Floweringtime Mutants

Arabidopsis: Most of the Arabidopsis mutants which are relevant to photoperiodism have been generated by chemical- or irradiation-induced mutagenesis of the Columbia (Co) and Landsberg erecta (Ler) ecotypes. Both are examples of early summer-annual races, i.e. they grow and flower in early spring as the days are getting longer, and have little vernalisation requirement. They behave as facultative LDP; flowering is induced by LD but although flowering is delayed it is not prevented in SD. The critical

FLOWERING-TIME MUTANTS Chromosome number 2 3 4


fha ~fb


fve f2

51 61

fwa fd

f5 le

FIG. 9.1. Location of induced late flowering loci in Arabidopsis thaliana. After Law et al. 1993.

daylength for induction is about 8-10 h and plants flower after about three weeks with 4-7 leaves under LD and 7-10 weeks with about 20 leaves under SD. The precise behaviour of the WT lines varies between laboratories, probably reflecting differences in growth conditions, especially light conditions and temperature.

Arabidopsis flowering time mutants fall into two classes. Firstly, late-flowering mutants in which flowering is delayed in LD and maybe also in SD, and secondly, early-flowering mutants where flowering occurs earlier in SD and perhaps in LD. Most of the initial attempts to isolate flowering time mutants were carried out with Co or Ler under greenhouse conditions using natural photoperiods which, in general, are in excess of the critical daylength of 8-10 h. The WT lines flower quite rapidly under these conditions and the most obvious mutants were those which flowered later than the non-mutated lines. From a series of flowering screens, induced mutations for late flowering were isolated by Redei (1962), Hussein and Van der Veen (1965, 1968) and Koorneef et al. (1991). The latter authors carried out the most comprehensive analysis, placing 42 independently isolated mutants in Ler into 11 complementation groups and placing them on the Arabidopsis linkage map (Fig. 9.1). The mutants were designated fca, fd, fha, fe, fb, fg, ft, fy, fpa, fwa and fve. The two mutants, to and gi, isolated by Redei (1962) in the Co background, were found to be allelic to fg and fb respectively. The mutants could be assigned to one of three phenotypic groups based on their response to daylength and vernalisation.

• Group I: Late flowering - vernalisation sensitive. This group includes fy, fpa, fve, fca, fe and is characterised by a much greater requirement for vernalisation than the WT in order to show early flowering under LD. A further gene, Id, originally identified by Redei (1962) and subsequently also by Lee et al. (1994) also falls into this group.

• Group II: Late flowering - daylength sensitive. This group, which includes fd, ft

Vegetative merister

Reproductive meristem

Vegetative merister

Reproductive meristem


9i fwa fe

FIG. 9.2. Possible causal model for the interaction of genetic and environmental factors controlling the initiation of flowering in Arabidopsis lhaliana. SD = short day treatment, V = vernalization treatment, I = inhibition of flowering, Isd = inhibitor produced by SD, P = flowering promoter. Gene symbols indicate the action of the wild-type allele with —» = promotive effect and —I = inhibitory effect. After Koornneef et al„ 1991.

and fwa flowers later than the WT lines in both SD and LD. They are not very sensitive to vernalisation but still retain a significant response to daylength. • Group III: Late flowering - daylength insensitive. This group includes fha, co (fg) and gi (fb). They flower at about the same time as WT controls in SD but although flowering is accelerated by LD it is much less so than the WT and thus the mutants are late flowering under normal greenhouse, i.e. LD, conditions. They show little response to vernalisation.

Of these mutants, only those in group III can be assumed to have genetic lesions directly in their daylength-response mechanism, i.e. they are potential photoperiod mutants. Two group III mutants, co and gi, have been characterised in some detail. The CO gene has been studied by Coupland's group (Coupland et al. 1993). Plants which have the mutant co allele flower with an increased leaf number under LD but with a similar leaf number to the WT under SD. This indicates that the CO gene encodes a product required in the promotion of flowering by LD. The earliest difference between co and WT can be seen in microscopical sections of the apex of plants grown in LD when the first 2-3 vegetative leaves are longer than 1mm, indicating that CO acts before this time in development. The CO gene has been isolated by Coupland's group and has the characteristic of a transcription factor. What is not known at the time of writing this book is the location in the plant where CO brings about its effects. This is a key question. If CO acts in the leaves, it would be a clear indication that CO is required for the induction process, whereas if it acts at the apex it must be required for the response to an inductive signal. A more detailed physiological evaluation of the GI locus was performed by Araki

TABLE 9.1 Comparison of photoperiodic responses of GI mutant alleles in Co or Ler backgrounds.


