SPP 1530: Flowering Time Control - from Natural Variation to Crop Improvement

State of the art

Since the pioneering work of Maarten Koornneef who identified a number of FTi mutants in Arabidopsis thaliana this species has been used as a model for genetic studies of FTi regulation. Many key regulators of FTi control have been cloned and characterized in the past 15 years2. Scientists in Germany have been at the forefront of FTi research. In the following we give a short overview of the key genes and mechanisms that have been described in Arabidopsis and other models and compare these with the current data from crop species.


Unraveling the mechanisms of flowering time control in model and crop species


Arabidopsis is an annual species which flowers in the first year without extended exposure to cold temperatures. Biennials like sugar beet or cabbage, in contrast, have an obligate requirement for cold exposure over winter for achieving floral competence in a process called vernalization. Perennials like fodder grasses and trees often need many years to flower. It is of utmost importance for this project that key regulators and pathways have been found to be highly conserved across species borders. However, important differences exist. In some cases important crop flowering time genes are absent in the model Arabidopsis (e.g. Ghd7 of rice or VRN2 of wheat). In other cases the same gene confers a different response to environment, for example CO of Arabidopsis promotes flowering in response to long summer days, whereas in rice its orthologue represses flowering in long days. In general, changes in FTi behaviour, e.g. winter and spring types, can be due to sequence changes within individual FTi regulators. 


Flowering genes are involved in the initial establishment of a flowering-competent state, e.g. in the process of vernalization, and in floral transition when meristems switch to reproductive growth and develop flowers. Mutant analyses under various environmental conditions and genetic and molecular interaction studies showed that several distinct regulatory pathways exist. These pathways are generally referred to by the exogenous or endogenous cues that they respond to, i.e. the vernalization, photoperiod, gibberellin, and autonomous pathway. These pathways converge to regulate a set of 'floral integrator' genes that integrate the outputs of the various pathways and, under favorable conditions, directly activate floral meristem identity genes (Figure 1). Subsequently, the vegetative shoot apical meristem is turned into a floral meristem that initiates one or several flower primordia.



Figure 1: Simplified scheme of genetic, endogenous (blue) and environmental (red) factors for flowering time control in A. thaliana and implications on plant production.

Conservation and divergence of vernalization pathways


Plants in temperate climates have evolved a signal perception and transduction pathway that senses prolonged periods of cold over winter and translates this environmental cue into an increased competence to flower, a process known as vernalization. This process, often combined with a day length-sensing mechanism, ensures flower development and flowering under favorable conditions in spring or summer.


Vernalization requirement and response possess two intriguing features: a) The temporal separation between the plant's exposure to cold in winter and the onset of flowering in spring or summer, and b), the renewed vernalization requirement for flowering in subsequent generations. Work in Arabidopsis has revealed the epigenetic nature of the underlying processes and established at the molecular level how the removal of a block to flowering by vernalization enables the plant to respond to the second major seasonal stimulus, daylength. One of the key genes in Arabidopsis that regulate vernalization requirement and response is FLC, a MADS-box transcription factor that acts as a repressor of floral transition. The expression of FLC itself is tightly controlled by a plethora of both positive and negative regulators (Figure 1). The response to vernalization is facilitated by a cascade of gene regulatory processes and results in chromatin-based and mitotically stable repression of FLC. In the following generation, FLC expression is reset around the time of early embryogenesis, thus ensuring a renewed requirement for vernalization.


While FLC orthologs in rapeseed and other Brassica crops are clearly functionally related to FLC the extent of conservation of FLC and the vernalization pathway as it has been described for Arabidopsis outside the Brassicaceae is still controversial. FLC-like ESTs were identified in the three major core eudicot clades (rosids, asterids, and caryophyllids), but evidence for functional conservation of the corresponding genes is scarce3. A notable exception is the BvFL1 gene in sugar beet, a caryophyllid, which was shown to act as a repressor of flowering when transformed into an Arabidopsis FLC null mutant.

Among monocots, vernalization requirement and response is best understood in temperate cereals. Identification of the key regulators of vernalization requirement in wheat, VRN1, VRN2 and VRN3, revealed a regulatory pathway whose components differ from those of the FLC-dependent vernalization pathway but are homologous to or share conserved domains with other floral regulatory genes in Arabidopsis. Recent data further indicate that the changes in expression of VRN1 in barley during vernalization and maintenance of its state of transcriptional activity post-vernalization involve chromatin modification. Thus, both Arabidopsis and the cereals use an epigenetic mechanism to bridge the temporal separation between induction of a flowering-competent state by vernalization and initiation of flowering in the spring, although the known targets, Arabidopsis FLC and wheat VRN1, are not orthologous and have opposite effects on floral transition. Despite some similarities in terms of protein domain organization and epigenetic regulation, the stark differences in vernalization pathways in Arabidopsis and temperate cereals suggests that vernalization requirement and response evolved independently in the dicot and monocot lineages. This creates a need for further research to unravel the FTi regulation in crop species.


