The idea of florigen, a substance which is generated by leaves and is supposed uniquely to evoke flowering at the shoot apex, underlies many considerations of the biochemistry of floral induction and evocation. Much of the experimental evidence in support of such a floral stimulus comes from physiological experiments on plants where flowering is controlled by exposure to appropriate daylengths but the idea has been extended to include plants where flowering is induced by other means. Thus, the original and still sometimes accepted concept of florigen is of a unique chemical substance which evokes flowering in a wide range of species and in all photoperiodic categories. However, florigen continues to remain a physiological concept rather than a chemical entity. Its elusive nature, together with the many effects of naturally-occurring growth regulators has led several workers to question whether it is a single substance (Cleland, 1978; Evans, 1971; Bernier, 1986). Certainly, other morpho-genetic processes, such as the initiation of roots and buds on callus, require the interaction of two or more hormones and, as we have seen, there is evidence that the auxin/cytokinin ratio can control the switch between vegetative and floral buds in florally induced tobacco tissue. Similarly, floral evocation at vegetative shoot meristems may depend on a particular ratio and/or temporal sequence of substances which are not themselves unique for flowering.
Despite the failure to characterise florigen, there is much circumstantial evidence for the existence of a transmissible substance, which is needed for floral evocation in a wide range of different plants. There are several reasons which might account for the failure, so far, to isolate and characterise such a substance; these include lack of suitable assay method, use of the wrong extraction solvents, difficulties of reintroduction into the plant at the right site and in the right quantity, lability, low concentration in the extracts, the presence of inhibitors from non-induced leaves of test plants or rapid breakdown in such leaves, and failure of phloem loading in leaves in non-inductive daylengths. In most attempts to extract a floral stimulus it has been assumed that it is a simple low-molecular-weight compound; however, preliminary attempts to look for larger compounds (such as soluble proteins and polypeptides) have also met with little success. A further possibility is that action of the floral hormone requires the presence of a complementary stimulus, such as a giberellin or cytokinin.
Chailakhyan's original florigen concept (1936) was replaced in 1958 by a hypothesis in which he proposed that flowering is regulated by a stimulus consisting of two components, namely gibberellins and a hypothetical substance, or group of substances, called anthesins (Chailakhyan, 1982). It was supposed that LDP produce anthesins in all daylengths but GAs only in LD, while SDP produce sufficient GAs in any photoperiod but anthesins only in SD; finally, the formation of florigen in the shoot apex requires both GA and anthesin. This scheme could also account for plants with a dual daylength requirement, if it assumed that LD are needed for GA production and SD for anthesin production, although it does not explain why the obligatory sequence should be different in LSDP and SLDP. Moreover, GA has been shown to substitute for LD in the LSDP Bryophyllum and for SD in the SLDP Scabiosa succisa (see Table 7.5), while in the SLDP Bromus and Poa, GA prevents the primary induction of flowering in SD (Heide et al., 1986a). It is also evident that interaction at the apex cannot occur in the LSDP Cestrum nocturnum since LD and SD must be given to the same leaf (Sachs, 1985), although this is not the case in Bryophyllum (see below). The idea that GA limits the flowering of LDP in SD, while anthesin limits the flowering of SDP in LD, requires that donors of one photoperiodic response type would evoke flowering in receptors of the other type, even when the donor itself is in non-inductive daylengths. However, in the majority of cases that have been examined, donors could only evoke flowering in receptors of other photoperiodic classes when they themselves were induced. There appear to be only two reported exceptions (Lang, 1965; Zeevaart, 1958). Some flowering response was obtained in Maryland Mammoth receptors (SDP maintained in LD) when grafted to vegetative shoots of N. sylvestris (LDP maintained in SD), which is surprising in view of the strong evidence for antiflorigens from non-induced leaves of N. sylvestris (see Table 6.15). In the LSDP Bryophyllum, flowers formed on receptor shoots situated between vegetative parts of the graft which were in SD and LD respectively. It was suggested that, in this case, flowering was evoked by GAs from LD- and anthesins from SD-treated leaves (Chailakhyan, 1975) but another and more likely explanation is that GA from LD leaves moved into SD leaves, thereby resulting in the production of florigen (Zeevaart, 1976). Thus the anthesin/GA hypothesis does not appear to be generally applicable to different photoperiodic response groups and, in any event, there is no information on the chemical identity of the putative anthesins. Even so, the possibility that one or more additional substances may be needed for the synthesis, transport, or action of florigen is not excluded. An appropriate supply of hormones and nutritional factors is necessary for the initiation, differentiation and growth of floral parts and, in many species, these further stages of floral development are also under photoperiodic control (see Chapter 10).
