Stress-induced flowering: the third category of flowering response

Abstract

The switch from vegetative growth to reproductive growth, i.e. flowering, is the critical event in a plant’s life. Flowering is regulated either autonomously or by environmental factors; photoperiodic flowering, which is regulated by the duration of the day and night periods, and vernalization, which is regulated by low temperature, have been well studied. Additionally, it has become clear that stress also regulates flowering. Diverse stress factors can induce or accelerate flowering, or inhibit or delay it, in a wide range of plant species. This article focuses on the positive regulation of flowering via stress, i.e. the induction or acceleration of flowering in response to stress that is known as stress-induced flowering – a new category of flowering response. This review aims to clarify the concept of stress-induced flowering and to summarize the full range of characteristics of stress-induced flowering from a predominately physiological perspective.

Abiotic stress, Arabidopsis, biotic stress, flowering, FLOWERING LOCUS T, pharbitis, salicylic acid.

Introduction

Plants switch to reproductive growth after a certain period of vegetative growth. The switch from vegetative to reproductive growth – flowering – is critical to plant development, because the proper timing of flowering ensures the success of the next generation and the continuity of the species. Flowering is regulated autonomously or by environmental factors. The primary environmental factors are the duration of the day and night periods, which regulate photoperiodic flowering, and low temperatures near the freezing point, which regulate vernalization. Photoperiodic flowering and vernalization have been well studied, and their regulatory mechanisms are well elucidated (Thomas and Vince-Prue, 1997; Bernier and Périlleux, 2005; Sung and Amasino, 2005; Turnbull, 2011; Romera-Branchat et al., 2014; Song et al., 2015). Flowering that cannot be classified as either photoperiodic flowering or vernalization has also been occasionally reported. Some plants in which flowering is largely regulated by photoperiod can flower under unsuitable photoperiods when grown under particular conditions. Whilst experienced flowering physiologists had noted that plants tend to flower when grown under unsuitable conditions, such atypical flowering was not studied systematically, presumably because it was considered an exceptional phenomenon that did not add to an understanding of the fundamental mechanisms of flowering. In recent years, however, surveys of the literature reporting these responses have found that the common factor responsible for such flowering could be regarded as stress (Wada and Takeno, 2010; Takeno, 2012), and several plant species have recently been reported to flower when different types of stress are applied. Thus, the evidence that stress is a flower-inducing factor is accumulating, and stress-induced flowering has recently received increased attention (Blanvillain et al., 2011; Yaish et al., 2011; Pieterse, 2012; Riboni et al., 2014; Kazan and Lyons, 2016). Stress-induced flowering can be now considered as the third category of flowering response in addition to photoperiodic flowering and vernalization.

Plants inhibit the processes of growth and development when they are stressed but they have evolved the ability to resist, endure, avoid, or adapt to stress such that they can protect themselves from harmful environmental conditions. Investment in these abilities by plants can moderate the reduction in yield of agricultural crops that occurs under stressful conditions. Plant stress physiology has developed from this perspective, and much knowledge has been accumulated (Qin et al., 2011; Rasool et al. 2015; Xia et al., 2015; Shabala et al., 2016). Stress-induced flowering is one response to stress and is the ultimate stress adaptation, because plants can survive as a species if they flower and produce seeds even when they cannot survive as individuals under severe stress. Despite this significance, stress-induced flowering is not the primary focus of plant stress physiology. Stress can be simply defined as a situation in which the processes of growth and development are inhibited (Lichtenthaler, 1998), and therefore stress may also inhibit the flowering process. Indeed, the inhibition of flowering by stress is not surprising and has been documented in many cases, as reviewed by Kazan and Lyons (2016). In contrast, the reverse response of induction or promotion of flowering by stress appears to be a paradox and may conceal some biological significance. Accordingly, this review focuses on the positive cases of regulation of flowering by stress, and specifically it examines their physiological aspects in order to clarify the concept of stress-induced flowering.

Generalities of stress-induced flowering

Plant species that can be induced to flower in response to stress

Nearly twenty plant species belonging to a wide range of taxonomic groups have been reported to flower in response to stress in previous reviews (Wada and Takeno, 2010; Takeno, 2012). Thus, stress-induced flowering is not an exceptional phenomenon in a restricted set of plant species, but is generally conserved in angiosperms.

