Regulation of Flowering Time by Arabidopsis Photoreceptors

See allHide authors and affiliations

Science  27 Feb 1998:
Vol. 279, Issue 5355, pp. 1360-1363
DOI: 10.1126/science.279.5355.1360


The shift in plants from vegetative growth to floral development is regulated by red–far-red light receptors (phytochromes) and blue–ultraviolet A light receptors (cryptochromes). A mutation in the Arabidopsis thaliana CRY2 gene encoding a blue-light receptor apoprotein (CRY2) is allelic to the late-flowering mutant, fha. Flowering in cry2/fha mutant plants is only incompletely responsive to photoperiod. Cryptochrome 2 (cry2) is a positive regulator of the flowering-time gene CO, the expression of which is regulated by photoperiod. Analysis of flowering in cry2 and phyB mutants in response to different wavelengths of light indicated that flowering is regulated by the antagonistic actions of phyB and cry2.

The blu e–ultraviolet A (UV-A) light receptors, cryptochromes, and red–far-red light receptors, phytochromes, mediate light-regulated plant growth and development from seed germination to flower initiation. Phytochrome A (phyA), phytochrome B (phyB), and cryptochrome 1 (cry1) function in both early photomorphogenesis (1-5) and floral induction (6-9). We report that in Arabidopsis thaliana, the blue-light receptor cry2 (10) plays a major role in floral induction.

We isolated two Arabidopsis mutant alleles that accumulate no CRY2 protein (Fig. 1A) (11). The mutations, cry2-1 andcry2-2, resulted from partial and complete deletion, respectively, of the CRY2 gene (Fig. 1C). Young seedlings of the cry2 mutants are impaired in blue-light–dependent hypocotyl inhibition and cotyledon opening (10), andcry2 mutant plants flowered later than normal (Fig. 1, E and F). The late-flowering phenotype of cry2 mutants is recessive. Plants heterozygous for thecry2 locus (cry2/CRY2) flowered at about the same time as the wild-type plants, although lower levels of CRY2 protein was detected in these plants (Fig. 1F). Because of prolonged vegetative growth, the number of rosette leaves of cry2 mutant plants was roughly twice normal at flowering (Fig. 1F), as with the late-flowering mutants (12).

Figure 1

Isolation ofcry2 mutants and characterization of cry2/fha. (A) cry2-1 and cry2-2 mutants accumulate no CRY2 protein. Samples from hy4(hy4-304, a cry1 mutant)cry2-1, cry2-2 mutant (11), and Columbia wild-type (wt) plants were prepared and analyzed, using immunoblot as described (10). The blot was probed with antibody to CRY2 (upper), and was then stripped to remove the anti-CRY2 (10) and reprobed with anti-CRY1 (lower). (B)fha-1 and fha-2 mutants accumulate little CRY2 protein. Samples of cry2-1, its Columbia parent (Col), fha-1, fha-2, and their wild-type parent (Ler) were analyzed as in (A). (C) A Southern blot of the genomic DNA of the wild-type (Col) and the cry2-1and cry2-2 mutant plants. Genomic DNA was isolated (28) using the cetyltrimethylammonium bromide method, digested with the restriction enzymes Eco RV and Eco RI, separated on a 1% agarose gel (10 μg per lane), transferred to a Nylon membrane, and hybridized with 32P-labeled CRY2 cDNA. (D) The diagram shows mutations in the CRY2gene (GenBank accession: U43397) of different cry2/fhamutant alleles (not to the exact scales). The CRY2 gene is boxed, genomic sequences surrounding CRY2 are represented by solid lines (5′) and thick dashed lines (3′), and the thin dashed lines represent deletions. The CRY2 sequences (29) at and flanking mutations found in fha-1and fha-2 are shown in comparison with the correspondingCRY2 sequence of the wild type (Ler). (E) Thirty-four-day-old plants of the cry2 mutant (cry2-1) and wild-type Columbia (wt) grown under continuous white light. (F) An immunoblot showing the absence or presence of CRY2 protein in the homozygous (cry2/cry2), heterozygous (cry2/CRY2) lines, and wild-type plants (CRY2/CRY2). The flowering time and number of rosette leaves at the emerging of the first flowering buds for each were the averages of a population with more than 20 plants.

Arabidopsis is a facultative long-day plant for which flower initiation is accelerated in long-day (LD) but delayed in short-day (SD) photoperiods (13, 14). We examined the effect of cry2 on photoperiod-regulated floral induction (Fig. 2) (15). A similar amount of total irradiation was provided for plants grown under LD and SD to minimize the effect of different day lengths on photosynthesis. The difference between the cry2 mutant and the wild-type plants in the flowering time (and the number of rosette leaves) was the greatest for plants grown in LD; this difference diminished when the plants were treated with fewer LD periods (Fig. 2, A and B). Under uninterrupted SD photoperiods, cry2 mutant plants flowered slightly earlier than the wild-type in SD (Fig. 2, A and B). Therefore, the mutation in the CRY2 gene results in a partial loss of photoperiodic regulation of flowering time. The transgenic plants overexpressing CRY2 (10) flowered slightly earlier than the wild type in SD, but at the same time as the wild type in LD (Fig. 2, C and D).

