circadian regulation. Altered timing and increased levels of GI gene expression correlate with early daylength-insensitive flowering in the elf3 mutant. Finally, we show that two genes influencing flowering in Arabidopsis, LHY and CCA1 are expressed at lower levels in the gi-3 mutant in long days.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References

Identification of the sequence of the GI gene

Previously, we identified a new gi mutant allele (now gi-11) carrying a T-DNA tag closely linked to the gi mutation (Richardson et al., 1998). Southern hybridization analysis using a probe from the left border of the T-DNA confirmed that there was a single T-DNA insertion in the genome (data not shown). Inverse polymerase chain reaction (IPCR) was used to amplify a 1.3 kb fragment flanking the left border of the T-DNA (see Figure 1A). The DDBJ/EMBL/GenBank DNA sequence database was searched with the sequence derived from the IPCR fragment. This revealed that the IPCR fragment sequence was identical to sequence of an Arabidopsis cosmid clone CC17J13 (DDBJ/EMBL/GenBank accession No. Y12227, Terryn et al., 1997) that is located on chromosome 1. This cosmid was contained on a sequenced bacterial artificial chromosome (BAC), T22J18 (DDBJ/EMBL/GenBank accession No. AC003979). Our previous genetic mapping experiments showed that GI mapped 0.36 cM from the restriction fragment length polymorphism (RFLP) marker TH1 (see Materials and methods), and TH1 is located on the adjacent sequenced BAC F19G10 (DDBJ/EMBL/GenBank accession No. ATAF000657), ∼50 kb from the cosmid sequence. Therefore, the CC17J13 sequence is present at a position consistent with high resolution mapping of GI.

image

Figure 1. Structure of the GI locus in mutant and wild-type plants and analysis of the GI protein. (A) Structure of the GI locus in the vicinity of the T-DNA insertion in gi-11. The black bar on the line represents the extent of the deletion caused by insertion of the T-DNA in gi-11 and the grey bar represents the left border IPCR fragment. The positions of XbaI (X) restriction enzyme sites are shown. The position of the λ3 clone spanning the T-DNA insertion and the full-length GI cDNA are shown as grey bars below the line. Gene 5 and the other genes predicted in this region by Terryn et al. (1997) are shown as black arrows below the line. (B) Pileup of the predicted amino acid sequence of the GI protein (DDBJ/EMBL/GenBank accession No. AJ133786) and rice EST sequence (DDBJ/EMBL/GenBank accession No. AJ133787). Putative transmembrane domains predicted in both of the proteins by membrane topology prediction programs are underlined (black line, GI protein; grey line, rice protein).

Download figure to PowerPoint

The sequence of the IPCR fragment indicated that the T-DNA in gi-11 was inserted 407 bp upstream of the predicted start codon of a novel gene, designated gene 5 (Terryn et al. 1997, Figure 1A). The IPCR fragment was used to screen a genomic library, and five overlapping λ clones from the region were identified. Southern hybridization analysis of wild-type and gi-11 DNA indicated that the λ clones (e.g. λ3, Figure 1A) spanned the site of the T-DNA insertion. Comparison of the restriction maps of wild-type and gi-11 genomic DNA in the region indicated that a deletion of 3–4 kb of plant DNA had occurred upon insertion of the T-DNA (data not shown). The deletion in gi-11 removed the 5′ half of gene 5 and upstream sequence, but did not affect any of the other predicted genes in the region. This strongly suggested that gene 5 corresponded to GI. Our final confirmation of the sequence of the GI gene by identification of the mutations in six classical gi mutant alleles is described in detail below.

Isolation and analysis of the GI cDNA sequence

The GI gene was predicted to encode 13 exons which are spliced together to form an ∼3.5 kb open reading frame (ORF) (Terryn et al., 1997). However, no Arabidopsis expressed sequence tags (ESTs) or cDNA clones corresponding to GI were identified in DDBJ/EMBL/GenBank database searches. Full-length GI cDNA clones were then isolated by RT–PCR and by screening a long insert Arabidopsis cDNA library (3–6 kb, Kieber et al., 1993). The DNA sequence of these cDNA clones indicated that the GI cDNA has a coding region of 3522 bp with a 5′-untranslated region of 318 and 217 bp of 3′-untranslated region. The structure of the gene 5/GI coding region predicted by Terryn et al. (1997) was correct, except that exon 6 was 18 bases longer in the GI cDNA. This had the effect of removing one amino acid and adding seven new amino acids to the predicted protein sequence.

The GI cDNA is predicted to encode a 1173 amino acid protein of 127 kDa (Figure 1B). Database searches using BLAST indicated that the GI protein has no significant homology to proteins of known function, but was similar to the predicted amino acid sequences of rice ESTs (Terryn et al., 1997). These rice ESTs all corresponded to a single gene. The longest EST clone (DDBJ/EMBL/GenBank accession No. C73052) contained an incomplete cDNA that was entirely sequenced (DDBJ/EMBL/GenBank accession No. AJ133787). The predicted protein product of this gene was 976 amino acids long and was 71% identical and 81% similar to the Arabidopsis GI protein (Figure 1B). Web-based membrane topology prediction programs (TopPred1, Von Heijne, 1992; PSORT and TMpred, Nakai and Kanehisa, 1992) predicted that the Arabidopsis GI protein contains up to 11 transmembrane domains, and that the rice protein contains up to six. Four of the most likely predicted domains were also conserved in the rice protein, suggesting that they might be functionally important, and a fifth, highly probable transmembrane domain in the Arabidopsis protein was not included in the region present within the incomplete rice sequence. The sequence analysis of the rice and Arabidopsis genes therefore predicts at least five transmembrane domains, all positioned within the N-terminal half of the protein (in the first 660 amino acids), indicating that this part of the protein may be inserted into a membrane. The PSORT program predicted that both the rice and Arabidopsis proteins are likely to be located in the plasma membrane.

