Environmental-Dependent Acceleration of a Developmental Switch: The Floral Transition

See allHide authors and affiliations

Science's STKE  08 Feb 2000:
Vol. 2000, Issue 18, pp. pe1
DOI: 10.1126/stke.2000.18.pe1


The transition from vegetative growth to reproductive growth in plants in which flowers are produced requires the activation of specific genes. Simpson and Dean discuss two recent reports that characterize the FLOWERING LOCUS T ( FT ) gene in Arabidopsis , which is part of the floral transition pathway. Unlike many of the known genes that initiate flower production, the FT gene appears to encode a membrane-associated protein that could function in signaling from the cell surface.

Plants control the time at which they flower by responding to environmental cues and endogenous developmental signals ( 1 ). This control contributes to both reproductive success and adaptation. One example of this is the acceleration of flowering in Arabidopsis that accompanies increasing day length. This environmental cue reliably signals the onset of the favorable conditions of spring and summer. Two recent pieces of work address the molecular mechanism of floral promotion in response to day length through the cloning of the floral promoting gene, FLOWERING LOCUS T ( FT ) ( 2 , 3). FT closely resembles another protein that regulates the floral transition, TERMINAL FLOWER 1 (TFL1), and both potentially function in plasma membrane-based signal transduction cascades.

Floral organ development is served by a population of stem cells known as the shoot apical meristem (SAM). The SAM is formed during embryogenesis, and vegetative development proceeds through the production of leaf primordia at its flanks. Upon the floral transition, a change in meristem identity is accompanied by the production of floral primordia. Floral fate depends on the action of floral meristem identity (FMI) genes. In Arabidopsis , these include LEAFY ( LFY ) and APETALA 1 ( AP1 ), both of which encode transcription factors ( 4 , 5). If the activity of FMI genes is reduced, flowers develop with various shootlike structures. The activity of the FMI genes is opposed by a group of genes that maintain shoot identity and includes TFL1 ( 6 , 7).

Several lines of evidence identify LFY as a key player in the floral transition: LFY expression precedes that of other FMI genes [ LFY directly activates the transcription of AP1 ( 8 ) and the floral homeotic gene AGAMOUS ( 9 )], and ectopic overexpression of LFY promotes the formation of determinate meristems, indicating that LFY is sufficient to determine floral fate ( 10 ). lfy loss-of-function mutations exhibit the most severe floral defects of the FMI gene mutants, and wild-type LFY copy number affects flowering time, indicating that there is a quantitative relation between LFY activity and the floral transition ( 11 ). Consistent with this, the increase in LFY expression is delayed and reduced in genetic backgrounds bearing mutations in a number of genes that control flowering time ( 12 ).

There is more to the floral transition than the simple up-regulation of LFY expression, however. For example, when compared to mutations in flowering time control genes, lfy loss-of-function mutations do not flower particularly late in long days (LDs). Likewise, the early-flowering phenotype of transgenic plants that ectopically overexpress LFY from the cauliflower mosaic virus 35 S promoter ( 35S::LFY ) in LDs is attenuated in short days (SDs) ( 10 ). These findings indicate that there must be (at least) another activity that functions in parallel with LFY to control flowering time (which is more important in LDs than in SDs). This notion is supported by the effect of 35S::LFY on the late-flowering phenotype of mutations in different flowering time control genes: 35S::LFY only partially suppresses the late-flowering phenotype of fca and co ( 12 , 13), suggesting that transcriptional up-regulation of LFY accounts for only part of the normal function of these genes in controlling flowering time. Furthermore, because 35S::LFY has no effect on the late-flowering phenotype of ft , fe , or fwa ( 12 ), it suggests that the normal function of these genes in controlling flowering time does not involve transcriptional up-regulation of LFY .

What then, might function in parallel to LFY ? Genetic analyses have identified FT as a likely candidate ( 14 ). ft mutants flower late in LDs and slightly late in SDs ( 15 ). The severity of meristem identity defects is significantly enhanced in ft lfy double mutants, indicating that FT functions in a parallel pathway to LFY to activate other FMI genes [ AP1 mRNA is not detected in ft lfy double mutants ( 14 )].