The response to stress differs depending on the species or cultivar. In pharbitis (Ipomoea nil, synonym Pharbitis nil) the cultivar Violet flowers respond to poor-nutrition, low-temperature, and high-irradiance stresses (Swe et al., 1985). However, cultivars Tendan and Kidachi are not induced to flower by exposure to poor nutrition or high-intensity light. Conversely, Tendan and Kidachi are more sensitive to low temperature than Violet (Ishimaru et al., 1996). Tendan is close to the original species, Violet is a domesticated form, and Kidachi is a dwarf mutant unable to synthesize gibberellin (GA). The flowering response to poor nutrition in Arabidopsis (Arabidopsis thaliana) is stronger in Landsberg erecta (Ler) than in Columbia (Col) accessions (Kolář and Senková, 2008). Among three tested strains of Lemna paucicostata (synonym Lemna aequinoctialis), strain 6746 is the most sensitive to poor nutrition (Shimakawa et al., 2102). The red-leaved form of Perilla frutescens var. crispa is sensitive to low-intensity light, but the flowering response of the green-leaved form is quite weak (Wada et al., 2010a). These plants, which share an almost identical genetic background but respond differently to stress, could be valuable for investigations into the regulatory mechanisms of stress-induced flowering.

Stress-induced flowering has been well studied using pharbitis, which is a qualitative short-day (SD) plant (Wada et al., 2013). This plant never flowers under long-day (LD) conditions unless the appropriate stresses are applied, and therefore it provides a suitable subject for the study of stress-induced flowering. In contrast, quantitative SD or LD plants flower even under non-inductive photoperiodic conditions, and therefore flowering responses to stresses applied under such conditions should be analyzed carefully. Furthermore, care must be taken when evaluating rosette plants, in which the number of rosette leaves is often used as an index of the flowering response. When rosette leaf formation itself is suppressed by stress, it has been concluded that flowering is accelerated even though the flowering process is not actually promoted. Bernier (2003) warned that rosette leaf number cannot be regarded as a critical determinant of the floral shift, and conclusions based on actual flowering time should be of prime importance. For example, pathogen infection does not affect the number of rosette leaves, but accelerates the flowering time in Arabidopsis (Korves and Bergelson, 2003). In this case, it could be concluded that the pathogen infection accelerates flowering.

Stress factors that can induce flowering

Many types of stress factors have been reported to induce flowering. These include high or low light intensity, UV light, high or low temperature, poor nutrition, nitrogen deficiency, drought, low oxygen, crowding, root removal, and mechanical stimulation, as summarized in previous reviews (Wada and Takeno, 2010; Takeno, 2012; Kazan and Lyons, 2016). Thus, stress generally functions to induce flowering. Although a clear definition of plant stress is lacking (Kranner et al., 2010), stress in this review is simply defined as conditions that suppress the growth or development of plants, following the definition of Lichtenthaler (1998). Many papers reporting the effects of stress consider poor nutrition, high or low temperature, high salinity and other conditions as a priori stress factors without providing evidence that these factors actually suppress growth or development of the tested plants. Such papers are considered in this review, but we should be careful to ensure that the plants in question are actually stressed. Because the rate of every biochemical reaction depends on it, ambient temperature – the physiological and non-stressful temperature range – also accelerates or delays flowering time (Capovilla et al., 2014). A flowering response accelerated by ambient temperature is not included in the category of stress-induced flowering covered in this review. On the other hand, flowering induced by exposure to 16–18 ºC in pharbitis is considered low-temperature stress-induced flowering because the vegetative growth of the plant is inhibited within this temperature range (Hatayama and Takeno, 2003).

Stress factors induce flowering in diverse plant species, and plant species flower in response to diverse stress factors. However, not all stress factors induce flowering in all plant species. Pharbitis flowers in response to poor nutrition, low temperature or high-intensity light stresses (Shinozaki et al., 1982; Swe et al., 1985; Hirai et al., 1993, 1994). Poor nutrition also induces flowering in L. paucicostata (Shimakawa et al., 2012) and accelerates the flowering of Arabidopsis (Kolář and Seňková, 2008), but it does not induce flowering in P. frutescens (Wada et al., 2010a). Drought stress promotes flowering under LD conditions in Arabidopsis (Riboni et al., 2013). Drought stress also induces flowering in Citrus latifolia (Southwick and Davenport, 1986), and the same basic flowering mechanism has been suggested to be shared by all citrus species (Tamim et al., 1996) and is reported in tropical and subtropical species. The perennial woody tree Sapium sebiferum blooms 3–5 years after germination, but 1-year-old seedlings bloom precociously when exposed to drought stress (Yang et al., 2015). In contrast, the effect of salt stress on flowering is largely unknown, but it has been reported that salt stress delays flowering in Arabidopsis (Kim et al., 2007; Ryu et al., 2013). Plants may have evolved to adapt to stresses, particularly those that frequently fluctuate in the natural environment and result in serious damage to plants. Poor nutrition, drought, and salinity are examples of such stress factors. This could lead to the conclusion that poor nutrition and drought stress induce flowering in many plant species. However, it is interesting to note that salt stress rarely induces flowering, although it is considered one of the most important stress factors affecting crop yield. Tropical and subtropical species may have evolved a mechanism to respond to the seasonal change in water availability, which can be used as an accurate calendar in place of changes in day length and temperature in those regions. UV-C light stress induces early flowering in Arabidopsis (Martínez et al., 2004). However, the ecological significance of UV-C light stress-induced flowering may be limited because only a small portion of UV-C radiation reaches the earth’s surface, and its influence on plants may be minimal under natural conditions.