Photoperiod and circadian clock control of flowering time


The extensive variation in seasonal temperature changes during the evolution of flowering plants and the divergence of vernalization pathways (s. above) stands in contrast to the stable annual photoperiodic conditions during the history of the earth, which is consistent with an apparently higher degree of functional conservation of photoperiodic control of flowering across taxa.


Photoperiod and seasonal changes in day length is a second environmental cue with major effects on the timing of flowering during the course of a year. In the (facultative) long-day plant Arabidopsis, a key role in the regulatory pathway involved is played by the plant-specific CCT domain transcription factor CO (CONSTANS). CO is regulated at the transcriptional level by the circadian clock. Photoperiod control of floral transition through CO and homologous genes is widely conserved among flowering plants. Supporting evidence includes the identification of CO-like genes from many monocot and dicot species4, and complementation of the co mutation in Arabidopsis by CO-like genes from species as distantly related to Arabidopsis as sugar beet or green algae5. However, the wide latitudinal variation between photoperiodic conditions and the migration histories of plant species are likely factors that contributed to the evolution also of differences in photoperiodic response mechanisms among flowering plants. Thus, although the CO-dependent promotion of flowering by activation of FT or its orthologs appears to be conserved among long-day plants, short-day plants and day-neutral plants differ from this scheme.


Interestingly, previous studies had shown that FLC and some of its regulators modulate the period of the circadian clock. Together with recent results on the regulation of the FLC-interacting protein SVP, these data may suggest a regulatory loop or feedback interactions involving circadian clock genes and FLC and SVP. Although on the basis of the current evidence the regulatory interactions would appear relatively weak, they may allow the fine-tuning of the plant's response to more subtle environmental changes such as changes in ambient temperature.



Integration of different pathways and environmental factors to one output


In Arabidopsis, the inputs from the vernalization and photoperiod pathways are integrated by floral integrator genes that include the MADS-box gene SOC1, FT and the FT homolog TSF, which are strong promoters of flowering. FLC represses FT in leaves and thus prevents production of FT protein, the long searched mobile signal ‘florigen’ that moves to the shoot apex and promotes flowering  by activation of AP14. In effect, the floral integrators are fully expressed only after elimination of FLC repression as a result of vernalization, and activation under long-day conditions by the photoperiod pathway. Like CO, FT and other floral integrator genes are largely conserved between dicots and monocots.



Clearly, other genetic factors than those discussed above also contribute to the regulation of floral transition. The effect of changing ambient temperatures on FTi is thought to be mediated by several genes, at least some of which are also regulated by other floral regulatory pathways and thus may contribute to the integration of different environmental signals. FTi is also affected by various environmental stresses such as drought and heat. The plant’s response to these stresses may be fine-tuned, at least in part, by differential interaction of CO with a HAP-like protein complex6.


Flowering time regulation in perennials


The flowering behavior differs significantly between perennials and annuals/biennials. Annuals and biennials go through senescence and die after flowering. In contrast, perennials live for many years and flower repeatedly. Perennials produce reproductive and vegetative meristems during one growth season. While the floral meristems terminate their growth by developing flowers, the vegetative meristems continue to grow during the subsequent season. Floral transition in meristems which stayed vegetative in one growth season is initiated in the following year after perception and transduction of seasonal floral promoting signals such as cold or specific photoperiods. Evidently, pathways linked to vernalization and photoperiod not only regulate flowering but also regulate seasonal growth cessation and release from dormancy in perennial plants.


Arabis alpina has been used as a model plant for perennials and first studies have shown that key genes are differentially regulated compared to annual Arabidopsis to confer characteristic patterns of perennial development7. Vice versa, downregulation of flowering genes in Arabidopsis resulted in phenotypes common to perennial plants suggesting only small molecular differences between perennials and annuals. Flowering time QTL were located in perennial and annual forage grasses (L. perenne, L. multiflorum) but no genes have been identified so far. Other studies with woody perennials identified genes involved in seasonal floral transition (e.g. strawberry, apple)8. Research is needed to find the links between floral inducing signals and genetic factors to control floral transition in woody perennials to tackle the problem of alternate flowering/fruit bearing in fruit trees.


Annual and perennial plants enter a juvenile stage after germination. Plants are unable to respond to flowering inducing signals during this stage. Juvenility is species-specific and can be very short or last several years as it is observed e.g. for trees. Overexpression of some floral promoting genes or suppression of floral repressor genes have broken or at least shortened the juvenile phase of trees like citrus, poplar or apple8. Also epigenetic mechanisms seem to have an impact on the juvenile-to-adult vegetative switch. However, factors that maintain the juvenile stage and thus repress flowering are largely unknown and need to be investigated in the future.


