Effects of High Temperature on Embryological Development and Hormone Profile in Flowers and Leaves of Common Buckwheat (Fagopyrum esculentum Moench)

Common buckwheat is a valuable crop, mainly due to the beneficial chemical composition of its seeds. However, buckwheat cultivation is limited because of unstable seed yield. The most important reasons for the low yield include embryo and flower abortion. The aim of this work is to verify whether high temperature affects embryological development in this plant species. The experiment was conducted on plants of a Polish cultivar ‘Panda’ and strain PA15, in which the percentage of degenerating embryo sacs was previously determined and amounted to 32% and 10%, respectively. The plants were cultivated in phytotronic conditions at 20 °C (control), and 30 °C (thermal stress). The embryological processes and hormonal profiles in flowers at various developmental stages (buds, open flowers, and wilted flowers) and in donor leaves were analyzed in two-month-old plants. Significant effects of thermal stress on the defective development of female gametophytes and hormone content in flowers and leaves were observed. Ovules were much more sensitive to high temperature than pollen grains in both genotypes. Pollen viability remained unaffected at 30 °C in both genotypes. The effect of temperature on female gametophyte development was visible in cv. Panda but not in PA15 buds. A drastic reduction in the number of properly developed embryo sacs was clear in open flowers at 30 °C in both genotypes. A considerable increase in abscisic acid in open flowers ready for fertilization may serve as a signal inducing flower senescence observed in the next few days. Based on embryological analyses and hormone profiles in flowers, we conclude that cv. ‘Panda’ is more sensitive to thermal stress than strain PA15, mainly due to a much earlier response to thermal stress involving impairment of embryological processes already in the flower buds.


Introduction
Common buckwheat (Fagopyrum esculentum Moench) belongs to the Polygonaceae family, but it is classified as a "pseudocereal" crop due to the cereal-like chemical composition of its seeds. The seeds are gluten-free; have high dietary fiber, high content of rutin and protein, and a well-balanced

Pollen Viability and Development
Pollen viability was generally high (>82%) in cv. 'Panda' and strain PA15, independent of temperature, suggesting regular meiosis. Once released from the tetrad (after meiosis), the microspores enlarged, forming vacuoles and thick sporodermis (Figure 1a). Then, they developed into pollen grains. Mature (three-celled) viable pollen exhibited dense cytoplasm (Figure 1b-e) and a positive purple-red or green stain in Alexander and fluorescein diacetate (FDA) tests, respectively (Figure 1b,c). 'Panda's' flowers showed significantly lower pollen viability at 20 • C (82%) than that of strain PA15 (96.3%) (compare Figure 1b vs. Figure 1e). We found no significant influence of thermal stress on pollen viability in either genotype compared with that of the control (compare Figure 1c vs. Figure 1d). Viable and non-viable microspores and pollen grains of common buckwheat cv. 'Panda' and strain PA15 under control (20 • C) and thermal stress (30 • C) conditions. Vacuolated microspores of 'Panda' at 20 • C with thick sporodermis (a, arrow); viable pollen grains of PA15 at 20 • C stained with Alexander dye (b) and at 30 • C stained with Ehrlich's hematoxylin combined with Alcian blue (c); and viable and non-viable (arrows) pollen grains of 'Panda' at 30 • C stained with fluorescein diacetate (d) and at 20 • C stained with Ehrlich's hematoxylin combined with Alcian blue (e). At least 7900 pollen grains per treatment were analyzed in two replicates.

Ovule Development
We observed regular megasporogenesis and different stages of female gametophyte development in buds, open flowers, and wilted flowers of both genotypes ( Figure 2). In buds, we noticed clear influence of temperature in cv. 'Panda' but not in strain PA15, in which embryological processes were slightly impaired at both 20 • C and 30 • C (83.3% and 79.5%, respectively) ( Table 1). In open flowers, the number of properly developed embryo sacs at the control temperature slightly decreased in both genotypes (88.2% in 'Panda' and 77.8% in PA15), and a strong impact of high temperature on female gametophyte development was evident in both genotypes. In wilted flowers, this effect was not so clear, probably because flowers with degenerated embryo sacs had been earlier aborted (Table 1).  Of the abnormal development patterns of female gametophyte observed under both temperatures, improper position of the nucleus in relation to the vacuole in the egg cell, degeneration of the entire embryo sac, and degeneration of the cells of the embryo sac and of entire ovules were the most frequent ( Figure 3).