Biotic stresses also promote flowering, but the details of how this occurs are unclear. Infection with the bacterial pathogens Pseudomonas syringae and Xanthomonas campestris, the oomycete Peronospora parasitica, and the root-infecting fungal pathogen Fusarium oxysporum accelerate flowering in Arabidopsis (Korves and Bergelson, 2003; Lyons et al., 2015). Arabidopsis plants transformed with the parasitism gene Mi8D05, which is critical for success of the parasitic root-knot nematode Meloidogyne incognita, show accelerated flowering (Xue et al., 2013), although it is uncertain whether the infection by the nematode itself induces early flowering. It is suggested that tobacco (Nicotiana tabacum) plants inoculated with tobacco mosaic virus tend to flower early (León, 2005), but the details have not been published. Plant growth-promoting rhizobacteria (PGPR) colonizing the rhizosphere or plant tissues generally promote plant growth, but the inoculation of Arabidopsis plants with the PGPR Burkholderia phytofirmans also accelerates flowering (Poupin et al., 2013). This acceleration of flowering should be distinguished from the stress-induced flowering caused by pathogens, because the flowering caused by PGPR is accompanied by promotion of plant growth and improvemed tolerance to abiotic stresses. Flowering induced by herbivory has not been reported. Flowering in response to crowding in Lemna perpusilla (Landolt, 1957) could possibly be categorized as flowering in response to biotic stress, but it could be due to abiotic stress from nutrient or oxygen deficiency as a result of competition between individuals, or a physical stimulus resulting from contact with neighboring plants. The influence of competition for light on flowering can be considered as flowering induced by low-intensity light stress, as discussed below.

Biological significance of stress-induced flowering

Plant stress physiology demonstrates that plants can survive stressful environmental conditions as individuals and also as a species via selection for adaptability or tolerance to stresses. Stress-induced flowering can be considered the ultimate adaptability to stress, because flowering results in the reproduction that ensures survival as a species. To test this interpretation, it is necessary to examine whether the plants in which flowering is induced by stress can produce fertile seeds. Pharbitis grown under conditions of poor nutrition throughout the life of the plant show induced flowering, achieve anthesis, and produce fruits and seeds (Wada et al., 2010b). The seeds germinate and the progeny develop normally. The third generation of the stressed plants also develops normally. Red-leaved P. frutescens grown under low-intensity light flowers and forms seeds (Wada et al., 2010a). The progeny grow normally and are induced to flower in response to SD treatments. Nutrient-deficient Arabidopsis Ler plants form normal flowers and fruits with seeds (Kolář and Senková, 2008). Arabidopsis in which flowering is accelerated by P. parasitica infection produces fruits (Korves and Bergelson, 2003). The stressed plants do not require the onset of a season when photoperiodic conditions are suitable for flowering, and such precocious flowering and seed production may assist in the preservation of the species. Thus, stressed plants can produce the next generation as an emergency response even when they themselves cannot adapt to unfavorable environmental conditions. Stress-induced flowering has a biological benefit, and therefore this response should be considered as important as photoperiodic flowering and vernalization.

The stress adaptation extends to the evolution of flowering traits. The influence of drought on the flowering traits of subsequent generations was studied in Brassica rapa by comparing the flowering time of plants produced from seeds collected before and after a natural drought (Franks, 2011). The post-drought lines flowered earlier than the pre-drought lines, showing an evolutionary shift towards earlier flowering. Drought-induced flowering in the woody tree S. sebiferum provides a possible means to shorten vegetative growth, which could be useful in genetic research and breeding (Yang et al., 2015).

Comparison between stress-induced flowering and other flowering responses

Stress treatment is required for a longer period than photoperiodic treatment to induce flowering. Pharbitis requires poor nutritional conditions for 20 d or low-temperature stress for 10 d to induce flowering, whereas only one SD treatment is required for photoperiodic flowering (Hatayama and Takeno, 2003; Wada et al., 2010b). Perilla frutescens requires low-intensity light for 3 weeks, whereas 5 d of SD treatment is sufficient for photoperiodic flowering (Wada et al., 2010a). It is known that the low-temperature time requirement for vernalization is generally longer than the SD or LD treatment for photoperiodic flowering, probably because the seasonal change in temperature on which vernalization depends is unreliable, whereas the day/night cycle on which photoperiodic flowering depends is quite stable and is thus an accurate indicator of season. Similarly, the change in the strength of stresses is not reliable, and therefore plants may require a sufficiently long period to ensure the correct response of to flower or not to flower. A strong stress treatment induces 100% flowering in pharbitis and P. frutescens, as does the SD treatment, whereas the flowering response to the stress treatment is weaker than the response to the SD treatment in L. paucicostata.