The role of FT has now been clarified following its identification and characterization by Araki, Weigel, and their co-workers ( 2 , 3). Consistent with its relative importance in LDs, FT mRNA increases in LDs before the floral transition and to a higher level than that found in SDs. FT appears to function in a branched pathway principally downstream of the photoperiod promotive pathway gene, CONSTANS ( CO ). This is based on a decrease in FT mRNA in a co mutant background ( 2 , 3), but an increase upon CO activation ( 2 ), and also by the fact that 35S::FT completely suppresses the late-flowering phenotype of co ( 2 , 3), whereas the early flowering of 35S::CO is only partly attenuated in an ft mutant background [cited in ( 2 )].

The phenotype of transgenic plants expressing either 35S::LFY or 35S::FT is similar in that both flower early and with a terminal flower, illustrating that ectopic overexpression of either is sufficient to control flowering time and floral fate ( 2 , 3). In both cases, the floral transition is preceded by a distinct vegetative phase. However, the phenotype of plants simultaneously expressing both 35S::FT and 35S::LFY is striking: These plants flower extremely rapidly after producing only two leaves. Because these first leaves are already formed in the embryo, the floral transition in these plants occurs almost immediately without passing through a substantial vegetative phase ( Fig. 1 ). This acceleration of flowering is more dramatic than the combination of either 35S::LFY 35S::AP1 or 35S::FT 35S::AP1 and indicates that FT and LFY function in parallel to perform partly nonoverlapping roles in addition to activating AP1 . FT has been described as a factor required to confer competence of the meristem to respond to LFY activity ( 12 ). The present data are consistent with this interpretation. However, an alternative possibility is that competence can be defined as the ability to coordinate the up-regulation of both LFY and FT expression, not because FT is required downstream for full LFY activity, but because they perform parallel partly nonoverlapping roles.

Fig. 1.

Scanning electron micrograph of a transgenic Arabidopsis plant ectopically overexpressing both 35S::FT and 35S::LFY . The SAM has been transformed into a single terminal flower flanked by the cotyledons (c), which were formed in the embryo.

So is FT a transcription factor like LFY? Probably not. Ectopic expression of AP1 is visible in plants expressing 35S::LFY , but not in plants expressing 35S::FT ( 3 ), indicating that FT activates AP1 in a less direct manner than LFY ( 8 ). Instead, the sequence of FT reveals that it is related to another previously identified gene that controls the expression of FMI genes: TFL1 ( 6 ). Both FT and TFL1 are related to a widely occurring family of (possibly) membrane-associated proteins named phosphatidylethanolamine-binding proteins (PEBPs). The mechanism of action of this class of proteins is currently poorly understood. Human PEBP (hPEBP) is the precursor of hippocampal cholinergic neurostimulating peptide (HCNP), which is generated from its precursor by (possibly autocatalytic) proteolytic cleavage ( 16 ). hPEBP was also recently identified as being identical to Raf kinase inhibitor protein, which regulates the activity of the RAF/MEK/ERK signal transduction pathway ( 17 ). This kinase cascade controls the proliferation and differentiation of different cell types. The crystal structure of hPEBP suggests that the ligand-binding site could accommodate the phosphate head of membrane lipids allowing the protein to adhere to the inner leaf of bilipid membranes ( 18 ). The structure also suggests that ligand binding may lead to coordinated release of the NH2-terminal part of the protein ( 18 ). It is not yet clear that FT functions in a related manner to hPEBP, but there are notable relations between hPEBP structure-function and ft mutant alleles. Two of the strong ft alleles, ft-3 and ft-4 , affect two of the three residues implicated in possible autocatalytic cleavage in hPEBP, and the missense mutation of ft-1 is in a region that may regulate access to the PEBP ligand-binding site ( 2 , 3 , 18 , 19 ).

The structural similarity between FT and TFL1 and their regulation of FMI gene expression raise the possibility that they function in a mutually antagonistic manner. However, this is not the case. TFL1 is still active in 35S::FT plants because 35S::FT tfl1-1 plants flowered even earlier than 35S::FT plants ( 3 ). Therefore, these related proteins function in separate pathways to control FMI gene expression (it is unclear whether they have the same molecular targets).

Most genes involved in the floral transition, identified to date, encode proteins involved in transcriptional or posttranscriptional control ( 1 ). Because cell-cell communication is a fundamental principle of development ( 20 ), the molecular characterization of FT should add much to our understanding of flowering time control, if indeed it turns out to be a component of a membrane-associated, signal transduction cascade. Specifically, the identification of the positive and negative regulators of both FT and LFY, as well as the targets of their action, will provide information on meristem competence and the very essence of the floral transition.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
View Abstract

Navigate This Article