Several hypotheses for the control of floral evocation have been put forward during the past 70 years. These include control by a single specific flower-promoting hormone, a relatively simple stoichiometric interaction between two substances, florigen and antiflorigen, a complex multicomponent stimulus with different factors being limiting in different conditions, and a nutrient-diversion theory that assumes only a secondary role for hormones. If there is a specific florigen, its identity and even the class of substance remains a mystery. The idea of a multicomponent stimulus is attractive and it has, for example, been suggested that the floral stimulus may consist of a mixture of known growth regulators. However, there is no evidence that any of the known plant growth substances, or any combination of them, can replace the floral stimulus and evoke flowering in a wide range of plants in all photoperiodic categories. Even in Lemna, where flowering can readily be modified by several growth regulators, the endogenous levels were not correlated with flowering and it was concluded that changes in hormone levels were of secondary importance in the photoperiodic control of flowering (Fujioka et al., 1986b). No plant growth regulator was able to substitute for photoperiodic induction in test plants of the qualitative SDP Chenopodium rubrum\ their action on flowering was strictly stage-dependent and rather non-specific growth effects (such as inhibition of cell division in axillary meristems by auxin and enhancement of cell division by CK) resulted in stimulation or inhibition of flowering depending on the stage of floral transition when they were applied (Ullman et al., 1985). In contrast, extracts from induced tobacco plants were able to cause flowering in strictly non-inductive conditions, suggesting the participation of an unknown compound. This does not mean that growth substances do not play a role in the overall flowering process, but indicates that they are probably non-limiting for flowering in most plants. Nevertheless, many are under photoperiodic control and this, together with their numerous effects on flowering, makes it likely that they are at least modifying factors for floral evocation and expression. All classes of known growth regulators or their precursors are transported over long distances and thus could, in theory, be part of a transmissible signal exported from the leaf.
Studies of the changes that occur at the shoot apex in response to the arrival of the floral stimulus following photoperiodic induction have indicated that the initiation and growth of floral primordia is probably controlled by a complex system of interacting factors including, among others, gibberellins and cytokinins (Bernier, 1986). Moreover, the different factors may act in sequence since some of the changes seem to occur before the floral stimulus arrives at the apex. For example, in the SDP Pharbitis, a first wave of enhanced uridine incorporation into RNA at the shoot apex (which was 2-fold greater than in the LD controls) occurred even before the end of the critical dark period (Vince-Prue and Gressel, 1985). An even more extreme situation is seen in Sinapis, where bark-ringing studies indicated that a leaf signal was exported to the roots before the 12th hour of an LD and an increase in cytokinins was detected in root exudates at the 9th hour (i.e.,within 1 h of extending the 8 h SD, Bernier et al., 1993). The role of these early increases is not immediately evident since they occurred under conditions that are not inductive for flowering. In some cases, changes normally associated with flowering can be caused by treatments which do not themselves bring about floral initiation. For example, in Sinapis alba, one of the earliest events at the shoot apex following photoperiodic induction is an increase in mitotic activity and any treatment which abolishes this also abolishes flowering. The application of CK to the shoot apex resulted in an increase in mitotic activity similar to that caused by exposure to a single LD. However, the long day induces flowering while the cytokinin treatment does not. These results have led to the proposal that cytokinin is a component of the floral stimulus, although it is insufficient alone to evoke