Effect of gi mutations on the predicted GI protein and on flowering time

To confirm that gene 5 was GI, the corresponding gene from six classical gi mutant alleles (gi-1 to gi-6) was sequenced. The region from bp 14 795 to bp 21 169 (Terryn et al., 1997) containing the gene was amplified by PCR from genomic DNA of the mutant alleles, in three overlapping fragments, which were sequenced directly. Alterations in the gene 5/GI region were identified in all six alleles. These were all predicted to alter the GI coding region (Figure 2A) providing confirmation that gene 5 was GI. Four of the mutations (gi-2, gi-6, gi-3 and gi-1) are predicted to introduce premature stop codons into the GI sequence. The most severe of these is the gi-2 allele which is predicted to encode a 144 amino acid protein rather than the 1173 amino acid GI protein. The remaining two mutations (gi-4 and gi-5) are predicted to alter the C-terminus of the GI protein (Figure 2A). gi-4 has a mutation in the 3′ splice acceptor site of intron 12 of the GI gene and gi-5 has a point mutation and a single base deletion in exon 13 of the GI gene. This causes a frameshift which is predicted to both change the last eight amino acids and add 27 amino acids to the C-terminus of the GI protein (Figure 2A).

image

Figure 2. The effect of gi mutations on the GI protein, GI transcript levels and flowering time. (A) Effect of the gi mutations on the GI protein. The position of the mutations in gi-1 to gi-6 on the transcribed region of the GI gene (top). The positions of the gi mutations correspond to the genomic sequence of the GI gene with position 1 at the A of the translation start codon. Exons in the transcribed region are in black and introns are in white. Sequence deletions are indicated by a triangle. The size and structure of the predicted gi mutant proteins are compared below. In gi-4, a mutation in the 3′ splice acceptor site of intron 12 is expected to cause aberrant splicing of the intron with unknown effects on the C-terminal end of the protein (?). In gi-5, the last eight amino acids of the predicted GI protein are altered and 27 amino acids are added to the C-terminus of the GI protein (light grey bar). (B) Northern hybridization analysis of GI expression. Plants were grown in LD conditions until the six leaf stage. Total RNA (10 μg) was extracted from aerial parts harvested 8 h after dawn (ZT 8) and analysed by Northern hybridization using a GI cDNA probe. Results are presented as a proportion of the highest value after normalization with respect to 25/26S rRNA levels. (C) Flowering time of wild-type and gi mutants under long day (LD) and short day (SD) conditions. The genotype tested is shown along the horizontal axis and the leaf number at flowering is plotted on the vertical axis. Data are presented as mean ± SE with n of 9–19 plants except for Ws in LD (n = 2) and gi-3 in SD (n = 2).

Download figure to PowerPoint

To determine if the expression of GI was altered in the gi mutants, tissue from mutant and wild-type plants grown in LD was harvested and GI transcript levels were analysed by Northern hybridization (Figure 2B). The samples were harvested 8 h after dawn (onset of illumination) which is close to peak GI expression in wild-type plants (see below). Three gi mutants (gi-1 to gi-3) had lower levels of GI expression than wild-type plants, while three of the mutants (gi-4 to gi-6) had GI expression levels similar to wild-type. No GI transcript was detected in the T-DNA allele, gi-11, which is consistent with deletion of the 5′ half of the GI gene and upstream region in this mutant.

To determine the functional importance of different regions of the GI protein, we analysed the effect of the six classical gi mutant alleles (gi-1 to gi-6) on Arabidopsis flowering time. Two T-DNA mutant alleles (gi-11 and gi-12) were also included in the experiment. The gi-11 allele is highly likely to be a null mutation, because the 5′ half of the GI gene and part of the upstream region is deleted in gi-11 and no GI transcript was detected by Northern hybridization analysis (Figure 2B). The gi-12 allele is also likely to be a strong allele as it contains a T-DNA insertion in the GI coding region (M.Aukerman and R.Amasino, personal communication). The mutants and corresponding wild-type ecotypes were grown in LD and SD conditions and scored for the numbers of leaves visible at flowering (Figure 2C).

The differences in flowering time between all six classical alleles and the T-DNA mutants were relatively slight in LD (Figure 2C). Interestingly, while the gi-4 and gi-5 mutations are likely to affect only the most extreme 3′ end of the GI gene, they resulted in long delays to flowering compared with Ler wild-type (Figure 2C). One explanation for the delayed flowering seen in gi-4 and gi-5 is that these mutations might have caused decreased stability of the GI transcript. However, we saw no difference in GI transcript levels in LD in gi-4 and gi-5 mutants compared with Ler, which suggests that the C-terminus of the GI protein is functionally important in flowering. The null T-DNA allele gi-11 did not abolish flowering, indicating that even though GI promotes flowering, it is not essential for it to occur.

The gi mutations in the Ler and Ws ecotypes caused late flowering with greatly reduced daylength sensitivity compared with wild-type plants, as previously reported for gi mutants (Redei, 1962; Koornneef et al., 1991; Araki and Komeda, 1993). However, the severe gi-2 and gi-12 mutants in the Col ecotype responded quite strongly to daylength because they flowered much later in SD than in LD (Figure 2B). This is inconsistent with a previous report in which gi-2 was found to flower more rapidly in SD than in LD (Araki and Komeda, 1993). In addition, in our experiment, the gi mutant plants all showed some response to SD because they flowered slightly later than wild-type in SD. This result is consistent with the main function of GI being to promote flowering in LD, but indicates that GI also has some role in promoting flowering in our SD conditions.

