- © 1999 American Society of Plant Physiologists
Abstract
Winter-annual ecotypes of Arabidopsis are relatively late flowering, unless the flowering of these ecotypes is promoted by exposure to cold (vernalization). This vernalization-suppressible, late-flowering phenotype results from the presence of dominant, late-flowering alleles at two loci, FRIGIDA (FRI) and FLOWERING LOCUS C (FLC). In this study, we report that flc null mutations result in early flowering, demonstrating that the role of active FLC alleles is to repress flowering. FLC was isolated by positional cloning and found to encode a novel MADS domain protein. The levels of FLC mRNA are regulated positively by FRI and negatively by LUMINIDEPENDENS. FLC is also negatively regulated by vernalization. Overexpression of FLC from a heterologous promoter is sufficient to delay flowering in the absence of an active FRI allele. We propose that the level of FLC activity acts through a rheostat-like mechanism to control flowering time in Arabidopsis and that modulation of FLC expression is a component of the vernalization response.
INTRODUCTION
The transition of shoot apical meristems from vegetative growth to flowering is a major developmental switch in the plant life cycle. The timing of floral initiation is critical for reproductive success, and many plant species have evolved multiple pathways to regulate flowering time. These pathways monitor both the developmental state of the plant and environmental cues such as photoperiod and temperature.
In Arabidopsis, which is a facultative long-day plant, studies of induced mutations have identified several loci that delay flowering when mutated and other loci that cause early flowering when mutated (Koornneef et al., 1998; Levy and Dean, 1998). The late-flowering mutants define genes that promote flowering, whereas the early-flowering mutants define genes that inhibit the transition to flowering. Physiological analyses of the effects of such mutations support a model in which these genes operate in parallel flowering pathways. For example, one group of late-flowering mutants (e.g., constans, fd, fe, fha, ft, fwa, and gigantea) exhibits little difference in flowering time when grown in short days compared with long days. Thus, these mutants define genes whose products promote flowering in a pathway that responds to inductive photoperiods (Koornneef et al., 1998; Pineiro and Coupland, 1998). In another group of late-flowering mutants (e.g., fca, fpa, fve, fy, and luminidependens [ld]), flowering is delayed under inductive (long-day) conditions and even more severely delayed during short days. Thus, this group retains a photoperiod response. Mutants in this group also exhibit a flowering response to extended exposure to cold (vernalization); the late-flowering phenotype of these mutants is suppressed by vernalization. The pathway in which these genes act is referred to as the autonomous or constitutive pathway to indicate that this pathway acts independently of photoperiod (Koornneef et al., 1998; Levy and Dean, 1998). The suppression of autonomous pathway mutants by exposure to cold indicates that a vernalization pathway can bypass blocks to the autonomous pathway.
The mutational analyses described above to identify flowering-time genes have been largely conducted in early-flowering ecotypes of Arabidopsis. However, many other Arabidopsis ecotypes are relatively late flowering when grown under inductive long-day conditions, and they become early flowering only after vernalization (Napp-Zinn, 1979; Sanda and Amasino, 1996a). Thus, the flowering behavior of these late-flowering ecotypes is similar to the flowering behavior of early-flowering ecotypes that have late-flowering mutations in autonomous pathway genes. Genetic analyses demonstrate that the difference in flowering behavior between late- and early-flowering ecotypes is due to allelic variation at one or both of two loci, FRIGIDA (FRI) and FLOWERING LOCUS C (FLC) (Napp-Zinn, 1979; Burn et al., 1993b; Clarke and Dean, 1994; Koornneef et al., 1994; Lee et al., 1994b). Late-flowering ecotypes contain dominant alleles of FRI and FLC that suppress flowering, whereas early-flowering ecotypes contain recessive fri and/or flc alleles. Thus, FRI and FLC act synergistically to delay flowering, and the effects of FRI and FLC are suppressed by vernalization (Lee and Amasino, 1995). These genes, therefore, have been critical in the evolution of the vernalization-responsive, winter-annual habit of certain Arabidopsis ecotypes (Napp-Zinn, 1979; Levy and Dean, 1998).
