The Synergistic Activation of FLOWERING LOCUS C by FRIGIDA and a New Flowering Gene AERIAL ROSETTE 1 Underlies a Novel Morphology in Arabidopsis

, and Vojislava Grbi

Corresponding author: Vojislava Grbi, University of Western Ontario, London, ON N6A 5B7, Canada., vgrbic{at}uwo.ca (E-mail)

Communicating editor: V. SUNDARESAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The genetic changes underlying the diversification of plant forms represent a key question in understanding plant macroevolution. To understand the mechanisms leading to novel plant morphologies we investigated the Sy-0 ecotype of Arabidopsis that forms an enlarged basal rosette of leaves, develops aerial rosettes in the axils of cauline leaves, and exhibits inflorescence and floral reversion. Here we show that this heterochronic shift in reproductive development of all shoot meristems requires interaction between dominant alleles at AERIAL ROSETTE 1 (ART1), FRIGIDA (FRI), and FLOWERING LOCUS C (FLC) loci. ART1 is a new flowering gene that maps 14 cM proximal to FLC on chromosome V. ART1 activates FLC expression through a novel flowering pathway that is independent of FRI and independent of the autonomous and vernalization pathways. Synergistic activation of the floral repressor FLC by ART1 and FRI is required for delayed onset of reproductive development of all shoot meristems, leading to the Sy-0 phenotype. These results demonstrate that modulation in flowering-time genes is one of the mechanisms leading to morphological novelties.

THE establishment of body plan in plants occurs throughout their postembryonic development as a result of the continuous generation of new organs initiated by shoot apical meristems. While plants display a great variability in form, they are composed of a series of repeating body segments, called metamers, that have the same basic structure. Each metamer consists of an internode and a node, which is generally composed of a leaf and its subtended axillary meristem. Variations in metameric structure, like variation of the internodal length, suppression of leaf or axillary meristem development, or transformation of leaves and axillary meristems into specialized structures, lead to variation in plant form. Even metamers formed on a single plant differ in their morphology depending on the phase of the life cycle. For example, Arabidopsis plants initiate leaves with associated secondary shoots during vegetative growth. These metamers differ in leaf shape and trichome density, defining the juvenile and the adult vegetative phases (TELFER et al. 1997 ). A more pronounced morphological alteration of vegetative metamers is elongation of their internodes. While most vegetative metamers have short internodes, metamers that form at the end of the vegetative phase have elongated ones. Reproductive metamers form upon transition to flowering. They consist of an elongated internode, a suppressed leaf, and an axillary meristem that converts to a floral meristem. Therefore, three types of metamers characterize an Arabidopsis body plan (Fig 1A). The vegetative metamers with compressed internodes (V1) form a basal rosette of leaves, the vegetative metamers with elongated internodes (V2) form the bottom of the inflorescence stem, and reproductive metamers (R) form solitary flowers at the top of the inflorescence. The sequence of these metamers is fixed (V1 V2 R), which is a reflection of the irreversible transition of Arabidopsis plants to flowering.

View larger version (45K): In this window In a new window Download PPT slide   Figure 1. Morphology of Ler and Sy-0 plants. (A) A 30-day-old Ler plant. Its body plan can be described as V1 V2 R, where V1 and V2 vegetative metamers form a basal rosette and the bottom of the inflorescence stem and reproductive metamers (R) designate solitary flowers at the top of the inflorescence stem. (B) An 80-day-old Sy-0 plant that can be described as V1 V2* R* R. V1 metamers form a basal rosette of leaves, V2* marks aerial-rosette-bearing nodes at the bottom of the inflorescence stem, and R represents solitary flowers at the top of the inflorescence. (C) Vegetative rosette of a 55-day-old Sy-0 plant. (D) Detail of an aerial rosette borne on an Sy-0 plant illustrating V2* metamer. (E) Detail of the apex of the Sy-0 primary shoot that has reverted to vegetative development. (F) Detail of the flower from Sy-0 that has reverted to the inflorescence meristem, an R* metamer. (G) A 50-day-old Ler plant. (H) A 1.5-year-old old Sy-0 plant.

A large number of genes that control the timing of the transition to flowering have been identified in Arabidopsis by mutant analysis (PEETERS and KOORNNEEF 1996 ). They were grouped in genetic pathways that mediate responses to multiple environmental and developmental cues (SIMPSON and DEAN 2002 ). The photoperiod and the vernalization pathways act to promote flowering by mediating environmental responses to light and cold. The autonomous and the gibberellin pathways promote floral transition probably by mediating endogenous cues that reflect the developmental state of the plant. The FRIGIDA (FRI) gene identifies a floral repression pathway whose input signal is at present unknown. The autonomous and the vernalization pathways promote the floral transition by reducing the level of the floral repressor FLOWERING LOCUS C (FLC). FRI acts as floral repressor by promoting the expression of FLC. Thus, the inputs from the autonomous, vernalization, and FRI pathways integrate to determine the level of FLC floral repressor. The inputs from the FLC floral-repressing pathway and floral-promoting photoperiod and gibberellin pathways converge to regulate the expression of floral pathway integrator genes (FT, AGL20/SOC1, and LFY). These genes upregulate the function of floral meristem identity genes (AP1, CAL, FUL, and LFY), which act as a genetic switch to specify the floral developmental fate.