The stress response is dependent upon photoperiodic cues in certain cases. Drought stress induces early flowering under LD conditions and delays flowering under SD conditions in Col-0 (but not in Ler) of Arabidopsis (Riboni et al., 2013). Expression of the flowering gene FLOWERING LOCUS T (FT) is involved in both stress-induced flowering and photoperiodic flowering, as discussed below. The photoperiodic pathway and the stress-induced pathway achieve the same goal of flowering, but the former may be the primary pathway under normal conditions and the latter may be a secondary pathway used in an emergency. Because of the close link between the two responses, knowledge gained from the study of the regulatory mechanism of photoperiodic flowering could be useful for the analysis of stress-induced flowering.

Perilla frutescens is induced to flower by low-intensity light stress (Wada et al., 2010a), which resembles a shade-avoidance response that is observed in plants shaded by neighbors. The major phenotypes produced by the shade-avoidance response are etiolation and rapid stem elongation, and early flowering is also a part of this syndrome (Smith and Whitelam, 1997; Franklin, 2008). The low-intensity light stress-induced flowering in P. frutescens appears different from the shade-avoidance response because it is not accompanied by etiolation and rapid stem elongation (Miki et al., 2015). In Arabidopsis plants with specific genetic backgrounds, however, early flowering can occur in the absence of other responses such as rapid stem elongation, and the other (vegetative) responses occur without flowering under shaded conditions (Chincinska et al., 2008; Adams et al., 2009). Early flowering is not necessarily associated with these other responses. Therefore, low-intensity light stress-induced flowering may essentially be one component of the shade avoidance syndrome.

Regulation of stress-induced flowering

Perception of stress factors

The analysis of perception of external stimuli could provide a means to understand the signaling cascade in developmental processes regulated by environmental conditions. However, the mechanism by which stress factors are perceived in stress-induced flowering is uncertain, whereas the light and dark signals in photoperiodic flowering and the low temperature involved in vernalization are perceived by leaves and shoot apices, respectively. Although the presence of cotyledons is essential for stress-induced flowering in pharbitis (Shinozaki and Takimoto, 1982; Shinozaki, 1985), this does not necessarily mean that the cotyledon is the stress-perceiving organ. Information on the sensors that detect abiotic stress factors is very limited (Hirayama and Shinozaki, 2010), although work on simpler organisms such as Synechocystis and Bacilus subtilis indicates that two-component transmembrane systems have a potential role in transducing abiotic stress, and it has been postulated that cold is sensed when membrane fluidity changes (Yang et al., 2015).

It cannot be assumed that there are receptors for stress factors similar to the photoreceptor in photomorphogenesis or the statolith in geotropism. In the case of low-intensity light stress-induced flowering, however, photoreceptive pigments may be involved, as in the shade-avoidance response. The shade-avoidance response is triggered when phytochrome B in the leaves detects the reduced ratio of red-to-far-red wavelengths in the light passed through the green leaves of nearby plants (Franklin, 2008). The shade-avoidance response is also induced by green light, which is detected by cryptochrome (Keller et al., 2011). Low-intensity light stress-induced flowering in the red-leaved P. frutescens is suggested to involve the influence of green light (Miki et al., 2015). The leaves of the red-leaved P. frutescens contain anthocyanins at high concentration under normal light conditions, and the concentration decreases under flower-inducing, low-intensity light conditions. The high-concentration anthocyanins may play the role of a red-colored optical filter under normal light conditions, and this filtering effect may be lost under low-intensity light, increasing the relative quantum of light in the green wavelengths. This change may induce a photobiological effect leading to flowering. This hypothesis is supported by the report that cryptochrome, which is inactivated by green light, is involved in the shade-avoidance response (Zhang et al., 2011; Wang et al., 2013).

Regulation by plant hormones

Defoliated scions grafted onto root stocks with cotyledons are induced to flower in response to poor nutritional conditions in pharbitis (Wada and Takeno, 2010). This suggests that a transmissible flowering stimulus such as florigen, which is involved in photoperiodic flowering, is also involved in stress-induced flowering. It was expected that cultivar Violet would not flower when grafted onto cultivar Tendan, because Tendan does not flower in response to poor nutrition. However, defoliated Violet scions grafted onto Tendan rootstocks with cotyledons were induced to flower under conditions of poor nutrition. Conversely, Tendan scions grafted onto Violet rootstocks were not induced to flower. Tendan may produce a transmissible flowering stimulus but may not respond to it. Analysis of behavior such as that exhibited by Tendan may be useful for understanding the nature of the flowering stimulus.