In this study, we report the molecular cloning of FLC and the effect of flc null mutants on flowering. FLC encodes a novel MADS domain protein, and the levels of FLC mRNA are controlled by FRI and other genes in the autonomous flowering pathway. Furthermore, overexpression of FLC from a heterologous promoter is sufficient to delay flowering. We propose that the level of FLC activity acts by using a rheostat-like mechanism to control flowering time in Arabidopsis.
RESULTS
Analysis of FLC Mutants
Naturally occurring late flowering in Arabidopsis is caused primarily by the interaction of two genes, FRI and FLC. The late-flowering phenotype conferred by FRI is suppressed in the Landsberg erecta (Ler) ecotype of Arabidopsis due to the recessive allele of the FLC gene (flc-Ler) present in Ler (Koornneef et al., 1994; Lee et al., 1994b). Similarly, late flowering caused by mutations in LD is also suppressed by flc-Ler (Koornneef et al., 1994; Lee et al., 1994b). To study the nature of the Ler allele of FLC and to determine specifically whether the ability to suppress the late-flowering phenotype of both FRI and mutations in ld could be conferred by flc loss-of-function mutations, we obtained additional flc alleles. In one screen, late-flowering ld plants (ld-3) were mutagenized with ethyl methanesulfonate (EMS), and one early-flowering M2 plant was isolated that in complementation tests proved to contain a mutation allelic to the recessive allele of FLC found in the Ler background. This allele was designated flc-1.
A second screen was performed using the late-flowering line homozygous for the late-flowering FRI allele from the ecotype San Feliu-2 (SF2) (FRI-SF2) in the Columbia (Col) background described in Figure 1A. Ninety-thousand seeds were treated with 5 to 6 krad of fast-neutron radiation, and 300,000 M2 plants were screened for early flowering. Because FLC-Col and FRI-SF2 synergistically interact to cause late flowering, it was likely that mutations in either gene could cause early flowering. Early-flowering plants were crossed to two “tester” lines to determine whether mutations were present in FRI or FLC. One tester was homozygous for the FRI-SF2 in the Ler background (FRI-SF2 in Ler), and the other was homozygous for the SF2 allele of FLC in the Ler background. Because these alleles of FRI and FLC behave dominantly, M2 plants containing a lesion in FRI would be early flowering when crossed to the FLC-SF2–containing line and late flowering when crossed to the FRI-SF2–containing line. Conversely, plants with a mutation in FLC would be early flowering when crossed to the FRI-containing line and late flowering when crossed to the FLC-containing line. From this screen, four flc alleles (designated flc-2 to flc-5) and three fri alleles were isolated.
Flowering Phenotype of Lines Described in This Study.
(A) Homozygous FRI-SF2 in the Col background (left) and a fast-neutron-induced flc allele in the same genetic background (right).
(B) Homozygous FRI-SF2 in the Ler background (left) and introduction of a genomic clone containing a late-flowering allele of FLC into a FRI-SF2 in Ler (right).
(C) Ler wild type (left) and introduction of constitutively expressed FLC into the Ler background (right).
The effect of all the flc mutations on flowering time is similar to that of the early-flowering allele of FLC found in Ler (Koornneef et al., 1994; Lee et al., 1994b). These flc alleles do not interact with FRI, and the extreme late flowering of FRI-SF2 in Col is eliminated in the presence of flc mutations, as shown in Figure 1A and Table 1. All flc alleles behave indistinguishably and represent the null phenotype (the lesions in these alleles are described below). Similarly, FRI mutants also become nearly identical in flowering time to Col, suggesting that Col contains a nonfunctional allele of FRI. Thus, our mutational analysis confirms the models developed from work with naturally occurring early- and late-flowering alleles of FRI and FLC: FRI and FLC interact synergistically to delay flowering, and a loss of function in either gene causes the late-flowering phenotype to be eliminated.
Isolation of FLC by Positional Cloning
To generate a segregating population useful for high-resolution mapping and positional cloning of FLC, we crossed the F1 plants generated from a cross of Ler to Col to the tester line described above containing the FRI-SF2 allele in Ler. This line contains a late-flowering allele of FRI but also contains the flc-Ler allele and is therefore early flowering. The progeny of the cross of the F1 plants to FRI-SF2 in Ler, which segregate 1:1 for late and early flowering, are late flowering when FLC-Col is present due to the interaction between FRI and FLC, but early flowering in the presence of flc-Ler. Test-cross progeny (4500 plants) were screened with the microsatellite markers nga158 and nga151 (Bell and Ecker, 1994), which previously had been shown to flank FLC. Plants containing recombination events between nga158 and nga151 were then tested with a third microsatellite marker nga249, which revealed that FLC resided in the interval between nga249 and nga151.