Figure 2

The photoperiodic response of the cry2 mutant and CRY2-overexpressing plants. Seeds of cry2 and the corresponding wild type (Col) (A and B) or the CRY2- overexpressing line H2-9 (CRY2+) and the corresponding wild type (ws) (C andD) were sown on compound soil with a similar density (24 seeds per 3-inch by 5-inch pot), kept at 4°C for 4 days, and grown in LD [18 hours of light (∼100 μmol s−1m−2), 6 hours of darkness]. One pot of each line was transferred to SD [9 hours light (∼200 μmol s−1m−2), 15 hours darkness] at the time shown on the abscissa or was grown in continuous LD (Cont. LD). “Days to flower” are measured as the days between the date plants were placed under light to the date the first flower bud appeared (A and C). The number of rosette leaves were scored at the day the first flower bud appeared (B and D). The flowering time and the number of rosette leaves shown were the averages of total plants in each pot; the standard deviations are shown.

We surveyed other Arabidopsis late-flowering mutant lines (gi, co, fha, fca, fd, fe, ft, fwa, fve, andld) (12, 16) for the expression of CRY2. One of the late-flowering mutants with reduced sensitivity to photoperiod,fha (12, 14), contained mutations in theCRY2 gene. Little CRY2 protein was detected infha mutants (Fig. 1B), although CRY1 expression was unaffected (Fig. 1B). All other late-flowering mutants examined showed normal CRY2 expression. cry2 and fha both flower later, and cry2-1 failed to complement fha-1(17). CRY2 and FHA mapped to the same region of chromosome 1 (17). DNA sequence analysis revealed a premature stop codon of CRY2 in fha-1 (Fig. 1D). The mutation of fha-2 converts Gly254 of CRY2 to Arg (Fig. 1D). It is unclear why such a missense mutation would result in loss of CRY2 protein in the fha-2 plants (Fig. 1B), but Gly254 is conserved in both photolyases and cryptochromes (10, 18) and is separated by five residues from another conserved cluster (Thr-Ser-Xaa-Leu-Ser, with Xaa indicating any amino acid). In Escherichia coli photolyase, this motif forms polar interactions between the apoprotein backbone and the phosphate oxygen of the FAD chromophore (19). The change of the aliphatic residue Gly to a basic amino acid Arg at position 254 may have a detrimental effect on the binding of flavin chromophore and thus result in the instability of the mutant protein.

Other Arabidopsis late-flowering mutants,co and gi, also flower late in LD but not in SD (12, 20, 21). The CONSTANS (CO) gene encodes a GATA-1–type transcriptional regulator required for the accelerated flowering of Arabidopsis in LD (20). The expression of CO itself is regulated by day length; the abundance of CO mRNA is higher in plants grown in LD than in plants grown in SD (20). CO activates the expression of floral meristem identity genes and leads to flower initiation (22). CO has been proposed to be downstream in a signaling pathway from a hypothesized blue light receptor (14). To investigate the relationship of CO and CRY2, we analyzed CO mRNA levels (20) in cry2mutant plants grown under different photoperiods (Fig.3). The expression of the AP2 gene (23), which is not regulated in response to photoperiod (20), was used as the control (Fig. 3). CO mRNA level detected in the cry2 mutant plants grown in LD was at least three times lower than that in the corresponding wild-type plants (Fig. 3). CO message amounts in cry2 mutant grown in SD were only slightly reduced (Fig. 3), which may explain whycry2 mutant plants flowered late only in LD (Fig. 2, A and B). Transgenic plants overexpressing CRY2 had CO mRNA levels significantly higher than the wild type in SD but not in LD (Fig. 3), correlating with the flowering time of the transgenic plants (Fig. 2, C and D). These results indicate that cry2 is a positive regulator ofCO in response to photoperiod. cry2 is apparently not the only photoreceptor regulating CO expression: there was a twofold increase of CO mRNA in the hy1 mutant impaired in the biosynthesis of the phytochrome chromophore (22), and activity of CO is required for the early-flowering phenotype of hy1 and phyB mutants (20). Thus, cryptochrome 2 and phytochromes appear to function antagonistically in the regulation of CO gene expression.