Total Hormone Content
We determined the hormone content in the flower buds, open flowers, and wilted flowers and in the donor leaves. Table 2 presents the total contents of the studied hormones in the plant organs for strain PA15 and cv. 'Panda' at both temperatures. We decided to show the differences between hormone amounts in particular organs independently of growth conditions, as such data are lacking in the literature.
Leaves accumulated the lowest amounts of cytokinins (zeatin and kinetin) as compared with buds and open and wilted flowers. Active forms of gibberellin (GA) and non-active GA 8 occurred mainly in the open and wilted flowers. Among non-active forms of gibberellins, GA 9 was the most abundant, and it was detected mainly in the open flowers. This gibberellin also constituted the greatest share in the total pool of gibberellins. Non-active GA 20 was present in scarce amounts in the flowers but reached its highest level in the leaves. The greatest amount of indole-3-acetic acid (IAA) was found in the wilted flowers, whereas the leaves contained only minute amounts of this hormone. Open flowers accumulated greater amounts of active abscisic acid (ABA-free) than the other organs. In contrast, leaves contained the highest levels of ABA glucosyl ester (ABA-glc). All the developmental stages of flowers showed higher levels of salicylic acid (SA) than leaves, whereas leaves were the richest in jasmonates. The results described above prompted us to conclude that it is difficult to state how or if hormone contents in the donor leaves affect hormone contents in the flowers. GA 6 and GA 1 were the most common in the total pool of active gibberellins in all studied organs, whereas GA 7 level was the lowest (Table 3). No significant differences in percentage content of particular gibberellins in cv. 'Panda' and strain PA15 were observed.   4 5.1 ± 0.5 a 4.5 ± 0.5 b 5.7 ± 0.7 a 4.7 ± 0.5 b GA 5 8.6 ± 1.0 c 8.3 ± 1.0 c 11.3 ± 1.3 bc 15.1 ± 1.8 a GA 6 36.6 ± 5.3 b 44.4 ± 7.7 a 43.6 ± 5.2 a 42.5 ± 5.1 a GA 7 0.9 ± 0.1 b 1.

Hormone Content in the Flowers and Leaves
The  The IAA level increased drastically in the wilted flowers of the PA15 plants exposed to thermal stress ( Figure 5a), but in the other cases, high temperature did not change the amount of this hormone. The IAA level in flower buds of cv. 'Panda' at both temperatures was significantly higher than that in the PA15 plants. We found no effects of temperature on IAA amounts in the leaves of any of the studied plants ( Figure 5b).  The GA level in the leaves of cv. 'Panda' was higher at both temperatures than in the PA15 leaves, but thermal stress did not affect the GA amounts in either genotype ( Figure 6b).
Generally, greater amounts of non-active GA 8       In the leaves of the PA15 plants, thermal stress reduced the SA content, contrary to that in the leaves of cv. 'Panda', where no effect of high temperature was observed (Figure 11b).
The jasmonic acid (JA) level was much higher in the buds of the PA15 and 'Panda' plants at both temperatures than those at the other stages of flower development. However, the buds at 30 • C contained less JA than the control buds (Figure 12a). 'Panda' responded differently to high temperature: in the PA15 leaves, the JA amount declined, but in the 'Panda' leaves, it was higher than that at 20 • C (Figure 12b). In general, JA was more abundant in the leaves than in the flowers of all the plants.
The amounts of methyl ester of JA (JA-Met) detected in the flowers (expressed as fmol g −1 dry weight (DW)) were much lower than those in the leaves (expressed as µmol g −1 DW) ( Figure 13). We noticed huge differences in the accumulation of this JA form at different temperatures. In the PA15 flowers of all the developmental stages, no significant temperature effect on JA-Met amount was visible, whereas in the 'Panda' flowers, high temperature increased its level as compared with that of the control. The highest amount of JA-Met was detected in the 'Panda' open flowers exposed to high temperature. In the leaves of all the studied plants, thermal stress drastically lowered the JA-Met amount.