Flowering Times of FLC- and FRI-Containing Plant Lines
The region between nga249 and nga151 was contained within four yeast artificial chromosome (YAC) clones (see Methods). To generate additional markers, we determined the DNA sequence of YAC end clones and designed primers to amplify the corresponding sequences from Ler and Col. These DNAs from Ler and Col were sequenced to identify single-nucleotide changes, which then were used to create derived cleaved amplified polymorphic sequence markers (Michaels and Amasino, 1998; Neff et al., 1998). Markers derived from the left and right ends of YAC CIC1B8 detected recombination events on either side of FLC, demonstrating that FLC resided in the 620-kb interval spanned by CIC1B8.
A group of 13 BAC, TAC, and P1 clones have been identified by the Kazusa Arabidopsis genome project that span CIC1B8. These clones have insert lengths of ~70 to 100 kb and were used as probes on DNA blots with EcoRV-digested DNA from plants containing fast-neutron-induced flc mutations. Two overlapping clones, K6M1 and MYB9, shown in Figure 2A, detected several deleted bands in flc-2. Random 10- to 20-kb fragments of K6M1 and MYB9 resulting from partial digestion with Sau3A1 were used to create a library in the binary vector pPZP211 (Hajdukiewicz et al., 1994), and individual clones from this library were used to transform the FRI-SF2 in Ler line. The library was constructed with DNA from the Col background, which contains a late-flowering allele of FLC. Thus, FRI-SF2 in Ler plants transformed with a construct containing the Col allele of FLC will be late flowering. One of the clones from this library, 211-31, produced T1 plants that were very late flowering. More than one-third of the plants underwent senescence without flowering after 8 months of growth (Figure 1B).
Sequencing revealed three putative genes in 211-31 with similarity to MADS box transcription factors, a major pollen allergen, or ETHYLENE INSENSITIVE3 (Figure 2A). To determine which gene represented FLC, we examined the three candidate genes from two additional fast-neutron flc alleles, flc-3 and flc-4, and one EMS-generated allele, flc-1. Both of the fast-neutron alleles showed polymorphisms in bands resulting from the MADS box transcription factor. Determination of the DNA sequence of flc-1, flc-3, and flc-4 revealed that all contained lesions in the first exon of the MADS box transcription factor. flc-3 contains a 104-bp deletion that removes the start codon, and flc-4 contains a 7-bp deletion that results in a frameshift after the first 20 amino acids. flc-1 contains a single-base transition at the first exon–intron junction that changes the conserved GT donor site to AT and presumably disrupts splicing.
The genomic organization of the FLC gene is shown in Figure 2A. FLC is composed of seven exons, and the first exon contains the conserved MADS box domain that is the site of the lesions in flc-1, flc-3, and flc-4. An FLC cDNA was isolated from the Col background. The deduced amino acid sequence is highly related to MADS box–containing transcription factors and is shown in Figure 2B, with the positions of the MADS, I, and K boxes noted.
Cloning and Characterization of the FLC Gene.
(A) Summary of the positional cloning of FLC, genomic structure, and positions of mutations. Exons are represented by filled boxes, and lengths are give above; introns are represented by open boxes, and lengths are given below. Asterisks denote the positions of mutations.
(B) The predicted amino acid sequence of FLC, with the MADS, I, and K domains indicated.
Expression of FLC
One model for the dependence of the late-flowering phenotype of FRI and ld mutations on the presence of an active FLC allele is that the role of FRI and LD is to control FLC expression. To investigate this possibility, we performed RNA gel blot analyses to compare the steady state levels of FLC mRNA in shoots of wild-type Wassilewskija (Ws) and ld-3 in the Ws background, and a line homozygous for FRI-SF2 in the Col background and a fri mutant from the Col background. The results are shown in Figure 3A. The FLC transcript was only detected in the late-flowering lines containing FRI and the ld-3 mutation. Thus, FRI acts to increase the level of FLC mRNA, and LD acts to decrease the level.