Figure 3

Expression ofCO in the cry2 mutant or CRY2-overexpressing transgenic plants. (A) The mRNA levels of CO in different samples were detected by reverse transcription–polymerase chain reaction using the primers specific for either CO orAP2 as described (20). We used 20 μg of total RNA, isolated from leaf tissues of different lines of plants grown under either LD or SD for 15 days, to synthesize the first strand of cDNA in a 100-μl reaction using a reverse transcription system according to the manufacturer's instructions (Promega); the reaction was diluted 10-fold, and 1 μl was used in a 50-μl polymerase chain reaction (preheat at 94°C for 2 min, then 25 cycles, each of 55°C for 30 s, 68°C for 2 min, and 94°C for 30 s); 5 μl of each sample was fractionated in 1% agarose gel, blotted to a Nylon membrane, and the DNA was hybridized with the 32P-labeled CO or AP2 probe, accordingly. (B) The relative intensity ofCO bands was calculated by normalization of the intensities of the CO2 bands with the intensities of the corresponding AP2 bands; both were quantified from the digitized autoradiography using the NIH Image program (National Institutes of Health, Research Service Branch, National Institute of Mental Health, Bethesda, Maryland).

Blue light (wavelength of ∼400 to 500 nm) and red light (∼600 to 700 nm) promote and inhibit flowering ofArabidopsis, respectively (24, 25), suggesting different functions of phytochromes and cryptochromes in the flowering-time determination. Consistent with previous reports (24, 25), wild-type plants grown under continuous blue light flowered earlier (within 15 days after germination) than plants grown under a similar intensity of red light (more than 30 days after germination) (Fig. 4; red, blue). Considering that blue light promotes flowering and the cry2mutant flowered late, it may be expected that cry2 mutants might flower later than the wild type if grown in continuous blue light. To our surprise, cry2 mutant plants grown under continuous blue (or red) light flowered at about the same time as the wild type (Fig. 4; red, blue). Because cry2 mutant plants flowered late under white light, we examined the flowering time ofcry2 mutant plants grown under light containing both blue and red wavelengths. Under this condition, cry2 mutant plants flowered significantly later than wild-type plants (Fig. 4; red + blue). Thus, the delayed flowering of cry2 mutant plants under white light can be phenocopied by growing the mutant plants under blue-plus-red light.

Figure 4

Effect of different wavelengths of light on the flowering time of Arabidopsis cry2 andphyB mutants. The flowering time was measured as described (Fig. 2) for the Columbia wild type (Col) and the cry2(cry2-1) and phyB (phyB-9) mutant plants grown under continuous red (75 to 90 μmol s−1m−2), blue (75 to 85 μmol s−1m−2), or blue-plus-red light (60 to 80 μmol s−1 m−2, with a ratio of red-light intensity to blue-light intensity of approximately 2 to 3) (10). Means of three independent experiments (individual samples contain more than 20 plants) with slightly different fluence rate from one experiment to another and the standard errors are shown.

Our results suggest that phytochromes mediate the red-light–dependent inhibition of flowering, whereas cry2 mediates the blue-light–dependent inhibition of phytochrome function. Phytochromes inhibit flowering in the absence of blue-light–dependent CRY2 activity such that red-light–grown wild-type plants flower late. In blue light, wild-type plants flower early, implying either the presence of a blue-light–dependent activator or the absence of a red-light–dependent inhibitor. Normal flowering of cry2 mutant plants in blue light indicates that the function of cry2 alone does not promote flowering under blue light. Thus, accelerated flowering of wild-type plants in blue light can be at least partially explained by the absence of the activity of the red-light–dependent inhibitors, phytochromes. Under white light or blue-plus-red light, red-light–dependent phytochrome activity and blue-light–dependent cry2 activity function in an antagonistic manner. In these light conditions,cry2 mutant plants flower late because the red-light–dependent phytochrome activity inhibiting floral initiation remains untamed as a result of the lack of the blue-light–dependent cry2 activity in the mutant plants.

We suggest that the function of both phytochromes and cry2 in flowering-time regulation are mediated by CO. The function of phytochromes proposed in our model is consistent with the observation that Arabidopsis hy1 and hy2 mutants, defective in the biosynthesis of phytochrome chromophore, flower earlier than the wild-type plants (6). It is not clear how many phytochrome species are involved in mediating red-light–dependent inhibition of flowering, although phyA is probably not associated with the flowering inhibition because the phyA mutant does not flower early (6). phyB mutant plants flower earlier than the wild-type plants grown under white light (6, 7), an effect mediated by CO (20). Thus, phyB could be one of the phytochromes that mediates red-light–dependent inhibition of flowering (4). Indeed, the early-flowering phenotype ofphyB is dependent on red light (Fig. 4). In blue light, however, phyB mutant plants flowered at about the same time as the wild type (Fig. 4; blue). Consistent with our model,phyB mutation can suppress the late-flowering phenotype ofcry2 under blue-plus-red light, whereas the cry2mutation cannot suppress the early-flowering phenotype ofphyB in red light (26).

Although our model explains the mode of action of cry2 and phyB in the regulation of flowering time of Arabidopsis, phyA and cry1 appear to function in different ways in this process (6, 8,9), and the relative importance of individual photoreceptors in mediating photoperiodic signals may be different in other plant species (9). It will also be interesting to learn the relationship between cry2 in photoperiodism and the circadian clock associated with blue-light–entrained circadian rhythms in plants (27).

  • * These authors contributed equally to this work.

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


Stay Connected to Science

Navigate This Article