GI transcript levels are regulated by the circadian clock

The expression of some other flowering time genes such as LHY and CCA1 is regulated by the circadian clock and appears rhythmically through light/dark cycles and in entrained plants shifted to continuous light (Schaffer et al., 1998; Wang and Tobin, 1998). To determine if GI transcript levels fluctuated within a cycle, tissue of wild-type ecotype Ler was collected every 2–4 h over a 24 h period in LD and analysed by Northern hybridization (Figure 3A). Sampling time was expressed in hours as zeitgeber time (ZT, Zerr et al., 1990) which is the number of hours after dawn (the onset of illumination). GI transcript levels cycled, with the highest level 10 h into the light (ZT 10) and the lowest level at dawn (ZT 0, Figure 3A). A second experiment gave similar results.

image

Figure 3. Northern hybridization analysis of GI expression in different light regimes. Plants were grown until the six leaf stage. Total RNA (10 μg), was extracted from aerial parts of plants harvested at the times shown and analysed by Northern hybridization using a GI cDNA probe. Results are presented as a proportion of the highest value after normalization with respect to 25/26S rRNA levels. ZT 0 is at lights on. Horizontal bars under each graph represent the light (white) and dark (black) conditions provided. Hatched bars represent subjective night experienced in continuous light (LL) and continuous dark (DD) conditions. (A) Time course of GI expression in plants grown in LD (top), LL (middle) and DD (bottom). For the LL experiments, plants were entrained in LD and shifted to LL 24 h before tissue harvesting was initiated at ZT 0. (B) Effect of SD and the transition to darkness on GI expression. Plants were grown in SD and LD (top). In the bottom panel, plants were grown in LD and at ZT 10, half of the LD plants were shifted to darkness (indicated by the arrow).

Download figure to PowerPoint

To determine if the rhythmic cycling of GI transcript levels was under the control of the circadian clock, plants entrained in LD were transferred to either continuous light (LL) or continuous dark (DD) and then assayed for GI expression every 4–8 h over a 48 h period. GI transcript levels continued to cycle in a similar phase under LL and DD, indicating that they were controlled by the circadian clock (Figure 3A). However, in DD, peak levels of GI expression decreased, while trough levels increased compared with LL. This caused an overall reduction in the amplitude of GI expression.

A preliminary experiment carried out over 24 h indicated that GI transcript levels cycled in SD, but peak levels in SD were earlier, higher and of a shorter duration (not detected at ZT 12) compared with LD. To compare GI expression in SD and LD in more detail, tissue was collected every 2 h over a 14 h period (ZT 0–ZT 14) in both conditions and analysed by Northern hybridization (Figure 3B). In SD, GI expression peaked ∼8 h after dawn and reached trough levels ∼3 h later, which was ∼1 h after the transition to darkness (Figure 3B). In LD, peak levels of GI expression were slightly lower and occurred 2 h later than in SD (∼10 h after dawn) and reached trough levels later (∼6 h later) before the transition to darkness (Figure 3A and B).

The rapid reduction in GI transcript levels in SD coincided with the onset of darkness at ZT 10. To test if GI transcript levels might be down-regulated directly by the transition to darkness, plants entrained in LD were prematurely transferred to the dark at ZT 10 and then assayed for GI expression. One hour after the transfer to darkness (ZT 11), GI transcript levels were approximately half the level of control LD plants (Figure 3B). A more extreme down-regulation of GI transcript levels by the transition to darkness was obtained in a second experiment. This indicates that the premature transition to darkness does reduce GI transcript levels in LD-entrained plants. However, in the reverse experiment, SD-grown plants were shifted into LD, but no immediate change in the pattern of GI expression between these plants and control SD plants was detected (data not shown). This suggests that the rapid reduction of GI transcript levels in SD-entrained plants is not controlled directly by the transition to darkness, because GI transcript levels did not increase immediately when SD plants were exposed to LD.

Early flowering Arabidopsis elf3 mutants misexpress GI

Plants carrying mutations in the ELF3 gene show early daylength-insensitive flowering. To test whether the elf3 mutation altered GI transcript levels, elf3 and wild-type ecotype Col plants were grown in LD and SD and assayed for GI expression every 4 h over a 24 h period (Figure 4A and B). In elf3 mutants, unlike wild-type plants, GI transcript was detected at all time points analysed including the beginning of the light period (ZT 0) and after the transition to darkness in SD (ZT 10). GI transcript levels were also higher in elf3 mutants than in wild-type at all time points. GI transcript levels cycled in elf3 mutants in LD and SD, but with much reduced amplitude. GI expression was also assayed in LD in 11 late-flowering mutants (fca-1, fpa-1, fve-1, fy-1, co-2, fe-1, ft-1, fd-1, fha-1 and fwa-1, Koornneef et al., 1991; ld-3, Lee et al., 1994) and corresponding wild-type ecotypes, Ler, Col and Ws. No differences in GI expression between these mutants and wild-type plants were detected at the two time points analysed (ZT 0 and ZT 8, data not shown).

image

Figure 4. Expression of GI in elf3 mutants in LD, SD and LL. Wild-type and elf3 mutant plants were grown in LD or SD conditions until the six leaf stage. Total RNA (10 μg) was extracted from aerial parts of plants at the times shown and analysed by Northern hybridization using a GI cDNA probe. Results are presented as a proportion of the highest value after normalization with respect to 25/26S rRNA levels. Horizontal bars under each graph represent the light (white) and dark (black) conditions provided. Hatched bars represent subjective night experienced in continuous light (LL) conditions. ZT 0 h is at lights on. (A) Time course of GI expression in LD. (B) Time course of GI expression in SD. (C) Time course of GI expression in LL. For the LL experiments, plants were entrained in LD and shifted to LL 24 h before tissue harvesting was initiated at ZT 0.

Download figure to PowerPoint

The elf3 mutant has no detectable circadian rhythms in continuous light conditions, but the circadian clock does function in elf3 mutants in darkness (Hicks et al., 1996). This result suggests that ELF3 may be responsible for regulating light input signals to the clock. The elf3 mutant and wild-type plants were grown in LD, transferred to LL for 24 h and then assayed for GI expression every 4–8 h over a 48 h period. The rhythmic pattern of GI expression was disrupted in elf3 in LL (Figure 4C). GI transcript was detected at all the time points analysed and was present at higher levels in elf3 mutants than in wild-type (Figure 4C). GI transcript levels appeared to fluctuate in elf3 mutant plants, but further experiments are needed to test whether this is really a true cyclical pattern with greatly reduced amplitude. This result indicated that the circadian rhythm of GI expression in the light is regulated via ELF3.