The effects of developmental stage and photoperiod on the level of FLC mRNA were also examined in shoots of the FRI-SF2 in the Col line. Shoots were collected for RNA isolation from plants grown in long days at 4, 6, 9, 11, 13, 15, and 21 days after germination and from short-day-grown plants at 4, 9, 11, and 15 days after germination. These plants had not flowered and had formed ~18 leaves in long days and 10 leaves in short days. FLC mRNA levels were not affected by photoperiod or by plant age during vegetative development (Figure 3B).
Expression of FLC.
(A) Expression of FLC in Col wild type (Col), FRI-SF2 in Col (FRI), a fast-neutron allele of fri in Col (fri), Ws wild type, and the ld-3 mutant in the Ws background.
(B) Time course of FLC expression in RNA prepared from shoots. Time is measured as days after germination (DAG).
(C) Expression of FLC in various organs.
(D) Suppression of FLC expression by 30 days after cold treatment. Plants were grown for 10 days at 22°C before harvesting. Blots were probed with 18S rDNA as a control for loading.
The tissue distribution of FLC mRNA in FRI-SF2 in the Col line was examined by gel blot analysis of RNA isolated from leaves, vegetative apices, stems of flowering plants, roots, and the inflorescence (tissue from the first open flower to the shoot apical meristem) (Figure 3C). FLC is expressed most highly in the vegetative apex and in root tissue, but it is also detectable in leaves and stems. It is interesting that in young tissues of the inflorescence, FLC mRNA is not detectable, which is in sharp contrast with the high level in young tissues of the vegetative apex. This indicates that FLC expression is downregulated in the apex after the transition to flowering.
Effect of Vernalization on FLC Expression
As discussed above, the late-flowering phenotype of FRI and ld can be suppressed by loss-of-function flc mutations. The late-flowering phenotype of FRI and ld, in the presence of an active FLC allele, can also be suppressed by vernalization (Lee and Amasino, 1995). Therefore, the promotion of flowering by vernalization might involve a modulation of FLC expression. To explore this possibility, we determined the effect of cold treatment on FLC mRNA levels. Imbibed seeds of the FRI-SF2 in Col line were cold treated for 30 days and then transferred to soil with imbibed seeds of control plants that had not been cold treated. Shoots were collected after 10 days of growth for RNA gel blot analysis. The FLC transcript level was reduced after vernalization to undetectable levels (Figure 3D), similar to the difference in expression seen between FRI-SF2 in Col and fri null mutants (Figure 3A). Thus, the elimination of the late-flowering phenotype of FRI-SF2 in Col by vernalization is accompanied by a corresponding decline in FLC expression. It should also be noted that control plants, which were not cold treated and which exhibited high FLC expression, and the vernalized plants, which do not have detectable FLC expression, were both grown from seeds that were harvested from vernalized plants. This demonstrates that the reduction of FLC expression in vernalized plants is not meiotically stable (i.e., the decreased expression is not carried through to the next generation).
Increased FLC Expression Causes Late Flowering
As part of the positional cloning of FLC described above, the genomic clone containing FLC, 211-31, was transformed into a line homozygous for the SF2 allele of FRI in Ler. This line normally flowers in long days after forming ~12 primary rosette leaves (Table 1; Lee et al., 1994b), but lines in which 211-31 had been introduced into FRI-SF2 in Ler exhibited a severe delay in flowering due to the synergistic interaction of FRI with an active FLC allele (Figure 1B). Moreover, the severity of the late-flowering phenotype often greatly exceeded that of genetically constructed lines homozygous for the same alleles of FRI and FLC. Specifically, eight of 19 lines into which 211-31 had been introduced formed >80 rosette leaves and underwent senescence without flowering, even when grown under continuous far-redenriched light conditions that always cause natural FRI/FLC homozygotes to flower after forming <60 rosette leaves. The extreme late flowering of some of the 211-31 transformants might have been due to position effects or multiple insertion events that cause higher levels of FLC expression than that found in natural FRI/FLC homozygotes.