It is uncertain whether the transmissible flowering stimulus suggested above is the same as the FT protein that is known as florigen in photoperiodic flowering (Corbesier et al., 2007; Lin et al., 2007; Tamaki et al., 2007). The transmissible flowering stimulus could at least partially involve stress substances that are generated when plants are stressed. These substances include reactive oxygen species, nitric oxide (NO), jasmonic acid (JA), salicylic acid (SA), ethylene, and abscisic acid (ABA), which regulate gene expression to adapt to the stressful conditions (Kohli et al., 2013). Pathogen infection generally increases SA, JA, and ethylene levels, suggesting a possible involvement of these hormones in flowering accelerated by pathogen infection (Korves and Bergelson, 2003). Among them, SA and ethylene have been reported to induce flowering. Ethylene induces flowering in members of the Bromeliaceae, including pineapple (Trusov and Botella, 2006). However, this is an exceptional case, and ethylene generally inhibits flowering in many plant species, including Arabidopsis and pharbitis (Achard et al., 2007; Kęsy et al., 2010). NO also uncommonly induces flowering in L. aequinoctialis: the NO donors sodium nitroprusside, S-nitroso-N-acetyl penicillamine and 3-morpholinosydnonimine induce flowering under non-inductive conditions but inhibit flowering induced by SD (Khurana et al., 2011).

The most likely stress substance involved in stress-induced flowering is SA. When pharbitis and Lemna gibba are induced to flower in response to stress, they turn red due to the accumulation of anthocyanins, the biosynthesis of which is regulated by phenylalanine ammonia-lyase (PAL), suggesting an involvement of the metabolic pathway regulated by PAL in stress-induced flowering. The PAL inhibitors aminooxyacetic acid (AOA) and L-2-aminooxy-3-phenylpropionic acid (AOPP) inhibit flowering (Hatayama and Takeno, 2003; Wada and Takeno 2010; Shimakawa et al., 2102). PAL catalyzes the conversion of phenylalanine to t-cinnamic acid, from which both anthocyanin and SA are derived. When SA was applied together with AOA to pharbitis under stressful conditions, it negated the inhibitory effect of AOA. PAL gene expression, PAL enzyme activity, and SA content increase when flowering is induced by stress, and exogenous SA enhances the flowering response in pharbitis (Wada et al., 2014; Koshio et al., 2015). These results suggest that stress-induced flowering is at least partially induced by SA (Wada and Takeno, 2013). Cleland and Ajami (1974) found that exogenous SA induces flowering in L. gibba, and thereafter many articles have reported that SA and its immediate precursor benzoic acid induce flowering in several Lemnaceous species (Kandeler, 1985). Although no positive correlation has been found between the endogenous level of benzoic acid and photoperiodic conditions (Fujioka et al., 1983), greater quantities of SA are detected in plants flowering in response to poor nutrition than in non-stressed plants in L. paucicostata (Shimakawa et al., 2012). SA may be an endogenous regulating factor, not in photoperiodic flowering but in stress-induced flowering. The involvement of SA in stress-induced flowering is also suggested in Arabidopsis. UV-C stress promotes flowering in wild-type Arabidopsis Col but not in NahG transgenic lines that are unable to accumulate SA (Martínez et al. 2004). The expression of FT increases in UV-C-irradiated wild-type plants but not in NahG plants. UV-C increases expression of the SA-responsive Pathogenesis-related 1 gene and the SA induction deficient 2/isochorismate synthase 1 gene encoding the SA biosynthetic enzyme in Col plants but not in NahG plants. Exogenous SA accelerates the flowering of Col, but the NahG plants are not responsive to SA. Thus, SA is required for the UV-C light-activated flowering in Arabidopsis. The regulatory mechanism of stress-induced flowering that is mediated by SA may be common among different plant species.

In contrast to pharbitis and L. gibba, the leaves of the red-leaved form of P. frutescens normally turn green due to anthocyanin depletion when flowering is induced by low-intensity light stress (Wada et al., 2010a). The anthocyanin depletion is consistent with the above-mentioned anthocyanin filter hypothesis, but it is difficult to predict the involvement of PAL in the flowering of P. frutescens. PAL activity is induced by stress (Christie et al., 1994; Dixon and Paiva, 1995; Chalker-Scott, 1999) and also by high irradiance (Gong et al., 1997; Yamazaki et al., 2003). Does this suggest that low-irradiance light as a stress factor decreases or increases the PAL activity? In fact, the gene expression and enzyme activity of PAL decrease under low-intensity light stress (Miki et al., 2015). These results are in contrast to the report that PAL expression and PAL activity increase stress-induced flowering in response to poor nutrition in pharbitis (Wada et al., 2014). Low-intensity light stress-induced flowering in the red-leaved P. frutescens and nutritional stress-induced flowering in pharbitis may be regulated by different mechanisms.