Effect of CCA1-OX and LHY on GI expression

To investigate further how the circadian clock controls GI expression, we analysed the effect of two circadian clock-controlled genes that affect flowering time on GI expression. CCA1-OX transgenic plants were grown in LD, transferred to LL for 24 h and then assayed for GI expression every 4 h over a 24 h period. In CCA1-OX plants in LL, the rhythmic pattern of GI expression was disrupted as GI transcript was detected at all the time points analysed (Figure 5A). This result indicated that the circadian rhythm of GI expression in LL is disrupted by constant expression of CCA1.

image

Figure 5. Effect of the CCA1-OX transgene on GI expression and the lhy mutation on GI and CCA1 expression. (A) Time course of GI expression in CCA1-OX. Wild-type and CCA1-OX plants were grown in LD conditions until the six leaf stage and shifted to LL 24 h before tissue harvesting was initiated at ZT 0. Hatched bars represent subjective night. Total RNA (10 μg) was extracted from aerial parts at the times shown. GI transcript levels were analysed by Northern hybridization using a GI cDNA probe. Results are presented as a proportion of the highest value after normalization with respect to 25/26S rRNA levels. (B) Time course of GI expression in lhy. (C) Time course of CCA1 expression in lhy. Wild-type and lhy plants were grown in sterile conditions on GM medium in LD (16 h light, 8 h dark). Total RNA (10 μg) was extracted from whole 8-day-old plants at the times shown. GI and CCA1 transcript levels were analysed by Northern hybridization. Results are presented as a proportion of the highest value after normalization with respect to ubiquitin 10 mRNA levels.

Download figure to PowerPoint

Overexpression of the LHY gene also disrupted circadian clock-controlled expression of several Arabidopsis genes, and of leaf movements (Schaffer et al., 1998). GI gene expression in LD cycles is also affected in these lhy mutant plants, and does not show its characteristic peak in expression at ∼ZT 10 (Figure 5B).

The LHY and CCA1 genes are closely related in sequence, and overexpression of CCA1 was shown previously to repress expression of both the endogenous CCA1 gene and the LHY gene (Wang and Tobin, 1998). In Figure 5C, it is shown that the overexpression of LHY also represses CCA1 expression, and therefore both of these genes will repress their own and each other's expression.

The gi-3 mutation lowers the expression of the LHY and CCA1 genes

The effect of the gi-3 mutation on LHY expression was analysed in LD and SD by Northern hybridization analysis. The gi-3 mutant (Ler ecotype) was selected for this analysis because Ler was the ecotype in which the wild-type pattern of GI expression was analysed. The gi-3 mutation caused a 5- to 6-fold reduction in peak LHY transcript levels in LD compared with wild-type ecotype Ler in both conditions (LD, Figure 6A). A similar result was obtained in gi-3 mutants grown in SD (data not shown). Although the level of LHY expression was reduced by the gi-3 mutation, LHY transcript levels continued to cycle in LD and SD conditions in the same phase as observed previously in wild-type plants (Schaffer et al., 1998). Interestingly, the CCA1 gene was also down-regulated in the gi-3 mutant in LD (Figure 6B) and in SD (data not shown).

image

Figure 6. Expression of LHY and CCA1 in the gi-3 mutant. (A) Time course of LHY expression in gi-3. (B) Time course of CCA1 expression in gi-3. Wild type and gi-3 mutant plants were grown in LD conditions until the six leaf stage. Total RNA (10 μg) was extracted from aerial parts of plants at the times indicated. Horizontal bars under each graph represent the light (white) and dark (black) conditions provided. ZT 0 h is at lights on. Northern hybridization results are presented as a proportion of the highest value after normalization with respect to 25/26S rRNA levels.

Download figure to PowerPoint

GI is expressed throughout development

GI expression during plant development was analysed by Northern hybridization. Wild-type ecotype Ler was grown in LD and samples were harvested 8 h into the light period when GI was known to be expressed. GI transcript was detected at all the stages of plant development tested, from seedlings at the two leaf stage to mature plants with developed siliques (Figure 7). GI transcript levels were analysed in individual organs and tissues from mature plants. GI was detected in all of the tissues tested, with the highest level of GI expression in inflorescence apices, young flowers and young siliques, and the lowest level in mature siliques and roots.

image

Figure 7. Northern hybridization analysis of GI expression throughout development. Total RNA (10 μg) was extracted 8 h after dawn (ZT 8) from the aerial parts of LD grown plants at the stages shown (from two leaf stage to mature plant) and from specific organs of mature plants as indicated. Bolted plants had a primary inflorescence of ∼4 cm. Mature plants had siliques which were fully expanded but still green. Apices were inflorescence apices with floral buds and flowers pre-anthesis. Flowers were at anthesis. GI transcript levels were analysed by Northern hybridization using a GI cDNA probe. Results are presented as a proportion of the highest value after normalization with respect to 25/26S rRNA levels. C.L., cauline leaves.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References

GI sequence and spatial pattern of expression

The protein encoded by the GI flowering time gene does not show homology with genes of known function, but is predicted to be membrane localized. Membrane topology prediction programs found up to 11 membrane-spanning domains in the N-terminal half of the GI protein, and the protein most closely related to GI, a rice protein encoded by an EST, is predicted to share at least four of these domains with the GI protein. The possible conservation of these membrane-spanning domains through evolution suggests that they are likely to be important for GI function. The C-terminal end of the protein is not predicted to be localized within a membrane, but flowering time experiments with gi mutant alleles gi-4 and gi-5 suggest that it is required for GI function.

Previous experiments with the conditional gi-2 allele suggested that GI acted 4–7 days after germination in LD (two-leaf to four-leaf stage plants) to promote flowering (Araki and Komeda, 1993). Detection of GI transcript in very young plants is consistent with a role early in plant development to promote flowering. Classical grafting studies of photoperiodic flowering in plants suggested that daylength was perceived in the leaves and a graft-transmissible signal moved to the shoot apex to activate or repress flowering. However, the tissues in which GI and other Arabidopsis flowering time genes are needed to promote flowering have not been defined. For example, GI transcript is detected throughout development and in all parts of mature plants, and the long day floral-promotive gene CO is detected in both leaves and shoot apices (Putterill et al., 1995; Simon et al., 1996).