Plant hormones

A direct influence of plant hormones in controlling flowering time sets in with the initiation of floral meristems. In dependence of its concentration and timing, auxin acts as a trigger stimulating the initiation of the inflorescence meristem. Auxin might either derive from local auxin biosynthesis, or it might be provided through the polar auxin transport pathway originating from the shoot apical meristem9. On the other hand, differentiation of an axillary meristem into a floral meristem is suppressed by cytokinins, which balance between meristem differentiation and maintenance and thereby positively regulate meristem size. While a contribution of root-to-shoot translocation of cytokinins in maintaining meristem size is so far just weakly supported by physiological studies, the essentiality of local cytokinin biosynthesis within the meristem has been demonstrated in rice by mutations in the LOG gene, encoding an enzyme involved in the last step of cytokinin biosynthesis or in CKX encoding a cytokinin oxidase required for cytokinin degradation10. More recently, transcriptome analyses of differentiating inflorescence meristems also indicated an involvement of abscisic acid and jasmonic acid, which might provide a regulatory link to the strong influence of abiotic stress factors on flowering time (see “Flowering time and the adaptation to changing environmental conditions”).


The role for gibberellins (GAs) in floral transition has so far been characterized at a lower resolution. It is well established that under long-day growth conditions or in biennial plants gibberellins mediate a photoperiodic stimulus to flowering that relies on an upregulation of GA20-oxidase gene expression in leaves. Evidence for GAs themselves representing this stimulus has so far only been obtained in Lolium. Under short-day conditions, however, the contribution of GA-dependent regulatory pathways even increases and becomes obligatory. For instance in Arabidopsis, GAs promote flowering through the activation of genes encoding the floral integrators SOC1, LFY and FT in the inflorescence and floral meristems11.


Flowering time and the adaptation to changing environmental conditions


Sequence variations among FTi genes are important for the adaptation to changing environmental conditions, both natural and artificial. Plants from northern geographical regions germinate late and remain in the vegetative phase over winter whereas plants flower early under arid climate conditions to avoid drought stress. There is increasing evidence for a rapid evolution of FTi in response to a climate fluctuation, e.g. by shifting the onset of flowering to earlier dates12.


The great impact of FTi genes for the adaptation of crops to local environments and production methods has been demonstrated. In seed crops, flowering should be as early as possible to extend the seed filling phase, to avoid harsh environmental conditions which endanger seed production or harvest (e.g. drought, heat, frost), or to escape pathogen attack. By contrast, delayed flowering may be desirable to realize high yields in biomass for energy production. A number of flowering time QTL were located in association studies using natural exotic and elite populations, e.g. wheat, barley13, rice, maize, rapeseed. In cereals, where natural allelic variation was exploited from early on during domestication and selective breeding for FTi traits (as co-determinants of yield) led to the prevalence of different alleles in different growing areas, genetic mapping and association studies with landraces and breeding material readily identified key loci with major effects on FTi (e.g. VRN1 to -3, Ppd1).


Selection for FTi traits in the past was based exclusively on phenotypic characteristics and relied on natural variation present in the primary and secondary gene pool of a crop species. In the future FTi genes can be used by breeders as functional markers for selecting favorable genotypes, for quality control of seed lots, or for targeted manipulation of flowering traits by genetic modification.


At first glance the application of FTi markers appears to be of limited value because the onset of flowering is easy to score. Therefore, FTi markers will hardly be used for routine selection. However, marker selection may be superior over phenotypic selection after crossing adapted elite parents with non-adapted (exotic) parents with inadequate flowering behavior. If this is due to individual FTi genes, favorable plants can be easily identified in early generations by a marker test whereas phenotypic selection on single plants can be problematic due to environmental interaction.


Future work is needed in at least four areas: first, elucidation of how much functional allelic variation is present for these major genes in wild and domesticated cereal germplasm. Second, identification of new FTi regulators and their interaction with internal and external factors. Third, analyses of how these major genes interact both with other key players and with the environment in crop species compared to model species like Arabidopsis, rice and Brachypodium. And fourth, to find out, if newly characterized functional alleles can be used in plant breeding to re-adjust the traits flower initiation, flowering time and flowering duration to future breeding goals in the light of changing climate conditions and agricultural practices.