The timing of flowering depends primarily on inputs from floral-promoting and FLC-repressing pathways. It determines the length of the vegetative developmental phase influencing the number of vegetative metamers formed. Late-flowering plants form an enlarged basal rosette of leaves due to the increased number of V1 type metamers. Mutations in meristem identity genes eliminate reproductive metamers: in lfy ap1 plants, flowers (reproductive metamers) are replaced by shoots with subtending leaves (therefore with V2 metamers; WEIGEL et al. 1992 ). Overexpression of these genes causes conversion of lateral shoots into flowers, resulting in the modification of V2 metamers. Therefore, modulations of the flowering pathway have a profound effect on plant architecture. However, in most cases these alterations affect the number of metamers formed and not their identity. Only in 35S:LFY and 35S:AP1 backgrounds do the vegetative metamers (V2) alter their morphology due to the conversion of axillary meristems directly into flowers (MANDEL and YANOFSKY 1995 ; WEIGEL and NILSSON 1995 ).

In an effort to understand the mechanisms underlying the evolution of novel plant forms, we have initiated the study of the naturally occurring variant of Arabidopsis thaliana, Sy-0. The body plan of Sy-0 plants differs significantly from the morphology of most early or late-flowering Arabidopsis strains. The salient morphological feature of Sy-0 plants is the formation of aerial rosettes in the axils of cauline leaves. Therefore, in Sy-0 plants V2 vegetative metamers have altered morphology due to the prolonged vegetative development of axillary meristems (GRBIC and BLEECKER 1996 ). We previously identified two genes, ENHANCER OF AERIAL ROSETTE (EAR) and AERIAL ROSETTE (ART), that are required for the Sy-0 phenotype. In this report, we show that EAR is a dominant allele of FRI and that the ART locus represents a complex containing two linked genes, FLC and AERIAL ROSETTE 1 (ART1). ART1 is a new flowering gene that identifies a separate floral repression pathway. Its effect on flowering is mediated through activation of the floral repressor FLC. We provide evidence that the Sy-0 phenotype arises due to the synergistic activation of FLC by FRI and ART1, demonstrating that modulation of flowering-time genes can lead to alteration of metamer structure and to morphological novelties.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant material and growth conditions:
The seed for Sy-0, accession 1204, was obtained from the Arabidopsis Information Service Stock Center (Frankfurt, Main, Germany). Seeds for Landsberg erecta (Ler) and lines homozygous for morphological markers lutescence (lu), male sterile 1 (ms1), and transparent testa glabra (ttg) were obtained from the Arabidopsis Biological Resource Center at Ohio State University. FRI-Sf2, FLC-Col, and FRI-Sf2/FLC-Col homozygous lines in Ler background were kindly provided by Richard Amasino (University of Wisconsin, Madison). Monogenic lines containing Sy-0 alleles of FRI, FLC, and ART1 loci were backcrossed to Ler at least five times.

All plants were grown under 100–150 µE m^-2 sec^-1 cool-white fluorescent light at 22° under long-day conditions consisting of 16 hr of light followed by 8 hr of darkness. As variability in flowering time has been observed between experiments, control plants (parents used in the cross) were always grown in parallel to the progeny that had been tested. Seedlings for the RNA analysis were grown on half strength Murashige and Skoog medium (Sigma, Irvine, UK) under 75 µE m^-2 sec^-1 cool-white fluorescent light at 22° and were harvested when the first true leaves were 1–2 mm in length.

Mapping and the analysis of the FRI-Sy-0 allele:
Cleaved amplified polymorphic markers generated from left borders of MUG13, MAC12, MWD9, MDJ22, K5A21, MRN17, MYJ24, and K19M13, as well as NIT4, were used for mapping the ART1 and the upstream modifier of ART1 present in lu Ms1 ART1 Ttg recombinant lines. The chromosomal positions of these clones can be found at http://www.arabidopsis.org/servlets/mapper. Analysis of the FRI-Sy-0 allele was done by PCR amplification of genomic fragments containing two identified deletion sites within the FRI sequence. The primers used and PCR conditions were as described in JOHANSON et al. 2000 .

