ABSTRACT

Many plant species are induced to flower by stress. Stress-induced flowering has been studied mostly in the short-day plant pharbitis (also called Japanese morning glory; Ipomoea nil, formerly Pharbitis nil). In this article, physiological characteristics, the regulation by salicylic acid (SA) and the expression of flowering-related genes in stress-induced flowering in pharbitis are reviewed. Pharbitis flowered under long-days in response to poor nutrition or low temperature. The pharbitis plants induced to flower by stress reached anthesis, fruited and produced fertile seeds. The progeny of the stressed plants developed normally. Grafting experiments indicated that a transmissible flowering stimulus is involved in poor nutrition stress-induced flowering. Aminooxyacetic acid (AOA), a phenylalanine ammonia-lyase (PAL) inhibitor, inhibited the stress-induced flowering, and this inhibition was overcome by SA. Stress induced PAL activity and SA biosynthesis. PnFT2, a pharbitis ortholog of the flowering gene FLOWERING LOCUS T of Arabidopsis thaliana, was expressed when the plants were induced to flower by stress. The overexpression of PnFT2 induced flowering, and PnFT2RNAi inhibited it. AOA inhibited PnFT2 expression induced by stress, and SA eliminated this inhibitory effect. SA enhanced PnFT2 expression under poor nutrition but not under non-stressful conditions. Therefore, stress may induce the production of SA and other unknown factor(s) that may work in combination to induce PnFT2 expression and flowering.

1. Introduction

Flowering in many plant species is regulated by environmental cues such as night length in photoperiodic flowering and temperature in vernalization [1]. In addition, recent studies have indicated that stress is a cue to induce flowering [2-4]. The short-day (SD) plant pharbitis (also called Japanese morning glory; Ipomoea nil, formerly Pharbitis nil) flowered under long-days (LD) when grown under poor nutrition, low temperature or light conditions of high intensity [5-8]. Poor nutrition, low temperature and high-intensity light are stresses. Accordingly, we named such non-photoperiodic flowering stress-induced flowering [9]. The SD plants Perilla frutescens var. crispa and Lemna paucicostata (also called Lemna aequinoctialis) flowered under LD when grown under low-intensity light and poor nutrition, respectively [10,11]. The long-day plant Arabidopsis thaliana was induced to flower early under ultraviolet-C (UV-C) radiation and poor nutrition [12,13]. Similar non-photoperiodic flowering has been reported in several plant species, and a review of those reports suggested that most of the factors responsible for flowering are stresses, including highor low-intensity light, high or low temperature, drought, poor nutrition and mechanical stimulation [2,4]. However, not all types of stress induce flowering in all plants. For example, P. frutescens flowers upon lowintensity light stress, but not upon poor nutrition, low temperature, salt or drought stresses [10]. Plants can modify their development to adapt to stressful conditions, and stress-induced flowering is such an adaptation. Plants flower as an emergency response when stressed, thus ensuring their ability to produce the next generation. Through this mechanism, the species is preserved, even under unfavorable environmental conditions. Therefore, stress-induced flowering is universal and as important as photoperiodic flowering and vernalization [2,4]. This article reviews the studies reported on pharbitis because stress-induced flowering has been best studied in this plant species.

2. Physiological Characteristics of Stress-Induced Flowering in Pharbitis

Pharbitis cv. Violet was induced to flower when grown in 1/10- or 1/100-strength mineral nutrient solution or tap water under LD conditions for 20 days or longer. In contrast, no flowering occurred in the full-strength nutrient solution [14,15]. The vegetative growth of the plants grown in the diluted nutrient solutions was significantly inhibited. Because the suppression of vegetative growth indicated that the plants were stressed [9], the flowering was likely stress-induced. The flowering response was stronger under the stronger stress condition (1/100- strength nutrient solution) than it was under the weaker stress condition (1/10-strength nutrient solution). Another cultivar, Tendan, did not flower under the same poor nutrition conditions, although the vegetative growth was significantly inhibited. Thus, poor nutrient stress does not induce flowering in all cultivars. Cv. Violet was induced to flower when grown at 13˚C or 15˚C for 10 days or longer under LD conditions [9,15]. Vegetative growth was suppressed at low temperatures. Cv. Tendan was also induced to flower by low-temperature stress. The flowering response induced by low temperature was stronger in Tendan than it was in Violet. The flowering response induced by low temperature was weaker than that induced by a single SD treatment in Violet. This trend was reversed in Tendan.