SA alone does not induce flowering in pharbitis, indicating that SA may be necessary but not sufficient to induce flowering in this species. AOA and AOPP also inhibit the biosynthesis of indole-3-acetic acid (IAA; Soeno et al., 2010) and 1-aminocyclopropane-1-carboxylic acid synthase (Amrhein and Wenker, 1979), influencing the endogenous level of IAA, ethylene, and polyamines. The involvement of IAA and putrescine (Put) in stress-induced flowering has been suggested for pharbitis (Koshio et al., 2015), but the involvement of ethylene is unlikely (Hatayama and Takeno, 2003). Although SA, IAA, and Put may all be involved in the regulation of stress-induced flowering in pharbitis, none of them function as a flower-inducing factor in isolation. They may play a role in promoting stress-induced flowering through interactions among them, or some other unknown factors may be involved. SA-induced flowering is significantly reduced by the exogenous application of NO scavengers, NO synthase inhibitors, and nitrate reductase inhibitors in two strains of L. paucicostata, suggesting that NO is partially involved in SA-induced flowering (Khurana et al., 2014).

Although ABA is regarded as a general inhibitor of flowering, ABA promotes the drought-induced early flowering response under LD conditions in Arabidopsis (Riboni et al., 2013). ABA up-regulates FT/TSF and SOC1 expression in a photoperiod-dependent manner in the drought-induced early flowering response. When considering the effect of ABA, it is necessary to note that natural ABA, (+)-(S)-ABA, has a dual effect on the photoperiodic flowering of pharbitis; it may inhibit the time-measuring process as well as promote some processes that proceed after the generation of the flowering stimulus (Takeno and Maeda, 1996). GA, especially GA₅, is considered as a LD florigen in the LD plant Lolium temulentum (King et al., 2006), and sucrose is known as a flower-inducing factor in photoperiodic flowering (Bernier and Périlleux, 2005). However, the roles of GA and sucrose in stress-induced flowering are uncertain.

Regulation by FT and the flowering genes

Flowering of Arabidopsis is known to be accelerated by LD, vernalization, autonomous cues, and other factors that operate through a common pathway integrated by the floral pathway integrator gene FT (Boss et al., 2004). Homologs of FT are widely conserved, and their products are known to act as florigen in photoperiodic flowering (Corbesier et al., 2007; Lin et al., 2007; Tamaki et al., 2007). This suggests that FT is also possibly involved in stress-induced flowering. In fact, UV-C stress induces the expression of FT in Arabidopsis (Martínez et al., 2004). FT is also required for early flowering induced by drought stress and by F. oxysporum infection in Arabidopsis (Riboni et al., 2013; Lyons et al., 2015). Pharbitis has two orthologs of FT, PnFT1 and PnFT2, both of which are expressed when flowering is induced by SD. PnFT2 expression is induced in the cotyledons of cultivar Violet flowering in response to poor nutrition, but PnFT1 expression is not induced (Wada et al., 2010b; Yamada and Takeno, 2014). The expression of PnFT2 is also induced by low-temperature stress. Poor nutrition does not enhance PnFT2 expression in cultivar Tendan, which does not flower in response to poor nutrition but does flower in response to low temperature, and low temperature does induce its expression. These results suggest that PnFT2, not PnFT1, is the major regulatory gene involved in the stress-induced flowering of pharbitis. It is interesting to note that PnFT2 is involved in both photoperiodic and stress-induced flowering, whereas PnFT1 is involved only in photoperiodic flowering. The two PnFT genes may have different roles in the regulation of flowering depending on the inductive cue.

The early flowering of Arabidopsis in response to drought stress requires the flower-promoting gene GIGANTEA (GI) and the floral integrator SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) in addition to FT, and does not require the activity of CONSTANS (CO), a transcriptional regulator of FT that acts downstream of GI in the photoperiodic response (Riboni et al., 2013). Flowering accelerated by UV-C and by F. oxysporum infection also requires GI (Martínez et al., 2004; Lyons et al., 2015). The co-1 mutant is associated with accelerated flowering in response to UV-C, suggesting that the UV-C-mediated acceleration of flowering does not require CO function (Martínez et al. 2004). The pharbitis homologs of SOC1 (PnSOC1) and CO (PnCO) are constitutively expressed regardless of the growth conditions (stressed or non-stressed), suggesting that these genes may not be involved in the regulation of stress-induced flowering of pharbitis (Yamada and Takeno, 2014). This is consistent with reports that PnFT may not be regulated by the PnCO protein in photoperiodic flowering of pharbitis (Hayama et al., 2007; Higuchi et al., 2011). FD, the product of which represents a key FT interactor in the shoot apical meristem, is not required for the early flowering induced by drought stress in Arabidopsis (Riboni et al., 2013). The involvement of FD in stress-induced flowering of pharbitis is unlikely because the presence of the FD homolog has not been reported to date in pharbitis. These results suggest that the CO/FT module and FT/FD interaction are not involved in stress-induced flowering in either Arabidopsis or pharbitis. The above-mentioned flowering pathways are summarized in Fig. 1.