RNA gel blot analysis:
Total RNA was extracted using the RNAwiz RNA isolation reagent (Ambion, Austin, TX) according to the manufacturer's instructions. Ten micrograms of total RNA was run on a 1x TBE agarose gel and transferred to a nylon membrane. A partial FLC cDNA fragment that was kindly provided by Richard Amasino (University of Wisconsin, Madison) and described in MICHAELS and AMASINO 1999 was used as an FLC probe. A partial cDNA FRI fragment was synthesized by RT-PCR using the following primers: 5'-GATTTGCTGGATTTGATAAGG-3' and 5'-TTCAATGACCACCGTAAAGG-3'. Products were cloned into pGEMT-Easy (Promega, Madison, WI). The identity of the FRI cDNA fragment was initially confirmed with diagnostic restriction digests and was subsequently sequenced. It was used as a FRI probe. Probes were labeled using the Rediprime II labeling kit (Amersham, Uppsala, Sweden).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The morphology of Sy-0 plants:
Sy-0 is a late-flowering accession, which results in the formation of an enlarged basal rosette common to all Arabidopsis strains that have delayed flowering (Fig 1B and Fig C). The unique aspect of the Sy-0 body plan is the formation of aerial rosettes in the axils of cauline (stem) leaves and reversion of inflorescence and floral meristems (Fig 1, D–F). In Arabidopsis, the primary shoot apical meristem usually irreversibly switches from vegetative to reproductive development, giving rise to leaf-bearing nodes at the bottom of the inflorescence and flower-bearing nodes at the top. In Sy-0, leaf-bearing nodes occasionally form when the plant has already initiated 10 flowers (Fig 1E). This suggests that after switching to reproductive development, the primary shoot apical meristem has reverted to vegetative development. In addition, early flowers of Sy-0 plants regularly show reversion of the floral to an inflorescence meristem, which is seen as formation of a branch from the middle of the flower (Fig 1F). If Sy-0 morphology is viewed in terms of metamer structure, it can be described as V1 V2* R* R, where V2* corresponds to the aerial-rosette-bearing nodes, R* to reverting flowers, and the double-headed arrow to the reversion of an inflorescence meristem. Therefore, in Sy-0, all shoot apical meristems—the primary shoot apical, axillary, and floral meristems—have delayed establishment of reproductive development leading to modulation of either number or identity of Arabidopsis metamers.

The heterochronic shift in shoot meristem development has a profound effect on plant morphology (Fig 1A, Fig B, Fig G, and Fig H). A consequence of this morphological change is the extension of vegetative development beyond the transition to flowering. This change in the life history strategy increases the life span of the plant to >1.5 years compared to the three-month life span of the Ler plants (Fig 1G and Fig H). It also enables Sy-0 plants to form a greater number of secondary branches as a greater number of leaves and axillary meristems form. This affects overall plant fitness, since Sy-0 plants produce more seed per plant than many other Arabidopsis strains, giving Sy-0 an adaptive advantage at least under some environmental conditions.

Identification of genes underlying changes in morphology of Sy-0 plants:
The EAR locus corresponds to the FRI-Sy-0 allele: In our previous work, we determined that EAR, a gene required for the aerial rosette phenotype of Sy-0, maps on chromosome IV in the vicinity of the FRI gene (GRBIC and BLEECKER 1996 ). This led to the possibility that a dominant allele at the FRI locus might be involved in specification of the Sy-0 phenotype. To test this hypothesis, we investigated the nature of the FRI-Sy-0 allele. A survey of 40 Arabidopsis accessions by JOHANSON et al. 2000 established a correlation between the early flowering phenotype and loss of FRI function due to the two independent deletion events. When the FRI-Sy-0 allele was tested for the presence of these deletions, neither was found, indicating that FRI-Sy-0 could be an active allele (Fig 2A). In addition, Northern blot analysis revealed that FRI-Sy-0 RNA is expressed, further suggesting that the FRI-Sy-0 allele may be functional (Fig 2B).

View larger version (29K): In this window In a new window Download PPT slide   Figure 2. Characterization of the FRI-Sy-0 allele. (A) An analysis of the promoter and 16-bp deletions in Sy-0, ART line, Ler, and Col. The FRI-Ler allele has a promoter deletion and the FLC-Col allele has a 16-bp deletion. (B) Expression of FRI in Ler, Sf-2, and Sy-0. (C) Flowering time of FRI-Sy-0, FLC-Col, and their F1 progeny.

FRI is a putative transcription factor that activates the floral repressor FLC, leading to delayed flowering only if functional alleles at both loci are present (MICHAELS and AMASINO 1999 ; SHELDON et al. 1999 ). Therefore, we tested the ability of the FRI-Sy-0 allele to delay flowering in the presence of a reference FLC-Col allele. F₁ plants derived from a cross between FRI-Sy-0 and FLC-Col lines flowered much later than either of the parents, indicating that the FRI-Sy-0 allele is active and that the previously identified EAR locus specifies an allele of the FRI gene (Fig 2C).

ART1 is a novel flowering locus identified in the ART line: The ART locus has been identified as another factor required for the late-flowering aerial-rosette-bearing phenotype of Sy-0 plants (GRBIC and BLEECKER 1996 ). The ART-containing line (ART line) flowers late, after initiating 76 rosette leaves. The late-flowering phenotype segregates as a single semidominant gene (36 art/art; 56 ART/art; 23 ART/ART, chi-square = 3.02, P > 0.1) with the heterozygous genotypic class overlapping both early and late-flowering homozygous plants (Fig 3A).

View larger version (23K): In this window In a new window Download PPT slide   Figure 3. (A) Frequency distribution of flowering time of F2 segregating population of cross between ART line and Ler. (B) The location of ART1 on chromosome V. Distances were calculated using the Map Maker program.

ART has been mapped to the short arm of chromosome V in a region that contains several other genes implicated in transition to flowering (GRBIC and BLEECKER 1996 ). To further refine the position of ART, recombinants around it were selected from the cross between the ART line (ART/ART) and a line homozygous for the morphological markers lu, ms1, and ttg. ART was initially located between Ms1 and Ttg markers and later positioned on a 150-kb genomic fragment contained within the MYJ24 and MKD15 bacterial artificial chromosome clones (Fig 3B). As this region does not contain any genes previously implicated in flowering, ART identifies a novel flowering locus designated ART1.