Pleiotropic effects of flowering time genes


Yield potential, plant height and heading date are three classes of traits that determine the productivity of many crop plants. Of immediate relevance to yield potential, and therefore of great interest in plant breeding, is hybrid vigor or heterosis. A recently finished SPP was dealing with heterosis in plants and its genetic and metabolic reasons. There is increasing evidence that FTi regulators and genes showing functional similarity are key players in establishing heterosis in plants. Recently it was demonstrated that growth vigor in hybrids of A. thaliana and A. arenosa was caused by repression of the circadian clock genes CCA1 and LHY14. These findings may have some impact on future heterosis research. The rice gene Ghd7 was cloned from a major QTL for hybrid yield present in many Chinese rice hybrid cultivars15. This gene encodes a CCT domain, typical for CO and related proteins. It has major effects on an array of traits in rice, including number of grains per panicle, plant height and heading date. Likewise, a homolog of FT was found to account for heterosis in tomato16. Comparative QTL analysis of heterosis in Arabidopsis and oilseed rape suggested that key flowering time loci may coincide with significant QTL hotspots involved in regulation of biomass, metabolite levels and seed yield. These results open new horizons for breeding research because they suggest that FTi gene expression might trigger a cascade of regulatory effects with a broad global effect on plant development and yield.


Novel strategies by altering flowering time regulation


Genetic variation for FTi beyond the natural variation can be increased by mutagenesis and transformation. Two or more mutations at different FTi loci can be combined by crossing, and plants with novel FTi behavior can be selected. For example, assuming that FTi repressors can behave in an additive or synergistic manner, the hypothesis shall be tested that crossing two mutant plants with delayed flowering phenotypes will produce a non-flowering hybrid. If confirmed, this strategy will be followed to develop prototypes for breeding biennials and perennials with novel FTi characters2.


Novel FTi characteristics can be also generated through targeted genetic modification by transformation. The current knowledge of FTi control has been exploited through either overexpression or suppression of gene activity2. Several papers have been published describing transgenic plants with either up- or down-regulation of FTi genes. In many cases the phenotypic effects were as desired resulting in earlier or delayed flowering or even complete avoidance of flowering 2. This demonstrates that manipulation of single genes can have drastic effects on FTi and, as demonstrated above, other agronomic characters in crop plants. The forage grass of the future should produce non-flowering or reduced culms because leaf blades are more digestible and of higher feeding value than sheaths and culms. Likewise, the future beet should be planted before winter which requires full FTi control.


Literature Cited


2. Jung,C. and Müller,A. (2009) Flowering time control and applications in plant breeding. Trends in Plant Science 14, 563-573

3. Reeves,P.A. et al. (2007) Evolutionary conservation of the FLOWERING LOCUS C-mediated vernalization response: evidence from the sugar beet (Beta vulgaris). Genetics 176, 295-307

4. Turck,F. et al. (2008) Regulation and identity of florigen: FLOWERING LOCUS T moves center stage. Annual Reviews in Plant Biology 59, 573-594

5. Chia,T. et al. (2008) Sugar beet contains a large CONSTANS-LIKE gene family including a putative CO homologue that is independent of the early-bolting (B) gene locus. Journal of Experimental Botany 59, 2735-2748

6. Wenkel,S. et al. (2006) CONSTANS and the CCAAT box binding complex share a functionally important domain and interact to regulate flowering of Arabidopsis. Plant Cell 18, 2971-2984

7. Wang,R.H. et al. (2009) PEP1 regulates perennial flowering in Arabis alpina. Nature 459, 423-U13820

8. Flachowsky,H. et al. (2009) A review on transgenic approaches to accelerate breeding of woody plants. Plant Breeding 128, 217-226

9. Barazesh,S. and McSteen,P. (2008) Hormonal control of grass inflorescence development. Trends in Plant Science 13, 656-662

10. Kurakawa,T.e.al. (2007) Direct control of shoot meristem activity by a cytokinin-activating enzyme. Nature 445, 652-655

11. Mutasa-Göttgens,E. and Hedden,P. (2009) Gibberellin as a factor in floral regulatory networks. Journal of Experimental Botany First published online 5 Mar 2009;

12. Sherry,R.A. et al. (2007) Divergence of reproductive phenology under climate warming. Proceedings of the National Academy of Sciences of the USA 104, 198-202

13. Stracke,S. et al. (2009) Association mapping reveals gene action and interactions in the determination of flowering time in barley. Theoretical and Applied Genetics 118, 259-273

14. Ni,Z.F. et al. (2009) Altered circadian rhythms regulate growth vigour in hybrids and allopolyploids. Nature 457, 327-3U7

15. Xue,W.Y. et al. (2008) Natural variation in Ghd7 is an important regulator of heading date and yield potential in rice. Nature Genetics 40, 761-767

16. Lifschitz,E. et al. (2006) The tomato FT ortholog triggers systemic signals that regulate growth and flowering and substitute for diverse environmental stimuli. Proceedings of the National Academy of Sciences of the United States of America 103, 6398-6403


Latest Events