Fig. 1.

Hypothetical pathways of photoperiodic (thin arrows) and stress-induced (thick arrows) flowering in Arabidopsis (left) and pharbitis (right). Long-day conditions induce early flowering of Arabidopsis via the functions of GIGANTEA (GI), CONSTANS (CO), FLOWERING LOCUS T (FT) and SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1). The FT product interacts with the product of FD in the shoot apical meristem to induce flowering. Stress also induces flowering via GI, FT and SOC1, but does not require the activity of CO and FD. Pharbitis has two orthologs of FT, PnFT1 and PnFT2, both of which are involved in flowering induced by short days, whereas only PnFT2 is involved in flowering induced by stress. Pharbitis has homologs of GI (PnGI), CO (PnCO), and SOC1 (PnSOC1), but none of them is involved in both photoperiodic and stress-induced flowering. Presence of the pharbitis homolog of FD has not been reported. Note that the CO/FT module and the FT/FD interaction, which are important components in photoperiodic flowering of Arabidopsis, are not involved in stress-induced flowering of both species and in photoperiodic flowering of pharbitis. Knowledge of the gene regulation in stress-induced flowering is very limited, especially in pharbitis, and therefore the scheme illustrated here is just tentative. Salicylic acid is involved in stress-induced flowering in both species, but the interaction between salicylic acid and flowering genes is still unclear.

The expression of the circadian clock-regulated gene Pathogen and Circadian Controlled 1 (PCC1) is strongly activated by UV-C in Arabidopsis Col (Segarra et al., 2010). This UV-C-activated expression of PCC1 is dependent on CO. The over-expression of PCC1 does not accelerate flowering, but suppression of its expression by RNAi delays flowering, suggesting that PCC1 alone is not sufficient to accelerate flowering. PCC1 is not activated by UV-C in SA-deficient NahG plants, suggesting that the up-regulation of PCC1 by UV-C is SA-dependent (Segarra et al., 2010). This is additional evidence for the involvement of SA in UV-C stress-induced flowering. UV-C-accelerated flowering requires the reduced expression of the flower-inhibiting gene FLOWERING LOCUS C (FLC) in addition to the enhanced expression of FT in Arabidopsis (Martínez et al., 2004). However, infection with F. oxysporum enhances FLC expression (Lyons et al., 2015). The up-regulation of both floral promoters and repressors suggests that floral transition reprogramming in response to F. oxysporum infection is under complex genetic control. The expression of the pharbitis homolog of FLC has not been detected under either poor or normal nutritional conditions (Yamada and Takeno, 2014).

The requirement of FT activation for stress-induced flowering is common to Arabidopsis and pharbitis, but the genes that act upstream and downstream of FT to progress the flowering process are believed to be diverse depending on the type of stress and the plant species. Stress-responsive genes, particularly those in the cross-talk between stress tolerance and flowering time (Kazan and Lyons, 2016), are possibly involved in the regulation of stress-induced flowering. However, the application of stress does not necessary induce flowering, as reported in several studies that describe the cross-talk between gene networks involved in stress perception and flowering time (Kazan and Lyons, 2016). Further studies are necessary to clarify the gene network regulating stress-induced flowering.

Regulation of FT expression

As mentioned above, SA is the most likely common hormonal component regulating stress-induced flowering. SA induces the expression of FT in Arabidopsis, and the FT homolog HAFT in sunflower (Helianthus annuus) (Martinez et al., 2004; Dezar et al., 2010). FT is not induced by UV-C in NahG plants. These results suggest that FT expression is regulated by SA. The expression of PnFT2 induced by nutritional stress is suppressed by the PAL inhibitor AOA; SA reduces this suppression, and PnFT2 expression induced by stress is enhanced by SA compared with treatment with AOA alone in pharbitis (Yamada and Takeno, 2014). However, SA does not induce PnFT2 expression under normal nutritional conditions. SA alone may not be sufficient to induce PnFT2 expression. Stress may induce the production of SA and other unknown factors, which may work in combination to induce PnFT2 expression and flowering in pharbitis.

The genes upstream of FT in the stress response pathway remain unclear, but microRNAs (miRNAs) may be an important component (Yaish et al., 2011). MicroRNAs act at both the transcriptional and post-transcriptional levels to regulate key genes, and the miR169 family plays a key role in stress-induced flowering (Hong and Jackson, 2015). The miR169 family members are up-regulated under stress in Arabidopsis, rice (Oryza sativa), cotton (Gossypium arboreum), maize (Zea mays), and soybean (Glycine max), repressing the nuclear factor-YA transcription factor, which in turn reduces the expression of FLC, allowing the expression of FT to promote flowering (Xu et al., 2014). The molecule miR393 plays a role in the stress response by targeting mRNAs that code for the auxin receptors. The early flowering phenotype is observed in OsmiR393-overexpressing rice plants through hyposensitivity to the auxin signal, resulting from reduced expression of the auxin receptor genes OsTIR1 and OsAFB2 (Xia et al., 2012). The miRNA172E is up-regulated by drought stress, and its up-regulation is dependent on GI (Han et al., 2013). However, another miRNA induced by stress, miR156, delays flowering in Arabidopsis and tobacco (Cui et al., 2014; Zhang et al., 2015).