The ART1-induced delay in flowering requires the presence of additional gene(s): During the course of mapping the ART1 gene, we identified numerous recombinant lines. Their phenotypic classification and flowering time are shown in Fig 4. Plants that had the Lu Ms1 ttg phenotype (Fig 4A), and therefore having recombination between Ms1 and Ttg markers, were genotyped for the ART1 locus using the NIT4 PCR marker. Plants heterozygous at the ART1 locus displayed a range of flowering times, from early to late, in the same pattern as seen in the F₂ segregating population shown in Fig 3A. Similarly, art1/art1 plants were early flowering.

View larger version (23K): In this window In a new window Download PPT slide   Figure 4. Frequency distribution of flowering time of recombinant classes derived from the cross between homozygous lines Lu Ms1 ART1 Ttg (Sy-0) and lu ms1 art1 ttg (Ler). (A) Lu Ms1 ttg recombinant class that segregates for ART1/art1 and art1/art1 plants. (B) lu Ms1 ART1 Ttg recombinant class; double crossing over late-flowering plants is marked with an asterisk. (C) Lu ms1 art1 ttg recombinant class.

However, plants that had the lu Ms1 Ttg phenotype (Fig 4B), which were heterozygous at the ART1 locus, flowered early (like the art1/art1 genotypic class in Fig 4A) instead of displaying a range of flowering times. This indicated that ART1 is not sufficient for the delayed flowering seen in the ART line and that some other factor(s), uncoupled from ART1 in this recombinant class, must be present for late flowering.

To identify such additional factor(s), we examined the genetic organization of the lu Ms1 ART1 Ttg plants. The genomic region upstream of Ms1 was homozygous for Ler alleles (Fig 4B). Upon genotyping the ART line with markers spanning the Ms1-Lu region, we found that it contained an Sy-0 genomic segment despite the five backcrosses of the ART line to Ler. This indicated that an Sy-0 region upstream of Ms1 may contain a factor required for the ART1-induced late flowering. Within lu Ms1 ART1 Ttg plants, 25 out of 379 showed varying degrees of delayed flowering characteristic for the ART1/art1 heterozygotes. To test the possibility that these plants flowered late due to the presence of an upstream factor, we genotyped them by using a MUG13 PCR-based marker that maps upstream of the Lu locus. In these plants a double crossover had occurred (plants marked with the asterisk in Fig 4B), making them heterozygous for the upstream Sy-0 chromosomal fragment. This suggests that the region upstream of Ms1 harbors a factor(s) from Sy-0 that is required for the ART1-induced delay in flowering.

The Lu ms1 art1 ttg recombinant class contains only the upstream region from Sy-0 (Fig 4C). All of these plants flowered early, demonstrating that the upstream factor(s) alone has no effect on flowering.

FLC could be the upstream factor required for the ART1-mediated delay in flowering: FLC is a known repressor of flowering that maps 2 cM upstream of Lu (LEE et al. 1994 ; KOORNNEEF et al. 1994 ). Therefore, there is a possibility that Sy-0 contains a functional FLC allele, which may be the upstream factor required for the ART1-induced late flowering.

Three independent lines of evidence suggest that the FLC-Sy-0 allele is functional. First, FLC is expressed in Sy-0 (Fig 5A). Sy-0 accumulates approximately the same level of FLC transcript as San Feliu-2 (Sf2), an Arabidopsis accession from which the reference FLC-Sf2 allele has been isolated (LEE et al. 1994 ). Second, a cross between early flowering FLC-Sy-0 and FRI-Sf2 monogenic lines yielded late-flowering F₁ progeny, indicating that a typical synergistic interaction exists between active FRI and FLC alleles (Fig 5B). Finally, the F₂ population of a cross between the FLC-Sy-0 and ld segregated plants that flowered significantly later than either of the parents (Fig 5C). Since LD acts to suppress the function of FLC (MICHAELS and AMASINO 1999 ), in FLC ld plants the functional FLC allele is derepressed and thus capable of delaying flowering in these plants. These results indicate that the FLC-Sy-0 allele is functional and therefore FLC could be the upstream factor required for the ART1-mediated delay in flowering.

View larger version (23K): In this window In a new window Download PPT slide   Figure 5. Characterization of the FLC-Sy-0 allele. (A) Expression of FLC in Sf-2, Ler, FRI-Sf2 in Ler, and Sy-0. (B) Flowering time of FLC-Sy-0, FRI-Sf2, and their F1 progeny. (C) Frequency distribution of flowering time of F2 segregating population of cross between FLC-Sy-0 and ld.

Identification of the genetic interactions underlying morphological changes in Sy-0 plants:
The genetic analysis of the late-flowering aerial-rosette-bearing phenotype indicated that Sy-0 alleles at FRI, FLC, and ART1 loci are required for the morphology of Sy-0 plants. Monogenic lines of each of these genes are all early flowering (Table 1, Fig 6). Thus, none of the Sy-0 alleles of these loci act alone in their effect on flowering. To identify allelic interactions leading to the Sy-0 morphology, we combined some of these alleles in crosses and determined the phenotype of their progeny.

View larger version (66K): In this window In a new window Download PPT slide   Figure 6. Flowering phenotype of lines in Ler background derived from Sy-0. (A) FRI-Sy-0. (B) FLC-Sy-0. (C) ART1. (D) ART1 FLC. (E) ART1 FLC FRI.