GI expression in wild-type plants and photoperiodic flowering

Previous physiological experiments implicated the circadian clock in photoperiodic time measurement, and genetic experiments in Arabidopsis have also identified mutants that disrupt both circadian clock-controlled responses and photoperiodic responses (reviewed by Koornneef et al., 1998b). The GI gene has been implicated in the flowering photoperiodic response, because flowering time of gi mutants is delayed most severely under long days. The analysis of GI expression through day/night cycles, and in entrained plants shifted to LL or DD, further suggested that GI may be involved in controlling photoperiodic responses. First, GI expression was regulated by the circadian clock and its expression peaked during the light period in the light/dark cycles tested. Second, the pattern of GI expression differed under long and short days. GI expression responded to long days in three ways: the peak of expression was slightly lower and is shifted later, and the reduction in expression from peak levels was slower, thereby increasing the length of time that GI is expressed above trough levels. This difference may be due in part to entrainment of the oscillator to different daylengths (Millar and Kay, 1996).

GI is unique amongst genes acting in the long-day pathway because it is the only one known to control photoperiodic flowering, to be circadian regulated and to have differing patterns of expression in LD and SD. Whether these changes in GI expression between LD and SD have a direct effect in controlling flowering time is so far unknown, but could be tested by manipulating GI expression in transgenic plants.

Interactions between GI and ELF3

The elf3 mutation was shown previously to disrupt several circadian rhythms in plants in LL conditions. The ELF3 gene is predicted to provide light input signals into the clock, because elf3 mutations disrupt circadian rhythms in LL, but not in DD. As expected, circadian regulation of GI expression was disrupted by the elf3 mutation in LL, and GI expression was observed at all time points at a level above trough levels of wild-type plants. This contrasts with the previous observation that LHY expression fell rapidly to trough levels in an elf3 mutant in LL (Schaffer et al., 1998). Therefore, although both the LHY and GI genes are arrhythmic in an elf3 background in LL, the level of expression of these genes responds very differently to a loss of ELF3 function. This may be a consequence of the different phases in which these genes cycle, with LHY showing a peak in expression at dawn, but GI peaking in expression later during the light period.

In the elf3 mutant, the waveform of expression of another circadian-regulated gene LHCB2 (previously designated CAB2) was observed to be altered compared with wild-type in LD, particularly as the light period lengthened (Hicks et al., 1996). However, under short days, rhythmic LHCB2 expression was similar in the mutant and wild-type (Hicks et al., 1996). The cyclical pattern of GI expression was again most affected by elf3 under LD, but was also affected under our SD conditions. It is possible that the increased level of GI expression detected in an elf3 mutant plays a role in the early flowering of these plants.

Interactions between GI, CCA1 and LHY

CCA1 and LHY are both predicted to encode Myb-related proteins and are thought to function in the circadian clock or in a clock output pathway, because CCA1-OX and LHY-OX disrupt circadian rhythms in LL and DD. CCA1 and LHY are not simply redundant genes with identical functions because a CCA1 knock-out allele reduces the period of a circadian cycle even in the presence of a wild-type LHY gene (Green and Tobin, 1999). As was shown previously for several other circadian-clock controlled genes, CCA1-OX and LHY-OX disrupted circadian expression of GI. However, the gi-3 mutation also had an effect on CCA1 and LHY expression, reducing the abundance of both transcripts. Therefore, it is not possible to propose a straightforward linear arrangement for the action of these genes with GI acting before or after CCA1 or LHY, but rather they affect each other's expression.

It is not clear how LHY and CCA1 levels relate to flowering time. Low levels of LHY and CCA1 in gi-3 are associated with late flowering, but reduced levels of LHY correlate with daylength-insensitive early flowering in elf3 mutants, while high levels of LHY or CCA1 in lhy mutants and CCA1-OX plants, respectively, are associated with late flowering. However, all of these genotypes in which LHY or CCA1 expression is disrupted show altered responses to daylength, suggesting that rhythmic expression of these genes plays an important role in daylength responses.

Since gi-3 affects the expression of CCA1 and LHY, which are closely associated with the circadian clock in Arabidopsis, then gi mutations might also be expected to alter circadian rhythms in Arabidopsis. To test this, circadian regulation of LHCB2 was assayed in F2 populations segregating the T-DNA null mutant allele gi-11 and the LHCB2::LUC reporter gene (D.Somers and S.Kay, personal communication). There was a gradual breakdown of circadian clock control of LHCB2 cycling in approximately one-quarter of the F2 plants in constant white light. However, initially, LHCB2 cycled normally in these plants, suggesting that GI is not absolutely required for rhythmicity (D.Somers and S.Kay, personal communication).

Perspectives

The GI, LHY, CCA1 and ELF3 genes previously were proposed to act in the same genetic flowering time pathway (Koornneef et al., 1998b). The expression analysis presented here strengthens these connections by showing that gi-3 reduces the expression of LHY and CCA1, and that CCA1-OX, lhy and elf3 affect GI expression. However, the complexity of these interactions suggests that this is unlikely to be a straightforward linear pathway because there are examples of genes affecting each other's expression. How direct these effects are is unknown. For example, it seems unlikely that the GI protein, which is probably located in a membrane, directly regulates the transcription of other genes. Clues to the identity of intermediates should be gained by identifying proteins that interact with the GI protein and by identifying genetic suppressors of gi mutations.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References

Plant material and growth conditions

Arabidopsis thaliana L. Heynh wild-types were ecotypes Columbia (Col), Landsberg erecta (Ler, obtained from Lehle Seeds, Round Rock, TX) and Wassilewskija (Ws obtained from the Arabidopsis Biological Resources Centre, OH). The gi mutants used were gi-1 and gi-2 (Col ecotype, obtained from the Arabidopsis Biological Resources Centre, OH), gi-3 to gi-6 (Ler ecotype, obtained from Maarten Koornneef, Wageningen, The Netherlands), gi-11 (T-DNA mutant allele in Ws ecotype; Richardson et al., 1998) and gi-12 (T-DNA mutant allele in Col ecotype, a gift from M.Aukerman and R.Amasino, Wisconsin). The elf3 mutant (Col ecotype) was obtained from the Arabidopsis Stock Centre and has been described previously (Zagotta et al., 1996). The lhy mutant (Ler ecotype) was described previously (Schaffer et al., 1998). Seed of transgenic plants overexpressing the CCA1 gene (CCA1-OX, Col ecotype) was a gift from Elaine Tobin (UCLA, California) and has been described previously (Wang and Tobin, 1998). The late flowering mutants fca-1, fpa-1, fve-1, fy-1, co-2, fe-1, ft-1, fd-1, fha-1, fwa-1 and ld-3 were obtained from the Arabidopsis Biological Resources Centre (OH) and were described previously (Koornneef et al., 1991; Lee et al., 1994). Seeds were placed on moist filter paper at 4°C for 3 days, planted in soil, germinated in growth cabinets and scored for flowering time or harvested for Northern hybridization analysis as described below.