To determine whether enhanced FLC expression is sufficient to cause late flowering in the absence of FRI or a late-flowering mutant allele such as ld, we introduced into Ler the genomic coding region of FLC under control of the constitutive cauliflower mosaic virus (CaMV) 35S promoter. (Ler lacks FRI activity and contains functional alleles of the flowering-time genes defined by late-flowering mutants [Lee et al., 1994b; Koornneef et al., 1998].) Of the transgenic plants into which the 35S::FLC construct was introduced, 35% exhibited a moderate delay in flowering time (flowering after forming 10 to 20 leaves) and 35% were very late flowering (flowering after forming >45 leaves; Figure 1C). Under these growth conditions, Ler flowers with seven or eight leaves. Thus, overexpression of FLC alone is sufficient to cause a significant delay in flowering time.
DISCUSSION
FLC was first identified as a suppressor of the late-flowering phenotype conferred by FRI and ld mutations. The suppressor allele was uniquely found in the Ler ecotype of Arabidopsis (Koornneef et al., 1994). In this study, we describe the characterization of null mutants and demonstrate that suppression of the late-flowering phenotype is due to loss of FLC function. Thus, the role of active FLC alleles is to repress flowering. We have identified the FLC gene by positional cloning and found that it encodes a MADS domain protein. After this article was submitted to be considered for publication, Sheldon et al. (1999) reported that a DNA insertion into a MADS box gene caused delayed flowering. Sequence comparison showed that this DNA insertion creates an allele of FLC that dominantly delays flowering.
The MADS domain proteins comprise a large family of transcription factors in plants. In Arabidopsis, >28 family members have been identified (Riechmann and Meyerowitz, 1997), and they appear to play diverse roles in development. For example, AGAMOUS-LIKE15 (AGL15) is expressed preferentially in developing embryos (Heck et al., 1995; Rounsley et al., 1995), AGL3 is expressed in all tissues of the shoot (Ma et al., 1991), and AGL12, AGL14, and AGL17 are expressed only in roots (Rounsley et al., 1995).
The most well characterized of the MADS domain proteins are those involved in floral development, such as AG, APETALA1 (AP1), APETALA3 (AP3), PISTILLATA, and CAULIFLOWER (reviewed in Riechmann and Meyerowitz, 1997). Among the MADS domain genes of known function, the floral meristem identity genes AP1 and CAULIFLOWER may have a biological role most similar to FLC. These genes act early in the development of individual flowers downstream of FLC's role in floral induction. FLC exhibits amino acid sequence similarity to CAULIFLOWER (20%) and to AP1 (30%), but FLC is also 30% identical to AGL15, which is believed to not have a role in floral development. The sequence with the strongest similarity to FLC comes from a sequenced bacterial artificial chromosome clone (GenBank accession number AC002291), with a predicted amino acid sequence that is 58% identical to FLC. Thus, this gene and FLC appear to define a new subfamily of MADS domain proteins distinct from MADS domain proteins involved in floral development.
Based on the interactions of FRI, LD, and FLC, we proposed a model in which the role of LD is to counteract the FLC-mediated inhibition of flowering and the role of FRI is to augment this inhibition (Lee et al., 1994b). The results presented in this study are consistent with this model, shown in Figure 4, and reveal some of the molecular details. In late-flowering lines that contain dominant, late-flowering alleles of FRI, FLC mRNA levels are relatively high. In an early-flowering fri null mutant, FLC mRNA levels are below our limits of detection. LD acts to decrease FLC mRNA levels because in an ld null mutant, FLC mRNA levels are similar to those in a FRI-containing line. Thus, the similar flowering phenotypes of plants containing recessive “late-flowering” mutations in floral-promoting genes in the autonomous pathway and plants containing FRI are likely to be due, at least in part, to the increase of FLC expression in both situations. Indeed, we find that mutations in several genes in the autonomous pathway result in an elevation of FLC expression similar to that in the ld mutation (S.D. Michaels and R.M. Amasino, unpublished results), which provides a molecular model for the strong enhancement of the late-flowering phenotype of many autonomous pathway mutations by an active FLC allele (Sanda and Amasino, 1996b).
Except for ld and flowering locus d mutations, however, other autonomous pathway mutations as well as FRI cause a measurable late-flowering phenotype in the Ler genetic background, which appears to lack FLC activity (Koornneef et al., 1991; Sanda and Amasino, 1996b). Thus, FRI and other autonomous pathway genes may have roles in flowering-time control in addition to the regulation of FLC expression, whereas LD and FLD may act solely to control FLC activity.