Quantitative Ratio of the Studied Hormones
Taking into account the role of individual hormones in the generative development of plants at the applied temperatures, we calculated the ratio of active GAs to ABA-free, IAA to ABA-free, and IAA to CYT (Table 4). Gibberellins were more abundant than ABA-free in all the organs of the PA15 and 'Panda' plants at both temperatures. In all the organs, thermal stress significantly increased the amount of active gibberellins in relation to ABA-free, and thus, the GAs/ABA-free ratio increased. The ratio was the highest in the 'Panda' open flowers at 30 • C. High temperature significantly boosted the IAA/ABA ratio mainly in buds (six times for PA15 and 10 times for 'Panda') and almost doubled it in the leaves. In all the studied plants, CYT levels were much lower than those of IAA. A significant advantage of auxin over cytokinins was observed at 30 • C in the buds of PA15 and cv. 'Panda', whereas in the open and wilted flowers, IAA prevailed over CYT at 20 • C. In the buds of all the studied plants at 20 • C, the IAA/CYT ratio was much lower than those in the open and wilted flowers, whereas at 30 • C this proportion was reversed. In the 'Panda' flowers, the IAA/CYT ratio at higher temperature decreased to a greater extent than that in the PA15 flowers. High temperature did not affect the IAA/CYT ratio in the leaves.