Measurement of flowering time

Flowering time analysis of gi alleles and wild-type was carried out on plants grown in Gallenkamp cabinets either in SD (10 h light/14 h dark) or in LD (10 h light + 6 h day extension/8 h dark) as described in Putterill et al. (1995). Flowering time was measured by counting the number of leaves when floral buds were visible at the centre of the rosette. Data from one experiment are presented as mean ± SE where n = 9–19 plants, except where indicated.

Isolation of plant genomic DNA sequences flanking the T-DNA insertion in gi-11

Previously, we identified a late flowering mutant (now gi-11), carrying a T-DNA insertion at a single site in the genome closely linked to the gi mutation (Richardson et al., 1998). In order to isolate DNA sequences flanking the T-DNA insertion, IPCR was carried out on gi-11 plant genomic DNA. A 1.3 kb fragment was obtained from plant genomic DNA digested with BstYI and, following ligation, amplified by nested PCR using primers gkb8 5′-agctggtacattgccgtag (6581–6560) and gkb9 5′-tttttgcttggactataatacc (6654–6675) followed by gkb7 5′-tagatgaaagactgagtgcgat (6581–6560) and gkb10 5′-ctacaaattgccttttcttatc (6737–6750). The co-ordinates of the PCR primers used correspond to the pGKB5 sequence which is available at http://nasc.nott.ac.uk:8300/vol2ii/bouchez.html. The PCR product was purified with the High Pure™ PCR product purification kit (Boehringer Mannheim) and sequenced, confirming that the fragment contained T-DNA sequences as well as unknown DNA sequence.

Position of the TH1 CAPS marker relative to gi

The TH1 CAPS marker is located on sequenced BAC F19G10 (DDBJ/EMBL/GenBank accession No. ATAF000657) and is defined by the following primers: F 5′-gttacttacaccacggttattc and R 5′-caccaactttgcttcctc. After cleavage with Sau3AI, this gives rise to a single fragment of 530 bp in Col and to two fragments of ∼350 and 180 bp in Ler. Genetic mapping placed this ∼0.36 cM from gi-3. This was determined by analysing the DNA of 699 gi-3 homozygotes derived from the F2 of a cross between gi-3 (Ler) and Col. Only five of the 699 gi-3 homozygotes carried a recombinant chromosome with a cross-over between gi-3 and TH1.

Sequencing of gi mutant alleles

The GI gene was amplified by PCR from genomic DNA of six different gi mutant plants, gi-1 to gi-6. Three overlapping genomic fragments of 2.15, 1.84 and 2.52 kb in size were amplified by three pairs of PCR primers respectively from each gi mutant. These were oli1 5′-ggtaatggcgcataaagg (14 795–14 812) and oli2 (R) 5′-tggttcaagagctggaag (16 924–16 941), oli3 5′-atgagactgctacgctgac (16 835–16 853) and oli4 (R) 5′-cccacttacatctcatcaac (18 673–18 692), and oli5 5′-gttcagacgttcaaagg (18 634–18 651) and oli6 (R) 5′-aactccaatcccaaaacc (21 152–21 169). The primer co-ordinates correspond to the DNA sequence of a cosmid containing the GI gene (DDBJ/EMBL/GenBank accession No. Y12227, Terryn et al., 1997). PCR products were purified with the High Pure™ PCR product purification kit (Boehringer Mannheim) and sequenced on one strand using a series of sequencing primers designed from the GI genomic sequence. The DNA sequence from the gi mutant alleles was compared with the wild-type sequence to identify the position of the mutations. DNA polymorphisms between the wild-type sequence (Col ecotype) and gi-3 to gi-6 (Ler ecotype) were excluded as they occurred in all four of these gi mutants. Approximately 0.46% of bases were polymorphic between Ler and Col. Once mutations were found, a second independently generated, PCR product was sequenced to verify that the DNA sequence alteration was real rather than a PCR-derived error.

Screening of phage and cosmid libraries

GI genomic clones were identified by screening 20 000 plaques of the CD4-8 Arabidopsis genomic λ library (Voytas et al., 1990). Partial GI cDNA clones were identified by screening 500 000 plaques of the CD4-7 Arabidopsis cDNA library (Newman et al., 1994). A full-length GI cDNA clone of 4077 bp was identified by screening 500 000 plaques of the CD4-16 Arabidopsis cDNA library (insert size range 3–6 kb, Kieber et al., 1993). This clone provided the sequence of the 5′-untranslated region of the GI gene.

Isolation of GI cDNA by RT–PCR

A GI cDNA was amplified in three overlapping fragments by RT–PCR. These were ligated together to form a full-length GI cDNA. First strand cDNA synthesis on 1 μg of total Ler RNA was primed using the dT17 adapter primer as described by Frohman et al. (1988). Three overlapping cDNA fragments of 0.9, 1.2 and 1.7 kb in size were amplified by three pairs of PCR primers designed from the predicted GI coding sequence, with some alterations for creation of restriction enzyme sites in the primers. These were oli26 5′-gttaagcttcggttcctggatggc (15 457–15 480, contains the GI translation start codon and HindIII site) and oli2 (R) 5′-tggttcaagagctggaag (16 924–16 941), oli28 5′-tggagagctcaagccgccaaccat (16 529–16 552) and oli30 (R) 5′-ctcttgctacctctagactgtgcttc (18 182–18 207, contains the XbaI site), and oli29 5′-cacagactagtggtagcaagagac (18 186 −18 209 contains the SpeI site) and oli7R 5′-gtgggtgctcgttattgg (20 519–20 536 contains the GI translation stop codon). The primer co-ordinates correspond to the DNA sequence of cosmid containing the GI gene (DDBJ/EMBL/GenBank accession No. Y12227, Terryn et al., 1997). The SacI site (16 533, DDBJ/EMBL/GenBank accession No. Y12227, Terryn et al., 1997) was used to join fragments 1 and 2, and the XbaI (oli30) and SpeI (oli29) sites were used to join fragments 2 and 3. The PCR products were digested with restriction enzymes and ligated into pBluescript SK+ or pJIT60 (Guerineau and Mullineaux, 1993) and sequenced on one strand using universal primers and a series of sequencing primers designed from the GI genomic sequence. A PCR-derived error was discovered in the fragment amplified with oli29 and oli7 and corrected by site-directed mutagenesis with QuikChange™ site-directed mutagenesis kit (Stratagene).