Another common feature of plants containing late-flowering mutations in autonomous pathway genes and plants containing FRI is the suppression of the late-flowering phenotype by vernalization (Lee and Amasino, 1995; Koornneef et al., 1998). For example, the late-flowering effect of FRI is completely eliminated by 40 days of cold treatment (Lee and Amasino, 1995). Our results demonstrate that vernalization eliminates the FRI-mediated enhancement of FLC expression; after cold treatment of a FRI-containing line, FLC mRNA levels were undetectable. FLC mRNA levels were also undetectable in an early-flowering fri null mutant that had not been cold treated. Thus, the vernalization suppression of FLC expression could be mediated through the effect of cold treatment on FRI activity. However, the late-flowering phenotype of an autonomous pathway mutant, such as ld, in a fri null background is also completely eliminated by vernalization; thus, we favor a model in which the vernalization pathway can directly modulate FLC expression rather than acting through other autonomous pathway genes such as FRI and LD.
A model for the interactions of FLC with FRI and LD and other autonomous pathway genes and the effect of vernalization is presented in Figure 4. In this model, FLC activity plays a central role in the control of flowering time in the autonomous pathway, and FLC activity is negatively regulated by vernalization and positively and negatively regulated by other genes in the pathway. The results of overexpression studies are also consistent with this model: overexpression of FLC alone (i.e., in the absence of FRI or ld mutations) is sufficient to cause late flowering and phenocopy the effect of FRI and ld mutations.
Several lines of evidence indicate that FLC suppresses flowering by a rheostat mechanism in which the level of FLC activity is proportional to the lateness to flower. For example, our previous work has demonstrated that active FLC alleles delay flowering in a semidominant manner in FRI-containing or ld mutant lines (Lee et al., 1994a). Furthermore, in this study, the transgenic lines in which FLC has been introduced under the control of its native promoter or under the control of the CaMV 35S promoter exhibit a broad range of flowering times, indicating that variation in FLC expression, presumably due to transgene copy number and/or position effects, has a large influence on flowering time. In fact many of these transgenic lines never flowered under conditions that readily promote flowering of lines naturally homozygous for late-flowering alleles of FRI and FLC, indicating that expression at levels higher than those naturally occurring in FRI/FLC homozygotes can cause a substantial further delay in flowering.
Model for the Interaction of FLC, FRI, LD, and Vernalization in the Regulation of Flowering in the Autonomous Pathway.
Boldface lines indicate interactions with FLC, and dashed lines indicate other effects that occur in the absence of FLC activity.
Studies of FLC mRNA distribution reveal that FLC is expressed most highly in the vegetative shoot apex and in roots. This pattern of expression is similar to that of other autonomous pathway genes such as LD and FCA (Macknight et al., 1997; Aukerman et al., 1999). The high level of expression in the vegetative shoot apex of FRI-containing lines is consistent with a role for FLC in shoot apical meristem cells to locally inhibit the floral transition. The lack of detectable FLC mRNA in the inflorescence apex of FRI-containing lines indicates that a decrease in FLC expression is associated with the transition to flowering, and it is possible that this decrease is necessary for the floral transition to occur. The lack of detectable FLC mRNA in shoots of lines that have been vernalized also suggests that a decrease in FLC expression may be a necessary component of the vernalization response.
A hallmark of vernalization is that the vernalized state is maintained for long periods of time and through many mitotic cell divisions after the cessation of the inductive cold treatment. We have shown that prolonged cold treatment causes FLC mRNA levels in the shoot to decrease to undetectable levels, and this decline is maintained after cold-treated plants are returned to optimal warm growing conditions. Another hallmark is that the vernalized state is lost in the next generation after meiosis and sexual reproduction. The decrease in FLC expression after cold treatment also has this feature; vernalized plants produce seeds that develop into plants in which a high level of FLC expression is restored.
The expression of FLC in leaves and roots is also consistent with a role for a stable change in FLC expression contributing to the maintenance of the vernalized state. Although vernalization appears to cause a change that renders the shoot apical meristem competent to flower (Lang, 1965), cells from other regions of cold-treated plants, such as leaves (Metzger, 1988) and roots (Burn et al., 1993a), can regenerate into plants in which the shoot apical meristem exhibits the vernalized state. Thus, cells other than those of the shoot apical meristem must contain a system to affect a stable change in response to cold treatment, and a decrease in FLC expression may represent a component of that stable change in many cell types.