Discussion
Our earlier studies showed that the percentage of developmental disturbances occurring in mature female gametophytes resulting from premature degeneration of synergids or the egg apparatus in flowers of plants grown in natural conditions amounts to 40%-55% and depends mainly on the genotype [8]. It seems, therefore, that the key reason for the low yield is the defective development of female gametophytes, taking into account that the anthers of buckwheat produce viable pollen (>90%) [8]. Pollen is formed earlier than the female gametophytes, when plants still have abundant assimilates and air temperature is not high. Therefore, these factors do not affect microsporogenesis and microgametophytogenesis (data not shown). In contrast, the developmental processes in megasoporogenesis and megagametophytogenesis happen later and occur in older flowers and at higher temperatures; therefore, they may be affected by the availability of assimilates and temperature.
McGregor [17] revealed an approximate flower abortion rate of 45% in Brassicaceae. The cause of this phenomenon may be the genetic background or the influence of abiotic factors, such as temperature. For example, abortion causes smaller yields, especially in legumes [18]. Patric [19] suggests that in Vicia faba L., minor apical auxin probably regulates the disintegration of generative organs as a result of competition for assimilates rather than through polar and basal transport to the lower parts of the main shoot. Some plant species have the ability to abort low-quality embryos selectively, which raises the average quality of the surviving offspring [20].
According to Moe [21], the process of floral abortion is initiated during the early stages of shoot growth before the differentiation of floral parts is completed. Low temperatures (12)(13)(14)(15) • C) at this critical stage of development strongly promote blind shoot formation but do not affect stamen and pistil primordia formation in the apical flower bud. Apart from genetic control, the course of embryogenesis and the abortion of flowers and fruit is also influenced by plant hormones and growth regulators, and in particular the relationship between their concentrations [22].
In this experiment, we analyzed the changes in individual hormone content during buckwheat flower development from buds to wilting. Our attempt to explain the disturbances in the embryological development of buckwheat by changes in the hormonal profile of the flowers is an innovative approach and may account for poor seed yielding of buckwheat. The hormone content in individual organs of common buckwheat has not been reported yet.
We found numerous differences in hormone accumulation in common buckwheat organs. Contrary to ABA-free, ABA-glc was present in greater amounts in the leaves than in the flowers. Similarly, as reported by Wang and Irving [22], a higher level of jasmonates accumulated in the donor leaves than in the flowers. According to these authors, the insufficiency of jasmonic acid affects anther or ovule development and results in sterile flower organs. This is probably the reason for embryo sac degeneration in buckwheat plants grown under thermal stress.
Our results showed that both studied genotypes differ in their response to thermal stress. Cultivar 'Panda' seems to be more sensitive to high temperature than strain PA15, as its response to thermal stress was more rapid. Only in the 'Panda' buds did high temperature drastically increase the percentage of embryo sac degeneration. In the open flowers of 'Panda' at 30 • C, the number of degenerated embryo sacs doubled, and in the open flowers of PA15, the percentage of properly developed embryo sacs was 2.7 times lower than that of the control. These processes were accompanied by an increase in the content of cytokinins, active gibberellins, and JA-Met. Cytokinins control cell division and organ differentiation, for example, while gibberellins stimulate plant elongation and promote flowering. In Prunus avium, endogenous GA induced early embryo sac development, which resulted in a low seed set under high temperature [23]. Methyl jasmonate and jasmonic acid are important cellular regulators mediating diverse developmental processes, such as seed germination, flower and fruit development, leaf abscission, and senescence [24]. Interestingly, particularly in the 'Panda' ovules at 30 • C, female gametophytes appeared to be extremely luxuriant, but detailed analysis revealed their abnormal vacuolization. Abnormal vacuolization of common buckwheat embryos was observed at 32 • C in Japanese cultivars, displaying only 30% seed set [25,26].
Buckwheat flowers and leaves contain active and non-active forms of gibberellins. GA 9 occurred in the largest quantities of all the determined gibberellins. This gibberellin is a precursor in the GA 1 biosynthesis pathway [27]. Gibberellins GA 6 , GA 1 , and GA 3 occurred in larger quantities than other active gibberellins. The profiles of individual active gibberellins in buds, open flowers, and wilted flowers at both temperatures were similar. For this reason, we presented the changes in total sum of active GAs. Halińska and Lewak [28] and Chien et al. [29] reported that a combination of some gibberellins had a greater effect on plant growth and development than individual gibberellin levels.
ABA is a phytohormone affecting many physiological processes. Its role in flowering promotion is ambiguous. There are data showing negative [30], as well as positive [31], regulation of flowering in Arabidopsis. In our study, the pattern of changes in ABA content during flower development was concurrent with earlier reports that ABA plays a signaling role in flower senescence [32,33].
According to Aneja et al. [32] and Panavas et al. [34], the levels of endogenous ABA had increased markedly before any signs of senescence became visible and kept rising during petal senescence in such plant species as cocoa and daylily. Flower senescence in some plant species depends on another signal hormone-ethylene [35]. Zhong and Ciafré [36] described Iris as a species whose flowers are ethylene-independent and whose senescence is regulated by ABA. In our opinion, common buckwheat could also belong to this group, because a single flower of buckwheat lives very briefly. In the plants we investigated, ABA-free content was significantly greater in all open flowers than in the flower buds and wilted flowers. Only in the PA15 wilted flowers at 20 • C was the ABA-free level still high, and in the other variants, the wilted flowers demonstrated a considerable decrease in this hormone. In the 'Panda' open flowers, high temperature increased ABA-free content, which decreased in the open flowers of PA15.
The GAs/ABA-free ratio in open flowers of cv. 'Panda' demonstrated that at a high ABA level at 30 • C, the content of active gibberellins was even higher. Gibberellins and ABA can act antagonistically [37], and their role is very well known, especially in seed dormancy and germination [38]. We also observed changes in the IAA/ABA and IAA/CYT ratios. The IAA/ABA ratio increased drastically at 30 • C in the 'Panda' and PA15 flower buds. This could explain the lack of embryological disturbances under thermal stress but only in the PA15 plants. In the 'Panda' buds, only 58% of embryo sacs developed properly as compared with the control. In the buds of all the plants, the IAA/CYT ratio increased six times under thermal stress, but in the open flowers, it was lower at 30 • C than at 20 • C. This result seems to be in accordance with data reported for the effects of high temperature in Arabidopsis. These conditions trigger an activation of PHYTOCHROME-INTERATING FACTOR 4 (PIF4). This protein regulates auxin biosynthesis at higher temperatures [30,39]. PIF4 genes stimulate GAs accumulation, increase the transcription level of enzymes synthesizing GAs, and lower the transcription level of GAs inactivating enzymes [40]. Auxins and cytokinins also interact on metabolic levels, and auxins rapidly suppress the cytokinin pool [41]. These considerations could be supported by future experiments with developing flower explants treated with exogenous phytohormones.