We hypothesized that stressed plants flower to produce the next generation as an emergency response if they cannot adapt to unfavorable environmental conditions. If this is the case, the plants induced to flower by stress must produce fertile seeds, and the progeny must develop normally. Cv. Violet grown in 1/10-strength nutrient solution or tap water throughout its life cycle flowered, reached anthesis, fruited and produced seeds [14]. Those seeds germinated, and the progeny developed normally and responded to SD treatment by forming floral buds. Furthermore, a normal second generation was produced. These data indicate that the stressed plants do not need to await the arrival of a season when photoperiodic conditions are suitable for flowering, and such precocious flowering may assist in species preservation. Therefore, stress-induced flowering may have a biological benefit, and it should be considered as important as photoperiodic flowering and vernalization.

A transmissible flowering stimulus such as florigen, which is involved in photoperiodic flowering, has not been reported in stress-induced flowering. The cotyledons are necessary for the flowering of pharbitis seedlings in response to poor nutrition and low temperature [5,16]. This necessity suggests that a flowering stimulus such as florigen is also involved in stress-induced flowering and that it is produced in cotyledons. If the stress-induced flowering stimulus is transmissible, defoliated scions may flower when grafted onto rootstocks with cotyledons and grown under stressful conditions. Cvs. Violet and Tendan were grafted in several combinations, and the grafted plants were grown in tap water under LD conditions [14]. The Violet scions grafted onto the Violet rootstocks flowered. The flowering may have been due to the rootstocks because all of the leaves were removed from the scions. Therefore, a transmissible flowering stimulus is involved in stress-induced flowering. We predicted that Tendan would not produce such a flowering stimulus because this cultivar did not flower under poor nutrition conditions. If this were the case, Violet would not be expected to flower when grafted onto Tendan rootstocks. However, defoliated Violet scions grafted onto Tendan rootstocks with cotyledons flowered. Conversely, Tendan scions did not flower when grafted onto Violet rootstocks. These results indicate that Tendan produces a transmissible flowering stimulus but does not respond to it.

3. Regulation of Stress-Induced Flowering by Salicylic Acid (SA)

When pharbitis was induced to flower by poor nutrition, the plants turned red due to the accumulation of anthocyanins, whose biosynthesis is regulated by phenylalanine ammonia-lyase (PAL). The activity of PAL increases when plants are stressed [17,18]. These results suggest that the metabolic pathway mediated by PAL is involved in the regulation of stress-induced flowering. Aminooxyacetic acid (AOA) and L-2-aminooxy-3- phenylpropionic acid, which are PAL inhibitors [19,20], inhibited stress-induced flowering in pharbitis [9,14]. PAL catalyzes the conversion of phenylalanine to t-cinnamic acid, from which anthocyanins, chlorogenic acid and SA are produced. We inhibited stress-induced flowering by applying AOA in combination with various metabolites. Among them, t-cinnamic acid, benzoic acid (a precursor of SA) and SA negated the inhibitory effect of AOA, while p-coumaric and caffeic acids, which lead to the synthesis of anthocyanins and chlorogenic acid, did not [9,14]. Flowering was induced by growing the plants in a 1/100-strength nutrient solution, and PAL gene expression in the cotyledons harvested at the end of the stress treatment was examined by reverse transcription-polymerase chain reaction (RT-PCR). The expression was stronger in the plants that were induced to flower by stress than in the control plants [21]. The plants were then induced to flower in the same manner as above, and PAL enzyme activity in the cotyledons was measured. The PAL activity was higher in the stressed plants than in the control plants. The SA content in the cotyledons was also measured by high performance liquid chromatography-mass spectrometry. It was higher in the stressed plants than in the control plants. These results suggest that poor nutrition stress-induced flowering in pharbitis is induced by SA synthesis, which is promoted by PAL.