flowering (Havelange et al., 1986). If the floral stimulus does have more than one component (acting either sequentially or together) then, in a given species, all factors will not necessarily be absent in conditions that do not allow flowering. This could explain the wide range of substances and conditions that can cause flowering in different plants. Bernier (1986) has emphasised that floral induction is achieved by different sets of environmental factors, often in the same plant, and proposed that the idea of a specific leaf-generated stimulus is essentially erroneous in the general context of floral induction. However, the existence of a leaf-generated signal in plants where induction is achieved through exposure to the appropriate photoperiod can hardly be questioned, and the identity of this signal is presently unknown. Moreover, it is evident from grafting experiments that, at least in some cases, the resultant transmissible product from leaves is the same stimulus (or is functionally equivalent), despite the fact that induction may be achieved via different means (Table 6.10). There is, moreover, no a priori reason to assume that a floral stimulus is necessarily complex. Although flowering itself is a complex morphogenetic process and is likely to involve a number of regulatory substances, the role of the floral stimulus may only be to control the time at which it begins and the rate at which it progresses; thus, a simple molecule might be sufficient.
The best evidence for a multifactorial stimulus for flowering comes from detailed studies of photoperiodic induction in the quantitative LDP Sinapis alba, which is induced to flower by exposure to a single LD. Increases (often transient) in a number of different components exported from the leaf and/or reaching the apical bud have been detected during, or immediately following induction. These include sugars, cytokinins, polyamines and calcium ions, while auxin levels in the apical bud have been shown to decline (Bernier et al., 1993). Given that the application of CK and the elevation of sucrose levels by high-intensity light can bring about a number of the changes in the shoot apical meristem that are normally associated with the LD induction of flowering, it is reasonable to conclude that CK and sugars are components of a multifactorial signal for flowering in this LDP. However, neither sugars nor CKs (nor both together) can evoke flowering in SD; moreover GA is not florigenic for Sinapis in SD, nor is any other substance tested. Thus, even if the transition to flowering depends on several regulatory factors, an essential component (florigen?) of the daylength-dependent stimulus generated by leaves remains, as yet, unidentified.
One of the earliest theories for the control of flowering was that it depended on the carbon:nitrogen ratio, flowering being favoured by a high C/N ratio and vegetative growth by a high N/C ratio. The idea is generally attributed to a paper by Kraus and Kraybill in 1918, although their work did not specifically address the question (Cameron and Dennis, 1986). Debate on the role of carbohydrates in the initiation and development of flowers has continued since the early part of the century and several 'nutritional' hypotheses of different kinds have been put forward. For example, in chrysanthemum, the initiation of floral primordia appears to occur when the apex has reached a critical size (Horridge and Cockshull, 1979) and so might be a relatively simple consequence of apex enlargement which, in turn, could arise from the action of a number of factors including nutritional ones. The correlation of apex size with flowering is not invariable, however, and it is evident that apex enlargement alone is not an adequate explanation in many plants; in some cases, the apex may actually decrease in size during floral initiation while, in Chenopodium rubrum, older plants with larger apices are less sensitive to SD induction than younger ones (Seidlova and Opatrna, 1978). In excised apices from non-induced Pharbitis plants, however, floral bud formation was promoted by a high concentration of sucrose and also by a low nitrogen concentration (Ishioka et al., 1991b) indicating that nutritional conditions can affect the morphogenetic pathway taken by the apex.