ART1 activates FLC independently of FRI: The F₁ progeny of a cross between ART1 and FLC flowered later than either of the parents, indicating synergistic interaction between the alleles examined (Fig 7A). This indicates cooperative action between genes that may be in the same pathway or that act through parallel pathways to produce the same outcome. To test if ART1 acts in the same pathway as FLC and upstream of it, we analyzed the FLC mRNA levels in the ART1-containing lines. As seen in Fig 7B, FLC mRNA cannot be detected in Ler plants. Therefore, detectable FLC transcripts in any genotype signify the activation of FLC expression.

View larger version (46K): In this window In a new window Download PPT slide   Figure 7. Analysis of genetic interactions between ART1, FRI, and FLC. (A) Flowering time of ART1, FLC-Sy-0, and their F1 progeny. (B) Expression pattern of FLC in various genotypes. (C) Analysis of genetic interactions between ART1 and FRI. (D) Expression pattern of FLC in various genotypes. Sy-0 alleles of corresponding loci are marked by capital letters, Ler alleles are not shown, and Col and Sf2 alleles are marked.

FLC mRNA can be detected in the ART1 line, indicating that ART1 activates FLC (Fig 7B). The effect of ART1 on FLC expression is allele specific, as FLC-Col mRNA accumulates at a higher level than FLC-Ler when combined with ART1. Given that the transcripts encoded by these two alleles are the same (SHELDON et al. 2000 ), this difference in mRNA levels can be attributed to the differential capacity of these FLC alleles to be upregulated by ART1.

Activation of FLC by ART1 occurs in the FRI-Ler background and therefore in the absence of a functional FRI allele. This indicates that ART1 acts through a FRI-independent pathway to activate FLC expression. In addition, these results confirm that FLC is the downstream factor required for the ART1-mediated delay in flowering previously seen in the ART line.

ART1 acts cooperatively with FRI to activate FLC expression: A cross between early flowering lines homozygous for ART1 and FRI-Sf2 yielded F₁ progeny that flowered late (Fig 7C). The synergistic interaction between these loci indicates that they act on a common factor. Since this interaction occurs in the presence of the FLC-Ler allele, which has a reduced capacity to be upregulated, it was necessary to determine whether these genes interact through FLC or some other downstream factor. If they act through FLC, then one would expect that ART1 and FRI could activate the FLC-Ler allele.

Northern analysis of FLC expression shows that ART1 and FRI-Sf2 alone have a limited capacity to induce the FLC-Ler allele (Fig 7D). However, they cooperatively induce FLC-Ler to high levels, similar to those seen for the FRI-Sf2 FLC-Sy-0 line. This suggests that FRI- and ART1-specific pathways converge at or upstream of FLC to activate its expression.

ART1 acts independently of the autonomous flowering pathway: Several flowering pathways integrate to affect FLC expression. The results described above establish that ART1 activates FLC expression through a FRI-independent pathway. Likewise, this activation occurs despite the functional alleles at LD, FPA, FVE, and FCA loci that act to repress the FLC expression through the autonomous pathway, suggesting that ART1 acts to activate FLC independently of these genes. If ART1 acts independently of the autonomous pathway, then one would expect a synergistic interaction between ART1 and mutant alleles that eliminate the FLC repression imposed by the autonomous pathway.

This hypothesis was tested by analyzing the flowering time of an F₂ segregating population derived from a cross between the ART1 FLC line and lines homozygous for ld, fve, fpa, or fca recessive alleles. The effect of ART1 FLC and mutations that disrupt the autonomous pathway was additive regarding the flowering time. The latest-flowering F₂ plants flowered after producing an additional 10–15 leaves relative to the ART1 FLC line. However, the latest flowering plants also formed aerial rosettes phenocopying the Sy-0 phenotype (Fig 8, A–E), indicating synergistic interaction between these loci in respect to the aerial rosette phenotype. These interactions indicate that ART1 acts independently of the autonomous pathway to affect timing of flowering and aerial rosette formation. In addition, the formation of aerial rosettes in a fraction of F₂ plants suggests that modulation of FLC expression underlies this phenotype.

View larger version (90K): In this window In a new window Download PPT slide   Figure 8. Late-flowering aerial-rosette-bearing plants from the F2 segregating population of a cross between the ART1 FLC line and fca (A), fve (B), fpa (C), and ld (D) homozygous lines. (E) Chi-square test analysis of aerial rosette phenotype in crosses shown in A–D.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic bases of Sy-0 morphology:
Sy-0 plants form an enlarged basal rosette and develop aerial rosettes in the axils of cauline leaves. The aerial rosette formation was inseparable from the late-flowering phenotype. In addition, their inflorescence meristem infrequently displays reversion of flowering. Also, several floral meristems, the first ones to form upon transition of the plant to reproductive development, show floral reversion. Therefore, in Sy-0, all shoot meristems—primary, axillary, and floral—have delayed establishment of reproductive development. However, once the reproductive development is established, both inflorescence and floral meristems develop normally.