The effect of exogenously applied SA on flowering was examined. SA did not induce flowering under nonstressful conditions, but it promoted flowering under stressful conditions (unpublished data). Therefore, SA may be necessary but not sufficient to induce flowering. Stress may induce the production of not only SA but also other essential factors involved in the induction of flowering.

It has long been supposed that SA is involved in the regulation of photoperiodic flowering in the Lemnaceous plants because SA induces flowering in L. paucicostata, Lemna gibba and some other species belonging to the Lemnaceae [22-24]. However, no evidence has shown an increase in the SA content under flowering conditions. When L. paucicostata was induced to flower by poor nutrition, the SA content in the fronds increased [11]. This is the first evidence that the SA content increases when flowering is induced in a Lemnaceous species. SA may function as an endogenous regulating factor in stress-induced flowering but not in photoperiodic flowering. The involvement of SA was also reported in the stress-induced flowering of A. thaliana [12,25]. These data support the conclusions for pharbitis described above. The regulatory mechanism of stress-induced flowering mediated by SA may be common among various plant species.

4. Genetic Regulation of Stress-Induced Flowering

The molecular basis of the regulation of stress-induced flowering is not well understood. A. thaliana flowering is induced by LD, vernalization, autonomous cues and gibberellins, and these factors operate through a common pathway integrated by the floral pathway integrator gene FLOWERING LOCUS T (FT) [26]. This suggests that FT could also be involved in stress-induced flowering. Two orthologs of FT, PnFT1 and PnFT2, have been identified in pharbitis, and these genes are expressed under inductive SD conditions to promote flowering [27]. Therefore, we examined PnFT expression in stress-induced flowering by RT-PCR. Pharbitis cv. Violet was grown in a 1/100-strength nutrient solution under LD conditions, and a control group of plants was given a single SD treatment without the stress treatment as a positive control. PnFT1 and PnFT2 were induced in the cotyledons when flowering was induced by the SD treatment, but neither gene was expressed without the SD or the stress treatment [14,15]. PnFT2 was induced in the plants that flowered under the poor nutrition conditions for two weeks or longer. The level of mRNA expression closely correlated with the flowering response. Conversely, PnFT1 was not expressed when flowering was induced by poor nutrition. PnFT2 expression in cv. Tendan, in which poor nutrition does not induce flowering, was also examined. The expression level was similar in plants grown under both poor nutrition and non-stressful conditions and was much lower than in the SD-treated cotyledons. PnFT2 expression in the stressed leaves was weaker in Tendan than in Violet. PnFT1 expression was also not induced by poor nutrition in Tendan. These results suggest that PnFT2, but not PnFT1, is the major regulatory gene involved in the stress-induced flowering of pharbitis. Violet was induced to flower by growing it at 15˚C for 2 weeks or longer, and the expression of PnFT2 was detected starting at the first week of the low temperature treatment [15]. PnFT1 expression was also detected, but a significant increase was not observed until the fourth week of the stress treatment. The low temperature induced PnFT2 expression, and PnFT1 expression was quite weak in Tendan as well as in Violet. PnFT2 expression was stronger in Tendan than in Violet. These results indicate that PnFT2 is also involved in low temperature stress-induced flowering. The involvement of PnFT2 in stress-induced flowering in pharbitis is consistent with a previous report indicating that FT is involved in UV-C stress-induced flowering in A. thaliana [12]. It is interesting to note that PnFT2 is involved in both photoperiodic flowering and stress-induced flowering, whereas PnFT1 is involved only in photoperiodic flowering in pharbitis. The two PnFT proteins might have different roles in the regulation of flowering depending on the inductive cue. PnFT2 and PnFT2RNAi were transformed into A. thaliana and pharbitis to test whether PnFT2 has flower-inducing activity (Yamada, unpublished data). A. thaliana transformed with 35S:: PnFT2 flowered much earlier than the wild type under continuous light. Some pharbitis somatic embryoids transformed with 35S::PnFT2 flowered without regeneration of vegetative shoots during cultivation on regeneration medium. Stress-induced and SD-induced flowering were suppressed in the pharbitis plants transformed with PnFT2RNAi. These results indicate that PnFT2 has flower-inducing activity.