A related suggestion has been that the transition to flowering is a consequence of an enhanced flow of nutritional factors to the apex. A specific nutrient-diversion hypothesis is that photoperiod controls flowering by altering the pattern of nutrient movement such that, in favourable daylengths, nutrients are diverted to the shoot apex where floral evocation results (Sachs, 1977, 1987). In this context, it is evident that some of the effects of growth regulator application result from enhanced transport of substances to the apex, or of a restriction of transport to the apex; for example, the promotion or inhibition of flowering in Pharbitis by CK application to non-induced or induced cotyledons respectively (Ogawa and King, 1979b). Moreover, daylength has been found to influence the partitioning of assimilates in both LDP and SDP (see Chapter 13). Nevertheless, it is evident from the results of many different kinds of experiment that nutritional factors alone cannot account for the daylength-dependent switch to flowering. As discussed earlier in this chapter, studies of the effect of photoperiod on the carbohydrate content of leaf exudates have demonstrated that, although an increased supply of carbohydrate may be crucial for floral development, it is not alone sufficient to trigger the transition from vegetative growth to flowering (Houssa et al, 1991; Lejeune et al., 1991). Furthermore, although the shortening of phases of the cell cycle at the apex of LD-induced plants of Sinapis was better mimicked by high-intensity light together with CK application than by either treatment alone, plants remained vegetative (Bernier et al., 1993). In another approach, it was shown that GA3 and GA4/7 have equal effects to mobilise assimilates to potential cone buds but only the latter promoted flowering (Pharis et al., 1989). Moreover, many of the daylength-dependent changes in carbohydrate distribution and concentration observed in normal clover plants also occur in mutants which are not initiating floral development (Jones, 1990). Any nutritional hypothesis also needs to explain why, in Xanthium, a single leaf in SD results in flowering whereas all the LD leaves have no effect. Similarly, it is difficult to understand how assimilate supply can account for the fact that a single leaf of Nicotiana sylvestris in LD is sufficient to cause flowering in Maryland Mammoth tobacco, whereas all the leaves on the Maryland Mammoth plant (more than 20 in some cases) are ineffective. Such results make a simple nutritional hypothesis unlikely although, once evocation has occurred, the developing flower is a strong sink and nutrients are diverted to it.
Attention has been drawn to the fact that, unlike the situation in animals, developmental processes in plants do not seem to be strictly dependent on the presence or absence of highly specific molecules and it has been argued that there is no reason to suppose that the transition to flowering is different. In support of this point, the effect of auxin/cytokinin ratios to switch the developmental pathways in tobacco TLCs between the production of vegetative and floral buds can be cited (Mohnen et al., 1990). Perhaps one of the most persuasive arguments for a relatively non-specific stimulus is the enormous range of responses, in addition to floral initiation, that may be under daylength control (see Part II), indicating that specificity may be a function of the target site rather than of the leaf-generated stimulus itself. Nevertheless, despite these arguments, it is clear from many physiological experiments that the stimulus exported from a photoperiodically induced leaf differs in some as yet unidentified way from that exported from a non-induced leaf and that this difference is related to the switch between vegetative and reproductive growth at the shoot apex. It is also evident that the stimulus from an induced leaf is effective in evoking flowering in plants of all photoperiodic categories. However, the failure to isolate either florigen or anti-florigen has led to the concept that floral evocation may depend on a complex stimulus such as a particular ratio and/or temporal sequence of endogenous growth regulators, inhibitors and other factors not specifically associated with flowering, rather than on a specific floral stimulus and/or inhibitor (Bernier et al., 1993). Moreover, many of these have been shown to be regulated by daylength (e.g., gibberellins, cytokinins, carbohydrates, calcium ions) and so could be constituents of a leaf-generated photoperiodic signal. The possibility that some, at least, of these factors act by sensitising the apex to small amounts of a floral stimulus should not be discounted, expecially as the effects are often seen only under marginal conditions for flowering. Unfortunately, 21 years on from the first edition of this book and despite a great deal of research effort, the identity of the daylength-dependent stimulus which is graft-transmissible and evokes flowering in plants of different photoperiodic categories is still unknown; not has its specificity for the flowering process yet been resolved.
Was this article helpful?