In this article we establish that Sy-0 alleles at FRI, FLC, and ART1 loci are required for the Sy-0 phenotype. FRI and FLC are repressors of flowering previously identified from late-flowering Arabidopsis accessions. FRI is an activator of FLC, which acts to repress the onset of reproductive development (SIMPSON and DEAN 2002 ; Fig 9). ART1 is another locus required for the Sy-0 phenotype. It is a novel flowering locus that activates FLC expression. It maps 14 cM proximal to FLC on chromosome V. Thus, ART1 and FLC form a set of floral repressors that behave genetically as a single gene. Linked floral repressors on chromosome V, named FLF and FLG, have also been identified in the early flowering Cvi accession of Arabidopsis. These genes may correspond to late-flowering alleles of FLC and ART1 (ALONSO-BLANCO et al. 1998 ). In this context, it is interesting that Cvi flowers early, which may be attributed to two other quantitative trait loci, EDI and FLH, that promote flowering in such a way as to suppress the effect of FLF and FLG.

View larger version (18K): In this window In a new window Download PPT slide   Figure 9. Model for the interactions of FLC, FRI, and ART1 in the regulation of flowering time.

Monogenic lines homozygous for Sy-0 alleles at FLC, FRI, and ART1 loci introgressed into Ler background are all early flowering (Table 1 and Fig 6). However, each flowers slightly later then Ler, which correlates with the detectable expression of FLC in these lines (Fig 7). Early flowering of monogenic lines derived from Sy-0 indicates the importance of genetic interactions among these loci for establishment of the Sy-0 phenotype. We have shown that ART1 activates FLC through a FRI-independent pathway. As predicted from such a model, ART1 acts cooperatively with FRI to activate FLC expression. The interactions between Sy-0 alleles at these loci are required for the late-flowering aerial-rosette-bearing Sy-0 phenotype.

Vernalization suppresses the effect of FLC on flowering (SHELDON et al. 2000 ). It also suppresses the delayed flowering of Sy-0 plants (GRBIC and BLEECKER 1996 ), indicating that the effect of ART1 on flowering can be abolished by vernalization. Two possibilities may account for the effect of vernalization on flowering time in Sy-0. Vernalization can decrease the expression or the activity of ART1, in which case ART1 is a member of the vernalization pathway acting to increase FLC expression. Alternatively, ART1 activates FLC expression through a vernalization-independent pathway, in which case the interaction between these pathways regulates the level of FLC expression. We favor the latter possibility as ART1-Sy-0 and ART1-Ler alleles confer different FLC responses in the absence of vernalization (Fig 7), yet the presence of these alleles does not affect the plants' ability to respond to cold. In addition, we have provided evidence that ART1 activates FLC expression independently of genes in the autonomous FLC-repressing pathway. Therefore, we propose that ART1 identifies a novel FLC-activation flowering pathway (Fig 9).

Implication of Sy-0 morphology:
Upregulation of FLC expression underlies the late-flowering phenotype of many Arabidopsis accessions (JOHANSON et al. 2000 ). The common phenotypic feature among these accessions is formation of an enlarged basal rosette of leaves, which forms due to the prolonged vegetative development of the primary shoot apical meristem. However, these accessions still follow the basic body plan V1 V2 R, characteristic for the majority of Arabidopsis strains.

The Sy-0 morphology can be described as V1 V2* R* R, indicating formation of two new types of metamers, V2* and R*. The formation of aerial-rosette-bearing nodes (V2*) has also been observed in some short-day-grown 35S:TFL plants (RATCLIFFE et al. 1998 ). It has been proposed that TFL regulates the shoot apical phase transitions in such a way that its overexpression retards progression through phase transitions. In some 35S:TFL plants it resulted in prolonged vegetative development of the primary and axillary meristems leading to a late-flowering aerial-rosette-bearing phenotype. Floral reversion (marked by R* metamers) is another hallmark of the Sy-0 phenotype. This aspect of the Sy-0 phenotype has also been described for heterozygous lfy and homozygous ag plants grown in a short-day photoperiod. These plants have flowers that display the same heterochronic transformation of flowers into inflorescence meristems (OKAMURO et al. 1996 ). Recently, this phenotype has been attributed to the reduced floral meristem identity maintenance function provided by LFY (PARCY et al. 2002 ).

A phenotype similar to Sy-0 has been described for indeterminate1 (id1) maize mutant (COLASANTI et al. 1998 ). It has been proposed that ID1 regulates the synthesis of a floral-promoting signal or its transmission from leaves to shoot meristems, resulting in its absence at shoot meristems and their delayed conversion to reproductive development in id1 plants. Reversion of flowering in Impatiens balsamina provides an additional example. In this case, the reversion of flowering occurs due to the depletion of the leaf-borne flower-promoting signal (POUTEAU et al. 1997 ).

We have previously proposed a model to explain the heterochronic shift common to all shoot meristems in Sy-0 (GRBIC and BLEECKER 1996 ). According to the model, the Sy-0 phenotype arises due to either the deficiency of floral-promoting signals at shoot meristems or the lack of competence of shoot meristems to respond to these signals. This model can explain the phenotypes of short-day-grown lfy/+ and ag plants, the id1 maize mutant, and reverting Impatiens plants, in which presumably floral-promoting signals are lacking at shoot meristems that display reversion. Likewise, short-day-grown 35S:TFL plants may lack a competence to respond to floral-promoting signals, resulting in the extension of developmental phases.