SUPPRESSOR OF OVEREXPRESSION OF CO1 (SOC1) is thought to be another floral pathway integrator gene in A. thaliana [28,29]. Therefore, it was expected that SOC1 expression might also be induced when flowering was induced by stress in pharbitis. However, the pharbitis SOC1 homolog was expressed constitutively and showed no correlation with flowering response [15]. It was reported that UV-C, which induced flowering, did not induce SOC1 expression in A. thaliana [12]. These results suggest that SOC1 is not involved in stress-induced flowering. CONSTANS (CO), a gene upstream of FT, is involved in the regulation of photoperiodic flowering in many plants [30]. The CO protein directly induces the transcription of FT in A. thaliana and Oryza sativa [31-33]. It has been proposed that the CO/FT regulatory module (i.e., the CO protein activates FT transcription) is highly conserved in both dicot and monocot plants [34]. The expression of CO was also moderately induced in UV-C-induced flowering in A. thaliana [12]. The pharbitis homolog of CO, PnCO, complements the co mutation in A. thaliana [35]. We found, however, that PnCO is constitutively expressed regardless of the flowering status in pharbitis; it is expressed under both rich and poor nutrition and both SD and LD conditions [15]. These results suggest that PnCO may not be involved in the regulation of both stress-induced and photoperiodic flowering in pharbitis. This conclusion is consistent with the previous report that PnFT mRNA abundance was not related to PnCO expression and that therefore PnFT may not be regulated by the PnCO protein in pharbitis [27,36]. Furthermore, a night-break treatment inhibited flowering but did not influence PnCO expression [35]. The CO/FT module may not be conserved in pharbitis.

RT-PCR on the genes that act upstream of FT showed that the expression of the homologs of FLOWERING LOCUS C, FRIGIDA and FVE was not induced by poor nutrition, and the FCA homolog was expressed even in the absence of stress [15]. Thus, candidate genes that act upstream of and regulate FT expression were not identified among the known flowering genes. Among the genes downstream of FT, the expression of the APETALA1 homolog correlated well with the flowering response, but the expression of the FD homolog was not detected (Yamada, unpublished data).

5. Interaction between SA and PnFT2 in the Regulation of Stress-Induced Flowering

We hypothesized that stress-induced flowering is regulated by PnFT, which is induced by SA. Accordingly, we examined the influences of both a PAL inhibitor and SA on PnFT expression in pharbitis. AOA inhibited flowering and PnFT2 expression induced by poor nutrition, and SA eliminated the inhibitory effects of AOA [15]. SA enhanced PnFT2 expression under poor nutrition but not under non-stressful conditions. These results suggest that SA induces PnFT2 expression, which in turn induces flowering. This conclusion is consistent with the results observed in A. thaliana and sunflower, in which SA induced the expression of FT and HAFT, a sunflower ortholog of FT [12,37]. However, SA did not induce PnFT2 expression and flowering under non-stressful conditions in pharbitis, suggesting that SA alone may not be sufficient to induce PnFT2 expression. Stress may induce the production of SA and other unknown factor(s), which may work in combination to induce PnFT2 expression and flowering.

6. Conclusion

Pharbitis is typically induced to flower under SD conditions and can be induced to flower under unfavorable photoperiodic conditions when exposed to stress. Stress activates PAL and enhances SA levels, promoting PnFT2 expression to induce flowering. However, SA alone cannot induce flowering, and other factors are necessary. The regulation of stress-induced flowering by SA may be common, at least in L. paucicostata and A. thaliana.

REFERENCES