Short days

Long days

Continuous light

(8 h)

(16 h)

(24 h)

GI (Co)












GI (Ler)








Flowering response is expressed as the number of rosette leaves at flowering. From data of Araki and Komeda (1993).

Flowering response is expressed as the number of rosette leaves at flowering. From data of Araki and Komeda (1993).

and Komeda (1993). They compared three alleles of the gi mutation, gi-1 and gi-2 in the Co background and gi-3 in Ler (Table 9.1). The Co mutants gi-1 and gi-2 behaved similarly to co, flowering later than WT in LD but at about the same time in SD. The gi-3 allele was delayed in LD but also flowered later than the WT in SD. The flowering time was similar in LD and SD for gi-2 and gi-3, but gi-1 still flowered much earlier in LD than SD. This type of result indicates that the three phenotypic groupings used for the late-flowering mutants need to be interpreted with some care. Some reassignments, especially between groups I and II, may be needed when more alleles of the mutations have been compared. An interesting feature of the gi-2 allele, which showed the strongest phenotype, was that the response diminished at higher temperatures, being much less severe at 28°C than at 22°C. By transferring gi-2 from 28 to 22°C during development in LD a sharp break point was noted in the time taken to flower. Transferring before this time point caused a much greater delay in flowering, indicating that GI normally operates in the WT before that particular developmental stage. The critical period was the time at which the plant had produced between 2 and 4 leaves longer than 0.5 mm, i.e. very similar to the time at which CO is believed to act. This could mean that either the CO and GI genes are acting to make the plants competent to respond to induction by LD, which takes place at the 2-4 leaf stage under continuous LD, or it could mean that one or other of the genes is involved directly in the induction process itself.

Based on the flowering responses of the 11 mutant late-flowering alleles, either alone or in mutant X mutant crosses, Koornneef et al. (1991) proposed a model in which late or delayed flowering in LD was caused by interference with the synthesis or action of promoters or enhancement of the inhibition of promoters (Fig. 9.2). They further proposed that the mutants in which promotive factors are inhibited fail to produce suppressors of the inhibitors, thus explaining the gain of function, such as vernalisation requirement, through a loss of function in the mutants. While this might be important for the remaining loci, which act independently of the daylength response, co, gi and fha can be considered as simple loss of function mutants (loss of the ability to respond to LD) and could equally well be explained by an inability to make or respond to a flowering promoter.

The second class of flowering time mutants are those which flower earlier than WT

to ffl


Ler fun1 fun2 fun3 fun4 fun5 fun6 Genotype

FIG. 9.3. Leaf number at flowering of fun early flowering mutants of Arabidopsis thaliana. Data of Thomas and Mozley (1994).

under SD. These so-called early-flowering mutants have been isolated by at least four laboratories and are described by different nomenclatures. They include tfl (terminal flower) and elf (early flowering; Zagotta et al., 1992), fun (flowering uncoupled; Thomas and Mozley, 1994) and esd (early short day) (Coupland et al., 1993, Koorn-neef, unpublished) mutants. The tfl and elf-1 and elf-2 mutations were isolated as plants which flower early in both LD and SD but retain some photoperiodic sensitivity, whereas elf-3 flowers early in both LD and SD, with the same leaf number in both conditions. The elf mutants all flower earlier and with a lower leaf number than WT in LD and therefore do not appear to represent photoperiodic mutants as such. Although the lf-3 mutant does not show a response to daylength this appears to be because it becomes autonomously induced at a stage before LD induction takes place. The esd series describe four genes and appear to be similar to the elf series, except that they are irradiation induced as compared to the elf mutations which are chemically induced. The fun series of chemically induced mutants were isolated as plants which showed enhanced responsivity to LD inductive treatments or reduced delay by a SD treatment. All of the mutants flowered at about the same time as WT in LD, but earlier in SD (see Fig. 9.3) and in that respect may be distinct from the esd and elf/tfl mutants. The fun mutants fall into two distinct groups. The first, containing fun-1 and fun-2, flower at about the same time as LD controls in both LD and SD. The plants are smaller than controls, with leaves, in particular, being reduced in size. These mutants also flower early when grown in darkness on agar supplemented with sucrose, indicating that flowering is more or less uncoupled from light treatments. The second group, containing fun-3 -fun-6, flower at the same time as WT in LD but are not delayed as much as WT in SD. For fun-3 and fun-4, early flowering in SD occurred but with about the same leaf number as WT. In contrast fun-6 had about half as many leaves as the WT in SD and fun-5 was intermediate both for flowering time and leaf number. Unlike most early flowering mutants there was little difference in visible appearance between fun-6 and WT. The time at which full photoperiodic sensitivity became established was earlier in this mutant than in the WT but the ability to respond to daylength and the critical daylength were not altered. It is suggested that this second group of mutations are related to sensitivity to the photoperiodic stimulus and may represent transduction mutants (Thomas and Mozley, 1994). Most early flowering mutants are recessive, which suggests that they represent components required for the production or action of an inhibitor of flowering under SD.