Our data indicate that the late-flowering aerial-rosette-bearing Sy-0 phenotype requires synergistic activation of FLC by ART1 and FRI. Interactions between ART1 and mutations that disrupt the autonomous FLC-repression pathway also lead to formation of aerial rosettes, suggesting that modulation of FLC expression in these genotypes underlies the establishment of the late-flowering aerial-rosette-bearing phenotype. How can modulation of FLC expression be integrated into the above model?

The mechanism by which FLC specifies the Sy-0 phenotype is at present unknown. However, upregulation of FLC expression may not be sufficient for its establishment. Sf-2 plants have comparable FLC expression levels and flower at approximately the same time as Sy-0 plants, but do not form aerial rosettes (Fig 5A and data not shown). Moreover, while aerial-rosette-bearing plants were among the latest flowering ones in the segregating populations, there was an overlap in flowering time between plants that did and did not form aerial rosettes (data not shown). This implies that there is specific modulation of FLC expression that leads to Sy-0 phenotype.

The FLC pathway represses the expression of floral pathway integrators (SIMPSON and DEAN 2002 ). Its expression pattern has not been studied in great detail, but it has been shown that it localizes to the shoot and root tips (MICHAELS and AMASINO 1999 , MICHAELS and AMASINO 2000 ). This suggests that synergistic activation of FLC expression by ART1 and FRI in Sy-0 plants could act to decrease the expression of floral pathway integrators at shoot meristems. Therefore, the Sy-0 phenotype would arise due to the lack of floral-promoting signals at all shoot apical meristems, leading to their delayed conversion to reproductive development.

Modulated expression of flowering genes can lead to variability in plant form. However, in most cases it results in an altered number of metamers formed and not in the change of their identity. Experimental manipulations of expression of LFY, AP1, and TFL can cause novel metamer morphology, indicating that these genes could be targets for selection leading to novel plant form. Moreover, variations in the temporal or spatial expression of LFY and TFL may account for many of the different inflorescence types (COEN and NUGENT 1994 ). However, it is still unclear whether these genes were targets for the natural selection to derive new forms. Our analysis of the Sy-0 ecotype demonstrates that naturally selected alteration in FLC expression underlies the evolution of novel morphological form. In this case, the morphological novelty arose by upregulation of the floral repressor FLC by ART1 and FRI, demonstrating that modulation of flowering-time genes has been employed to produce novel plant morphology. Future molecular characterization of the new flowering gene ART1, which underlies specific modulation of FLC expression, should shed more light on the mechanism of plant morphological evolution.

	FOOTNOTES

¹ Present address: Center of Applied Genetics, University of Agricultural Sciences, Vienna, A-1190 Wien, Austria.

	ACKNOWLEDGMENTS

We thank Alan Noon and Ian Craig for photographic and art work. We thank Tony Bleecker and Caroline Dean for critical reviews of this manuscript. The work presented here was funded by grants from the European Molecular Biology Organization ALTF 695-1996 and the Natural Sciences and Engineering Research Council of Canada.

Manuscript received October 14, 2002; Accepted for publication January 7, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALONSO-BLANCO, C., S. E. EL-ASSAL, G. COUPLAND, and M. KOORNNEEF, 1998  Analysis of natural allelic variation at flowering time loci in the Landsberg erecta and Cape Verde Islands ecotypes of Arabidopsis thaliana.. Genetics 149:749-764.[Abstract/Free Full Text]

COEN, E. S. and J. M. NUGENT, 1994  Evolution of flowers and inflorescences. Dev. Suppl. 107–116.

COLASANTI, J., Z. YUAN, and V. SUNDARESAN, 1998  The indeterminate gene encodes a zinc finger protein and regulates a leaf-generated signal required for the transition to flowering in maize. Cell 93:593-603.[Medline]

GRBIC, V. and A. B. BLEECKER, 1996  An altered body plan is conferred on Arabidopsis plants carrying dominant alleles of two genes. Development 122:2395-2403.[Abstract]

JOHANSON, U., J. WEST, C. LISTER, S. MICHAELS, and R. AMASINO et al., 2000  Molecular analysis of FRIGIDA, a major determinant of natural variation in Arabidopsis flowering time. Science 290:344-347.[Abstract/Free Full Text]

KOORNNEEF, M., H. BLANKENSTIJN-DE VRIES, C. HANHART, W. SOPPE, and T. PEETERS, 1994  The phenotype of some late-flowering mutants is enhanced by a locus on chromosome 5 that is not active in the Landsberg erecta wild-type. Plant J. 6:911-919.

LEE, I., S. D. MICHAELS, A. S. MASSHARDT, and R. M. AMASINO, 1994  The late-flowering phenotype of FRIGIDA and mutations in LUMINIDEPENDENS is suppressed in the Landsberg erecta strain of Arabidopsis.. Plant J. 6:903-909.

MANDEL, M. A. and M. F. YANOFSKY, 1995  A gene triggering flower formation in Arabidopsis.. Nature 377:522-524.[Medline]

MICHAELS, S. D. and R. M. AMASINO, 1999  FLOWERING LOCUS C encodes a novel MADS domain protein that acts as a repressor of flowering. Plant Cell 11:949-956.[Abstract/Free Full Text]

MICHAELS, S. D. and R. M. AMASINO, 2000  Memories of winter: vernalization and the competence to flower. Plant Cell Environ. 23:1145-1153.