METHODS
Physical Map of the FLOWERING LOCUS C Region
The chromosomal positions of the microsatellite markers, yeast artificial chromosomes (YACs), bacterial artificial chromosomes, transformation-competent artificial chromosomes, and P1 clones described below are available at the Kazusa Arabidopsis data opening site: http://www.kazusa.or.jp/arabi/.
Creation of Lines
The following lines used in this study have been described previously: lines homozygous for FRIGIDA (FRI)-SF2 in Landsberg erecta (Ler), FLOWERING LOCUS C (FLC)-SF2 in Ler, FRI-SF2 in Columbia (Col; Lee et al., 1994b), Col and FRI in Col (Lee and Amasino, 1995), and Wassilewskija (Ws) and luminidependens-3 (ld-3) in Ws (Lee et al., 1994a).
Plant Growth Conditions
All plants were grown under cool-white fluorescent light at 22°C. Long-day conditions consisted of 16 hr of light followed by 8 hr of darkness; short-day conditions consisted of 8 hr of light followed by 16 hr of darkness. For experiments involving vernalization, imbibed seeds were incubated at 2°C on agar-solidified medium containing 0.65 g/L Peters Excel 15-5-15 fertilizer (Grace Sierra, Milpitas, CA). During cold treatment, plants were kept under short-day conditions.
RNA Gel Blot Analysis
Total RNA was isolated using RNA Isolator (Genosys Biotechnologies, The Woodlands, TX) according to the manufacturer's instructions. For RNA gel blots, 15 to 20 μg of RNA was separated by denaturing formaldehyde–agarose gel electrophoresis as described previously (Sambrook et al., 1989). RNA gel blots were probed with a 32P-ATP–labeled cDNA fragment that did not contain the conserved MADS box domain. Blots were also probed with an 18S rRNA probe as a control for the quantity of RNA loaded.
Fast-Neutron and Ethyl Methanesulfonate Mutagenesis
The ethyl methanesulfonate (EMS) mutagenesis of ld-3 seeds was conducted as follows. Eight thousand dry seeds were soaked in 0.5% EMS for 8 hr at 20°C and rinsed in eight changes of water over a 2-hr period. Seeds were then sprayed onto soil after being suspended in a solution of 0.1% agarose. Of the 8000 M1 seeds planted, ~5000 survived to set seed and were harvested in 50 pools.
For the mutagenesis of FRI-SF2 in Col by fast-neutron radiation, three aliquots of 50,000 seeds each were sent to the nuclear reactor at the Plant Breeding Unit of the Food and Agriculture Organization of the United Nations and International Atomic Energy Agency Program (Seibersdorf, Austria) where they were treated with 5, 5.5, or 6 krad of radiation. Seeds were subsequently sprayed onto soil in solution of 0.1% agarose. Only small differences were seen in viability in the M1 generation between the lots that had been exposed to different amounts of radiation and 60 to 70% of seeds germinated and survived to set seed. M1 seeds were harvested in 98 pools.
Analysis of Fast-Neutron Lines
Fast-neutron-generated lines known genetically to contain mutations in FLC were used to search for lesions in the three candidate genes contained on rescuing clone 211-31. Polymerase chain reaction was used to amplify 1- to 2-kb fragments that spanned the candidate genes. The amplified DNA was then cleaved with RsaI or TaqI to generate smaller fragments, separated on 5% nondenaturing acrylamide gels, and visualized with ethidium bromide.
ACKNOWLEDGMENTS
We are grateful to Ilha Lee for the genetic identification of FLC, to Caroline Dean for the analysis of FRI fast-neutron alleles, to Marty Yanofsky for assistance in the analysis of the FLC sequence, and to Satoshi Tabata for providing TAC and P1 clones. This research was supported by the College of Agricultural and Life Sciences of the University of Wisconsin and by grants to R.M.A. from the U.S. Department of Agriculture National Research Initiative Competitive Grants Program and the National Science Foundation.
- Received March 10, 1999.
- Accepted March 30, 1999.
- Published May 1, 1999.
REFERENCES