Triticum aestivum (wheat) is a classic quantitative LDP in which the flowering response is a function of daylength, light quantity and light quality (Carr-Smith et al., 1989). In genetic terms the wheat species cultivated for bread flour is an allohexaploid species in which there are 21 pairs of homologous chromosomes. It can be considered as three separate genomes, A, B and D, each having been derived from a different ancestral diploid species. Because genes are present in triplicate, wheat will tolerate the loss or acquisition of entire chromosomes to produce aneuploid lines (Sears, 1954). These lines provide useful tools for identifying chromosomes which carry genes important for particular processes. Two dominant genes, Ppdl and Ppd2, responsible for a level of insensitivity to daylength or early flowering under SD were first identified by using conventional crosses (Keim et al., 1973; Klaimi and Qualset, 1973). These genes were also detected in subsequent studies in which substituting chromosomes from the homeologous group 2 (chromosomes 2A, 2B and 2C) in the wheat variety Chinese Spring (CS) with those from the wild goat grasses Aegilops umbellulata and Ae. comosa or perennial rye, Secale montanum, produced significant effects on the interaction between the lines and daylength (Law et al., 1978). The Ppdl and Ppd2 genes are located on the short arms of chromosomes 2D and 2B respectively and are probably duplicate genes arising from the polyploid nature of wheat (Scarth and Law, 1983; Worland and Law, 1986). The Ppd genes are dominant for insensitivity to daylength. Increasing the gene dosage leads to early flowering under SD conditions but plants lacking chromosomes carrying Ppd genes flower late in SD. The major effect of the gene appears to be on the rate of development of the ear and spike rather than on the time of floral initiation (Scarth et al., 1985) and this makes it unlikely that Ppd is affecting the induction process. The ability to substitute 'alien' chromosomes, which contain similar loci, from different species such as Aegilops and Secale along with observations on additional lines of barley suggests that the homology of Ppd genes extends to a range of cultivated species.

Pisum sativum: A considerable amount is known about the genetics of flowering in peas based largely on the work of Murfet and co-workers. Pea is a quantitative LDP in which some genotypes also respond to vernalisation. A wide range of genotypes is available and the ability to combine genotypes by grafting enables the site of action of particular genes to be deduced. Also, because grafting experiments give information about the transmission of signals from sensitive to insensitive tissues and vice versa it is possible to interpret the function of certain genes in producing promoters or inhibitors of flowering. A number of genes have been found which influence the onset of flowering, some of which operate through an interaction with daylength


TABLE 9.2 Genes responsible for controlling flowering time in Pisum sativum.


TABLE 9.2 Genes responsible for controlling flowering time in Pisum sativum.





Sterile nodes

Interacts with Dne and Ppd to confer daylength sensitivity


Day neutral

Interacts with Sn and Ppd to confer daylength sensitivity



Interacts with Sn and Dne to confer daylength sensitivity



Reduces expression of the Sn Dne Ppd system during early development


High response

Acts later in development to prolong Sn Dne Ppd activity

Gi (fsd)

Gigas (flowering short days)

Recessive alleles greatly delay flowering


Late flowering

Confers sensitivity to flowering stimuli, four naturally occurring alleles



Required for flowering. Recessive homozygotes remain vegetative



Recessive alleles delay flowering and affect plant size and fertility

Based on Murfet (1990) and Arumingyas and Murfet (1994).

Based on Murfet (1990) and Arumingyas and Murfet (1994).

(Table 9.2). The most important for daylength responses seem to be Sn for sterile nodes (Murfet, 1971a) and Dne for daylength neutral (King and Murfet, 1985). These genes act in combination to modify plant response to photoperiod. The recessive alleles at these two loci give rise to daylength insensitivity whereas the combination of Sn and Dne gives delayed flowering under SD but response to LD and low temperatures is retained. Grafting experiments indicate that the combination of Sn and Dne is needed in the same part of the plant (e.g. leaf or cotyledon) for the production of an inhibitor which is required to suppress flowering under SD (King and Murfet, 1985). Recently, Ppd, a third gene required for the production of the inhibitor, has been identified (Arumingyas and Murfet, 1994). The Sn, Dne and Ppd system for inhibitor production probably acts in leaf tissues and may be switched off during photoperiodic induction. The inhibitor produced by the Sn, Dne and Ppd system may modify the pattern of assimilate distribution in the pea (see Chapter 6) leading to a pleiotrophic phenotype, modified in branching, flower number and life span, rate of flower and fruit development, maturity date and yield (Murfet, 1990). Two additional modifying genes, E (early) and Hr (high response) interact with the Sn Dne homozygotes to promote earliness or delay flowering further, respectively, under SD conditions. As suggested by their names, E operates in the cotyledons to reduce Sn Dne activity in the early stages of seedling growth but HR acts later in the life cycle to prolong the duration of Sn and Dne activity.