OKAMURO, J. K., B. G. DEN BOER, C. LOTYS-PRASS, W. SZETO, and K. D. JOFUKU, 1996  Flowers into shoots: photo and hormonal control of a meristem identity switch in Arabidopsis.. Proc. Natl. Acad. Sci. USA 93:13831-13836.[Abstract/Free Full Text]

PARCY, F., K. BOMBLIES, and D. WEIGEL, 2002  Interaction of LEAFY, AGAMOUS and TERMINAL FLOWER1 in maintaining floral meristem identity in Arabidopsis.. Development 129:2519-2527.

PEETERS, A. J. M. and M. KOORNNEEF, 1996  Genetic variation of flowering time in Arabidopsis thaliana.. Cell Dev. Biol. 7:381-389.

POUTEAU, S., D. NICHOLLS, F. TOOKE, E. COEN, and N. BATTEY, 1997  The induction and maintenance of flowering in Impatiens.. Development 124:3343-3351.[Abstract]

RATCLIFFE, O. J., I. AMAYA, C. A. VINCENT, S. ROTHSTEIN, and R. CARPENTER et al., 1998  A common mechanism controls the life cycle and architecture of plants. Development 125:1609-1615.[Abstract]

SHELDON, C. C., J. E. BURN, P. P. PEREZ, J. METZGER, and J. A. EDWARDS et al., 1999  The FLF MADS box gene: a repressor of flowering in Arabidopsis regulated by vernalization and methylation. Plant Cell 11:445-458.[Abstract/Free Full Text]

SHELDON, C. C., D. T. ROUSE, E. J. FINNEGAN, W. J. PEACOCK, and E. S. DENNIS, 2000  The molecular basis of vernalization: the central role of FLOWERING LOCUS C (FLC).. Proc. Natl. Acad. Sci. USA 97:3753-3758.[Abstract/Free Full Text]

SIMPSON, G. G. and C. DEAN, 2002  Arabidopsis, the Rosetta stone of flowering time? Science 296:285-289.[Abstract/Free Full Text]

TELFER, A., K. M. BOLLMAN, and R. S. POETHIG, 1997  Phase change and the regulation of trichome distribution in Arabidopsis thaliana.. Development 124:645-654.[Abstract]

WEIGEL, D. and O. NILSSON, 1995  A developmental switch sufficient for flower initiation in diverse plants. Nature 377:495-500.[Medline]

WEIGEL, D., J. ALVAREZ, D. R. SMYTH, M. F. YANOFSKY, and E. M. MEYEROWITZ, 1992  LEAFY controls floral meristem identity in Arabidopsis.. Cell 69:843-859.[Medline]

This article has been cited by other articles:

				M. Abe, M. Fujiwara, K.-i. Kurotani, S. Yokoi, and K. Shimamoto Identification of Dynamin as an Interactor of Rice GIGANTEA by Tandem Affinity Purification (TAP) Plant Cell Physiol., March 1, 2008; 49(3): 420 - 432. [Abstract] [Full Text] [PDF]

				C. Shindo, C. Lister, P. Crevillen, M. Nordborg, and C. Dean Variation in the epigenetic silencing of FLC contributes to natural variation in Arabidopsis vernalization response. Genes & Dev., November 15, 2006; 20(22): 3079 - 3083. [Abstract] [Full Text] [PDF]

				F. Tooke, M. Ordidge, T. Chiurugwi, and N. Battey Mechanisms and function of flower and inflorescence reversion J. Exp. Bot., October 1, 2005; 56(420): 2587 - 2599. [Abstract] [Full Text] [PDF]

				J. D. Werner, J. O. Borevitz, N. H. Uhlenhaut, J. R. Ecker, J. Chory, and D. Weigel FRIGIDA-Independent Variation in Flowering Time of Natural Arabidopsis thaliana Accessions Genetics, July 1, 2005; 170(3): 1197 - 1207. [Abstract] [Full Text] [PDF]

				S.-I Lin, J.-G. Wang, S.-Y. Poon, C.-l. Su, S.-S. Wang, and T.-J. Chiou Differential Regulation of FLOWERING LOCUS C Expression by Vernalization in Cabbage and Arabidopsis Plant Physiology, March 1, 2005; 137(3): 1037 - 1048. [Abstract] [Full Text] [PDF]

				I. R. Henderson and C. Dean Control of Arabidopsis flowering: the chill before the bloom Development, August 15, 2004; 131(16): 3829 - 3838. [Abstract] [Full Text] [PDF]

				S. Feng, Y. Shen, J. A. Sullivan, V. Rubio, Y. Xiong, T.-p. Sun, and X. W. Deng Arabidopsis CAND1, an Unmodified CUL1-Interacting Protein, Is Involved in Multiple Developmental Pathways Controlled by Ubiquitin/Proteasome-Mediated Protein Degradation PLANT CELL, July 1, 2004; 16(7): 1870 - 1882. [Abstract] [Full Text] [PDF]

				H.-S. Yoon and D. A. Baum Transgenic study of parallelism in plant morphological evolution PNAS, April 27, 2004; 101(17): 6524 - 6529. [Abstract] [Full Text] [PDF]