Detection of mRNA by Northern hybridization analysis

Unless otherwise stated, plants were grown in Percival AR-32L cabinets at 22°C providing either continuous light (LL), short days (SD) of 10 h light/14 h dark, or long days (LD) of 18 h light/6 h dark. Light intensity of 150–170 μmol/m2/s was provided by fluorescent tubes. Plants for the continuous dark (DD) experiment were entrained in LD of 18 h light/6 h dark in cabinets with light provided by metal halide lamps and then shifted to continuous darkness. The aerial parts of at least 10 plants, or at least three plants for specific organs, were pooled for RNA extractions (Whitelam et al., 1993). Total RNA (10 μg) was electrophoresed on agarose gels and transferred to Boehringer Mannheim positively charged nylon membrane or Hybond NX (Amersham) as described in Fourney et al. (1988). RNA was bound to the membrane using a UV Stratalinker (Stratagene). The GI probe used in Northern hybridization analysis was a 1817 bp cDNA fragment from the 3′ half of the GI cDNA (2235–4051 on GI cDNA, DDBJ/EMBL/GenBank accession No. AJ133786). This probe is specific for the GI transcript as it does not detect any transcript in gi-11 mutant plants which carry a deletion of the 5′ end of the gene and promoter. The LHY probe used in Northern hybridization analysis was the full-length LHY cDNA (DDBJ/EMBL/GenBank accession No. AJ006404). A 605 bp CCA1 fragment was generated by PCR using primer cca1f 5′-tccttcagagccagatagtc (3782–3801 on the CCA1 gene, DDBJ/EMBL/GenBank accession No. U79156) and primer cca1r 5′-gtgccatcctcttgcctttc (4387–4386 on the CCA1 gene, DDBJ/EMBL/GenBank accession No. U79156). For Northern hybridization analysis of the lhy mutant, a 1600 bp fragment from the 3′ end of the GI cDNA, a 2000 bp fragment of the CCA1 cDNA (derived from plasmid CCA1-24, a gift from Elaine Tobin) and a 110 bp fragment from the 3′-untranslated region of the UBQ10 cDNA (described in Schaffer et al., 1998, a gift from Elaine Tobin) were used as probes. The probes were radiolabelled by priming with random octamers (Gibco-BRL). The radiolabelled DNA was hybridized to the Northern blot membranes in hybridization buffer for 18 h at 65°C and then washed at moderate stringency using two washes of 0.5 or 1× SSC, 0.1% SDS at 65°C. After Northern hybridization, nylon membranes were exposed to a Fujifilm BAS-MP imaging plate at room temperature. The image was visualized using a Fujifilm FLA-2000 phosphorimager running Imagereader version 1.3E software. The expression levels were quantitated using the MacBAS version 2.5 program and background hybridization levels subtracted. Expression levels were normalized against the signal obtained by hybridizing the blot with an asparagus 25/26S rDNA probe. The normalized values were then expressed as a proportion of the highest value obtained and graphed.

The Northern analyses presented are the results of one experiment. The experiments presented in Figure 3 were replicated (except the DD panel). Similar results were obtained from the second experiment.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References