Low-temperature treatment can be replaced with DNA demethylation to induce flowering in vernalization-requiring plants, revealing that vernalization is regulated epigenetically (Burn et al., 1993; Finnegan et al., 1998; Michaels and Amasino, 2000). Similar epigenetic regulation is also suggested in stress-induced flowering (Yaish et al., 2011). Heat stress accelerates flowering in the Arabidopsis accession Sha-0, and the early flowering phenotype is heritable, probably through epigenetic mechanisms, when there is exposure to heat stress over three generations (Suter and Widmer, 2013). Epigenetic regulation can also be considered in pharbitis and P. frutescens, because they are induced to flower by treatment with the DNA demethylating reagents 5-azacitydine and zebularine (Kondo et al., 2006, 2007, 2010; Iwase et al., 2010). The DNA methylation status in these plants when they are stressed to induce flowering is uncertain, but it is known that several stress factors change DNA methylation status in maize, rice, and dandelion (Taraxacum officinale) (Steward et al., 2002; Verhoeven et al., 2010; Kou et al., 2011). The flowering traits of pharbitis and P. frutescens induced by DNA demethylation reagents are not heritable, although simultaneously induced dwarfism in P. frutescens is heritable. If the phenotype modified by stress is heritable, it is reminiscent of the classical concept of the inheritance of acquired traits (Sano, 2010; Kou et al., 2011).

Future prospects

It is apparent that stresses induce flowering in a wide range of plant species. Poor-nutrition stress-induced flowering has been well-studied in many plant species, but the definition of poor nutrition is as vague as the definition of plant stress. Castro Marín et al. (2011) have pointed out that little is known about how nutrients regulate flowering, because nutrients have large effects on vegetative growth, making it difficult to distinguish primary and secondary influences on flowering. A possible causal element of nutritional stress-induced flowering could be nitrogen deficiency, because nitrogen is the most important inorganic nutrient. Studies using experimental systems that were designed to ensure that vegetative growth was not affected found that low nitrate accelerates flowering (Tanaka et al., 1991; Castro Marín et al., 2011). These results suggest that low nitrate has a primary influence on flowering, and thus induced flowering is not stress-induced flowering because the plants were not stressed. If this interpretation is correct, flowering caused by the secondary influence of nutrients must be stress-induced flowering. This means that stress-induced flowering is essentially the result of general growth retardation, a systemic response. It is reasonable to assume that the retardation of vegetative growth accelerates flowering because flowering switches the plant from vegetative growth to reproductive growth. These interpretations need to be carefully examined in future studies.

SA and FT are the primary components involved in the regulation of stress-induced flowering. However, the genes functioning upstream and downstream of FT and the nature of the interaction between SA and FT are still unclear. The factor that operates in conjunction with SA to induce PnFT2 expression in pharbitis is unknown. These problems need to be addressed in the future. Analysis of why cultivar Tendan is not induced by nutritional stress to flower may be the most useful question to pursue. Pharbitis has contributed significantly to the physiological analysis of stress-induced flowering, but it has been subject to only limited molecular analysis. One of the reasons for this deficit may be that genetic transformation is not easy in this plant. Improvement of the transformation method in pharbitis is thus necessary.

Abbreviations:

Abbreviations:
 
• ABA

  abscisic acid

 
• AOA

  aminooxyacetic acid

 
• AOPP

  L-2-aminooxy-3-phenylpropionic acid

 
• CO

  CONSTANS

 
• Col

  Columbia

 
• FLC

  FLOWERING LOCUS C

 
• FT

  FLOWERING LOCUS T

 
• GA

  gibberellin

 
• GI

  GIGANTEA

 
• IAA

  indole-3-acetic acid

 
• JA

  jasmonic acid

 
• LD

  long-day

 
• Ler

  Landsberg erecta

 
• miRNA

  microRNA

 
• NO

  nitric oxide

 
• PAL

  phenylalanine ammonia-lyase

 
• PCC1

  Pathogen and Circadian Controlled 1

 
• PGPR

  plant growth-promoting rhizobacteria

 
• Put

  putrescine

 
• SA

  salicylic acid

 
• SD

  short-day

 
• SOC1

  SUPPRESSOR OF OVEREXPRESSION OF CONSTANS 1.

Author notes

Editor: Christine Raines, University of Essex