Plant Material
The study was carried out in common buckwheat plants of a Polish cultivar 'Panda' and strain PA15 in phytotronic conditions. Seeds were supplied by Plant Production Facility in Palikije (Małopolska Plant Breeding Station, Polanowice, Poland).

Experimental Treatments
The experiment was carried out in phytotronic chambers. Plants were cultivated in pots (20 × 20 × 25 cm; 9 plants per pot), containing commercial soil substrate (pH = 5.8) mixed 1:1 with perlite (v:v). Plants were grown for 3 weeks at the control temperature (20 • C) at a humidity of 50%-60% under 16 h photoperiod and 300 µmol m −2 s −1 of PPFD (photosynthetic photon flux density). Then, half of the plants (all at the vegetative stage) were transferred into a chamber with a temperature of 30 • C (heat stress) and the same humidity and light conditions. Then, from two-month-old plants, the flowers at three developmental phases (buds, open developed flowers, and wilted flowers) were collected, and their embryological development (embryo sacs) and hormonal profiles were analyzed. Pollen viability was determined in open flowers. Additionally, donor leaves (fully developed young leaves, closest to the flower cluster) were collected for hormone analysis.

Embryological Analyses Pollen Viability
Initially, several randomly chosen flowers per treatment were taken for pollen viability screening via a FDA test (fluorescein diacetate). FDA dye was prepared according to Dafni and Firmage [42], as follows: 2 cm 3 20% sucrose in H 2 O with several drops of stock solution of FDA (2 mg FDA/1 cm 3 acetone). Freshly stained pollen was kept in a humid chamber for 30 min at 24 • C and afterwards observed with a Nikon E80i microscope with a UV-2A filter. Viable pollen emits yellow-green fluorescence. No counts were performed for the stained pollen. For the pollen viability count using Alexander's test, 20 open flowers per treatment were randomly collected and fixed in FAA (10 cm 3 96% ethanol, 7 cm 3 H 2 O, 2 cm 3 37% formaldehyde solution, 1 cm 3 glacial acetic acid) solution. Alexander's dye is a mixture of malachite green staining the cellulose of pollen walls green and acid fuchsin staining the pollen protoplast red [43]. Viable pollen grains appear purple-red, and non-viable pollen grains stain green. At least 7900 (altogether, viable and non-viable) pollen grains per treatment were counted under a Nikon E80i microscope (Tokyo, Japan) in two replicates.

Ovule Development
Paraffin sections of ovules were obtained by fixing flowers at three stages of development (buds, fully developed, and wilted flowers) in FAA solution, dehydrating them in increasing series of ethanol and saturating them with chloroform (in increasing proportion-1:3, 1:1, 3:1, 1:0-with absolute ethanol, each for 2 h at room temperature (RT)) and with paraffin dissolved in chloroform (at 57 • for several days, until chloroform evaporated). Flowers prepared this way were embedded in paraffin blocks, sliced into 11-15 µm sections on a rotary microtome (Adamas Instrumenten BV, HM 340E, Leersum, Netherlands), and double stained with Ehrlich's hematoxylin and Alcian blue [44]. Finally, the slides were mounted in Entellane (Sigma-Aldrich, St Louis, MO, USA) and analyzed under a Nikon E80i microscope. The number of analyzed flowers per treatment at each stage of development was 15-32 in two replicates.

Statistical Analyses
Two-way analysis of variance (ANOVA) and Duncan's multiple range test (at p < 0.05) were performed using the statistical package STATISTICA 13.0 (Stat-Soft, Inc., Tulusa, OK, USA). Data were presented as means ± SE (standard error). Non-normal distribution data were analyzed using Chi-squared test (χ 2 , p < 0.05).

Conclusions
Ovules are much more sensitive to thermal stress than stamens in both genotypes. A drastic reduction in the number of properly developed ovules in open flowers is visible at 30 • C, but this temperature does not affect pollen development and wilted flowers. A considerable increase in ABA in open flowers ready for fertilization may serve as a signal inducing flower senescence observed in the next days. Based on the embryological analyses and hormone profiles in flowers, we conclude that cv. 'Panda' is more sensitive to thermal stress than strain PA15, mainly due to a much earlier response to this stress involving the disturbances in embryological processes already in the flower buds.