We are very grateful to the following people who provided us with materials: Milo Aukerman and Rick Amasino for the gift of the gi-12 mutant, Maarten Koornneef for providing us with seed of gi mutant alleles, Elaine Tobin for the gift of the CCA1-OX seed and CCA1 probes, and the Arabidopsis stock centre at Ohio for providing us with numerous seed stocks and Arabidopsis genomic and cDNA libraries. We thank Dave Somers and Steve Kay for sharing data pre-publication. We thank Robert Schaffer and Fran Robson for their comments on the manuscript and Caroline Hollis for confirming the sequence of the gi-4 allele. This work was supported by a stipend to S.F. from Agmardt New Zealand and by a fellowship to H.O. from Ciba Geigy, Japan.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Materials and methods
  7. Acknowledgements
  8. References
  • Araki T and Komeda Y (1993) Analysis of the role of the late-flowering locus, GI, in the flowering of Arabidopsis thaliana. Plant J, 3, 231239.
  • Dunlap JC (1999) Molecular bases for circadian clocks. Cell, 96, 271290.
  • Eimert K, Wang S-M, Lue W-L and Chen J (1995) Monogenic recessive mutations causing both late floral initiation and excess starch accumulation in Arabidopsis. Plant Cell, 7, 17031712.
  • Fourney RM, Miyakoshi J, Day RS,III and Paterson MC (1988) Northern blotting: efficient RNA staining and transfer. Focus, 10, 57.
  • Frohman MA, Dush MK and Martin GR (1988) Rapid production of full-length cDNAs from rare transcripts: amplification using a single gene-specific oligonucleotide primer. Proc Natl Acad Sci USA, 85, 89989002.
  • Green RM and Tobin EM (1999) Loss of the circadian clock-associated protein 1 in Arabidopsis results in altered clock-regulated gene expression. Proc Natl Acad Sci USA, 96, 41764179.
  • Guerineau F and Mullineaux P (1993) Plant transformation and expression vectors. In Croy,R.R.D. (ed.), Plant Molecular Biology Labfax. BIOS Scientific Publishers Ltd, Oxford, UK, pp. 121147.
  • Guo H, Yang H, Mockler TC and Lin C (1998) Regulation of flowering time by Arabidopsis photoreceptors. Science, 279, 13601363.
  • Hicks KA, Millar AJ, Carré IA, Somers DE, Straume M, Meeks-Wagner DR and Kay SA (1996) Conditional circadian dysfunction of the Arabidopsis early-flowering 3 mutant. Science, 274, 790792.
  • Hollis CA (1999) Characterising GIGANTEA; an Arabidopsis gene that regulates flowering-time. MSc thesis, University of Auckland.
  • Kieber JJ, Rothenberg M, Roman G, Feldmann KA and Ecker JR (1993) CTR1, a negative regulator of the ethylene response pathway in Arabidopsis, encodes a member of the Raf family of protein kinases. Cell, 72, 427441.
  • Koornneef M, Alonso-Blanco C, Blankestijn-de Vries H, Hanhart CJ and Peeters AJM (1998a) Genetic interactions among late-flowering mutants of Arabidopsis. Genetics, 148, 885892.
  • Koornneef M, Alonso-Blanco C, Peeters AJM and Soppe W (1998b) Genetic control of flowering time in Arabidopsis. Annu Rev Plant Physiol Plant Mol Biol, 49, 345370.
  • Koornneef M, Hanhart CJ and van der Veen JH (1991) A genetic and physiological analysis of late-flowering mutants in Arabidopsis thaliana. Mol Gen Genet, 229, 5766.
  • Kurepa J, Smalle J, Van Montagu M and Inez D (1998) Oxidative stress tolerance and longevity in Arabidopsis: the late-flowering mutant gigantea is tolerant to paraquat. Plant J, 14, 759764.
  • Lee I, Aukerman MJ, Gore SL, Lohman KN, Micheals SD, Weaver LM, John MC, Feldmann KA and Amasino RM (1994) Isolation of LUMINIDEPENDENS: a gene involved in the control of flowering time in Arabidopsis. Plant Cell, 6, 7583.
  • Levy YY and Dean C (1998) The transition to flowering. Plant Cell, 10, 19731989.
  • Lin C, Ahmad M, Chan AR and Cashmore AR (1996) CRY2: a second member of the Arabidopsis cryptochrome family. Plant Physiol, 110, 10471048.
  • Millar AJ and Kay SA (1996) Integration of circadian and phototransduction pathways in the network controlling CAB gene transcription in Arabidopsis. Proc Natl Acad Sci USA, 93, 1549115496.
  • Nakai K and Kanehisa M (1992) A knowledge base for predicting protein localization sites in eukaryotic cells. Genomics, 14, 897911.
  • Newman T et al. (1994) Genes galore: a summary of methods for accessing results from large-scale partial sequencing of anonymous Arabidopsis cDNA clones. Plant Physiol, 106, 12411255.
  • Nilsson O, Lee I, Blázquez MA and Weigel D (1998) Flowering-time genes modulate the response to LEAFY activity. Genetics, 150, 403410.
  • Piñeiro M and Coupland G (1998) The control of flowering time and floral identity in Arabidopsis. Plant Physiol, 117, 18.
  • Putterill J, Robson F, Lee K, Simon R and Coupland G (1995) The CONSTANS gene of Arabidopsis promotes flowering and encodes a protein showing similarities to zinc finger transcription factors. Cell, 80, 847857.
  • Rédei GP (1962) Supervital mutants of Arabidopsis. Genetics, 47, 443460.
  • Richardson K, Fowler S, Pullen C, Skelton C, Morris B and Putterill J (1998) T-DNA tagging of a flowering-time gene and improved gene transfer by in planta transformation of Arabidopsis. Aust J Plant Physiol, 25, 125130.
  • Saunders DS (1977) An Introduction to Biological Rhythms. Blackie and Son Ltd, Glasgow and London, UK.
  • Schaffer R, Ramsay N, Samach A, Corden S, Putterill J, Carré I and Coupland G (1998) The late elongated hypocotyl mutation of Arabidopsis disrupts circadian rhythms and the photoperiodic control of flowering. Cell, 93, 12191229.
  • Simon R, Igeño MI and Coupland G (1996) Activation of floral meristem identity genes in Arabidopsis. Nature, 384, 5962.
  • Somers DE, Webb AAR, Pearson M and Kay SA (1998) The short-period mutant, toc1-1, alters circadian clock regulation of multiple outputs throughout development in Arabidopsis thaliana. Development, 125, 485494.
  • Terryn N et al. (1997) Sequence analysis of a 24-kb contiguous genomic region at the Arabidopsis thaliana PFL locus on chromosome 1. FEBS Lett, 416, 156160.
  • Von Heijne G (1992) Membrane protein structure prediction, hydrophobicity analysis and the positive-inside rule. J Mol Biol, 225, 487494.
  • Voytas DF, Konieczny A, Cummings MP and Ausubel FM (1990) The structure, distribution and evolution of the Ta1 retrotransposable element family of Arabidopsis thaliana. Genetics, 126, 713722.
  • Wang ZY and Tobin EM (1998) Constitutive expression of the CIRCADIAN CLOCK ASSOCIATED (CCA1) gene disrupts circadian rhythms and suppresses its own expression. Cell, 93, 120.
  • Weigel D, Alvarez J, Smyth DR, Yanofsky MF and Meyerowitz EM (1992) LEAFY controls floral meristem identity in Arabidopsis. Cell, 69, 843857.
  • Whitelam GC, Johnson E, Peng J, Carol P, Anderson ML, Cowl JS and Harberd NP (1993) Phytochrome A null mutants of Arabidopsis display a wild-type phenotype in white light. Plant Cell, 5, 757768.
  • Zagotta MT, Hicks KA, Jacobs CI, Young JC, Hangarter RP and Meeks-Wagner DR (1996) The Arabidopsis ELF3 gene regulates vegetative photomorphogenesis and the photoperiodic induction of flowering. Plant J, 10, 691702.
  • Zerr DM, Hall JC, Rosbach M and Siwicki KK (1990) Circadian fluctuations of period protein immunoreactivity in the CNS and the visual system of Drosophila. J Neurosci, 10, 27492762.