Although early grafting experiments suggested the presence of a transmissible flowering promoter in peas (Murfet, 1971b) it is only recently that genetic evidence for a gene acting in the floral stimulus pathway has been found. The recessive allele gi (gigas) was originally obtained from the cultivar Virtus, where it causes plants to flower much later than the initial line (Murfet, 1990). Another recessive mutant, fsd


y WT

FIG. 9.4. The effect of number of WT (cv Porta) donor leaves on flowering in mutant (M) gif scions in an 18 h photoperiod. Grafts were made epicotyl to epicotyl of 7-day-old seedlings. The WT side shoots arose from the cotyledonary node; the number of leaves present is indicated next to the lateral shoot. Lateral shoots with 1, 2 or 4 leaves were obtained by decapitating the shoot above the specified leaf. The lateral shoots in the treatment on the far right were not decapitated and they produced a mean of 16 leaves. F = flowering, V = vegetative; the percentage of plants flowering is given above. After Taylor and Murfet 1994.

(flowering short days) shows obligate SD flowering behaviour and remains vegetative under LD (Taylor and Murfet, 1994). Allelism tests suggest that fsd is a more severe allele at the Gi locus and a new symbol gifsd has been proposed for the fsd locus. The delay in flowering in gi is prevented when the mutant is grafted to the WT stock. Similarly, grafting of gifsd scions on to WT stocks also overcomes the inability to flower in LD. The WT stock is effective providing it carries at least two foliage leaves (Fig. 9.4). This result is consistent with the WT shoots providing a transmissible substance necessary for flowering, i.e. a floral promoter, which is missing in the mutant shoots. The gi or gifsd phenotypes have many similarities with the veg (vegetative) mutants which do not flower under any environmental or genetic circumstances. Grafting evidence in this case indicates that veg acts at the apex and blocks some aspect of flower initiation (Reid and Murfet 1984). The late flowering phenotype of Gi mutants and its alleles is similar to the late flowering phenotype of co or gi in Arabidopsis, although the site of action of the latter two genes is not as yet known. A further gene, Lf for late flowering has been identified and controls the response to both vernalisation and LD. A number of alleles at this locus have been found and these are characterised by different minimum node number at which flowers are formed. In this case it is thought that the gene acts at the apex and establishes the sensitivity to flowering signals (Murfet, 1971b).

Hordeum vulgare L. (Barley): At least four recessive ea (early maturity) loci have been found to influence flowering time in barley. Of the recessive homozygotes, easp confers the earliest flowering phenotype while eak suppresses the other ea loci. A

dominant enhancing allele interacts with the eak homozygous recessive to give the earliest flowering phenotype ea*/EN. An apparently independently isolated flowering genotype of barley (BMDR-1) is allelic to the eak!EN genotype and is characterised by complete photoperiod insensitivity but retains FR mediated promotion of flowering (Principe et al., 1992). A biochemical comparison of the BMDR-1 and the corresponding WT showed differences in two peptides when protein extracts were separated by two-dimensional polyacrylamide gel electrophoresis. However, neither appeared to be pysiologically regulated in a manner which correlated with the phenotypic photoperiodic behaviour of the mutant alleles and they may not be important in establishing photoperiod sensitivity.

Soyabean: Almost all of the detailed analysis of flowering time genes in relation to photoperiodic mechanisms is based on the LDP described above. For SDP, genetic analysis has largely been limited to descriptive genetics in relation to breeding requirements. An example of this sort is soyabean (Glycine max) of which there has been extensive analysis. Four genes (Et E2, £3 and £4) which, as recessive alleles confer relative photoperiod insensitivity, have been identified (Palmer and Kilen 1987). Of these, the £3 locus appears to be very important in establishing photoperiod differentials. The dominant £3 gene causes a strong delay by long photoperiods and the recessive e3 makes plants insensitive to photoperiod and causes early flowering. The e4 locus also causes insensitivity to long photoperiods (Buzzell and Voldeng, 1980).

Was this article helpful?

0 0

Post a comment