Asymbiotic Germination of Mature Seeds and the Seed Development of Vanilla Planifolia


 Background: Vanilla planifolia is an important tropical orchid for production of natural vanilla flavor. Traditionally, V. planifolia is propagated by stem cuttings, which produces identical genotype that are sensitive to virulent pathogens. However, sexual propagation with seed germination of V. planifolia is intricate and unstable because of the extremely hard seed coat. A better understanding of seed development, especially the formation of impermeable seed coat would provide insights into seed propagation and conservation of genetic resources of Vanilla.Results: We found that soaking mature seeds in 4 % sodium hypochlorite solution from 75 to 90 min significantly increased germination and that immature seeds collected at 45 days after pollination (DAP) had the highest germination percentage. We then investigated the anatomical features during seed development that associated with the effect of seed pretreatment on raising seed germination percentage. The 45-DAP immature seeds have developed globular embryos and the thickened non-lignified cell wall at the outermost layer of the outer seed coat. After 60 DAP, the cell wall of the outermost layer of the outer seed coat became lignified and finally compressed into a thick envelope. These features matches the significant decreases of immature seed germination percentage after 60 DAP. Conclusion: We report a reliable protocol for seed pretreatment of mature seeds and for immature seeds culture based on a defined time schedule of V. plantifolia seed development. The thickened and lignified seed coat formed an impermeable envelope surrounding the embryo, and might play an important role in seed dormancy of V. plantifolia.

usually showed very low germination (Knudson 1950, Menchaca et al. 2011, which might result from the impermeability of the hardened seed coat (Withner 1955 This study aimed to establish an effective propagation method of V. planifolia via asymbiotic germination. We rst tested the effects of asymbiotic germination at different seed developmental stages. We then examined the effect of different combinations of sodium hypochlorite strengths and soaking time on the germination of mature seeds under asymbiotic germination treatments. We also investigated the morphological, histological, and histochemical changes of seed development within a de ned timescale, from fertilization to seed maturity. Based on this time frame, we determined the optimal timing for collecting seeds to inoculate in vitro. Such studies would bene t mass propagation to meet commercial needs and breeding programs for vanilla production.

Plant material and seed collection
The mature plants of V. planifolia were maintained in a greenhouse at Taoyuan District Agricultural Research and Extension Station at Taoyuan City, Taiwan. Anthesis generally occurs in late April each year ( Fig. 1A). For the pod setting, owers were hand-pollinated by transferring the pollinia onto the stigma of the same ower (Fig. 1B). Developing pods (Fig. 1C) were collected at regular intervals for morphological measurement, histology, and seed germination experiments. In January the next year, pods began to mature and turned yellow (Fig. 1D). In each experiment, at least three pods were collected at regular intervals after pollination.

Evaluation of asymbiotic germination percentage
The pods of different developmental stages were surface-sterilized with a 1.8% sodium hypochlorite solution with one drop of a wetting agent (Tween 20, Sigma-Aldrich) for 15 min. After surface sterilization, the capsules were cut open, and seeds were scooped out with forceps onto the culture medium. To ensure the seed quality and developmental stages of each capsule, the remaining seeds of each capsule were xed and examined under a microscope. The culture medium used was 1/2 Murashige and Skoog (MS) (Murashige and Skoog 1962) supplemented with 2 mg l − 1 glycine, 0.5 mg l − 1 niacin, 0.5 mg l − 1 pyridoxine HCl, 0.1 mg l − 1 thiamine, 1 g l − 1 tryptone, 20 g l − 1 sucrose, and solidi ed with 7 g l − 1 agar (plant cell culture, tested, powder; all Sigma-Aldrich). The pH value was adjusted to 5.7 before autoclaving at 121 °C and 1.2 kg cm 2 for 20 min. An amount of 10 ml medium was placed into each culture tube (20 × 100 mm). After sowing, the cultures were incubated in a growth room under a 16/8 hours photoperiod with daylight uorescent lamps (20W, China Electric, Taipei) at light intensity 30 µmol m − 2 s − 1 . Germination percentage was recorded 60 days after sowing.

Effect of seed pretreatments on germination
To examine the effectiveness of different concentrations and durations of sodium hypochlorite pretreatments in improving germination, mature seeds at 135 DAP were collected. The pretreatments included soaking the seeds in 0.5%, 1%, 2% or 4% sodium hypochlorite solution with one drop of Tween 20, for 15,30,45,60,75 or 90 min. In control, seeds were soaked only in water. After pretreatments, seeds were washed three times with sterilized distilled water, then inoculated on 1/2 MS medium as described above.
Histological and histochemical studies Developing seeds were collected and xed in 2.5% glutaraldehyde and 1.6% paraformaldehyde buffered with 0.1M phosphate buffer (pH 6.8) for 48 h at room temperature. Seeds were then dehydrated in an ethanol series, then in ltrated gradually (3:1, 1:1, and 1:3 100% ethanol: Technovit 7100, 24 hr each) by using Technovit 7100 resin (Kulzer & Co., Germany), followed by three changes of pure resin. Seeds were then embedded in resin, as described by Yeung and Chan (2015). Sections of 3-µm thick were cut using a Reichert-Jung 2040 Autocut rotary microtome. These sections were stained with periodic acid-Schiff (PAS) procedure for structural carbohydrates and counterstained with 1% (w/v) amido black 10B in 7% acetic acid for protein (Sigma-Aldrich, St. Louis, MO, USA) or 0.05% (w/v) toluidine blue O (TBO, Sigma-Aldrich) for general histological staining (Yeung, 1984). For detecting the deposition of cuticular material in developing seeds, sections were stained with 1 µg ml − 1 Nile red (Sigma-Aldrich) for 1 min, then washed in running tap water for 3 min. The uorescence pattern of Nile red was viewed under an epi uorescence microscope (Axioskop 2, Carl Zeiss AG, Germany) equipped with the Zeiss lter set 15 (546/12 nm excitation and 590 emission). All images were recorded by using a CCD camera attached to the microscope.

Rooting and acclimatization of in vitro seedlings
After 150 days of culture, developing protocorms with roots were transferred onto seedling growth medium: 1/2 MS medium supplemented with 20 g l − 1 sucrose, 1 g l − 1 activated charcoal, 20 g l − 1 potato homogenate, and 7 g l − 1 agar for growing seedlings described by Lee (2011). The potato was boiled for 10 min, then peeled and cut into about 1-cm 3 sections, then homogenized with a kitchen blender. The pH of the medium was adjusted to 5.6 before autoclaving at 121 °C for 20 min. An amount of 100 ml medium was dispensed into a 500-mL culture ask. After transferring to the seedling growth medium, asks were placed in a growth room under a 16/8 hours photoperiod with daylight uorescent lamps (20W, China Electric, Taipei) at light intensity 30 µmol m − 2 s − 1 . After 90 days of culture in the seedling growth medium, seedlings of about 10 cm tall with 4 leaves were taken out of asks.

Experimental design and statistical analysis
All experiments were performed in a completely randomized design and repeated three times; 12 replicates (culture tubes) were used for each treatment, with one explant planted in each plate. Data were analyzed by using analysis of variance (ANOVA) with Fisher's protected least signi cant difference test at P < 0.05. All data were analyzed with SAS v9.0 (Cary, NC, USA).

Results
Seed germination can be achieved using asymbiotic germination technique at early seed developmental stages of V. planifolia DAP and their germination percentages under the asymbiotic germination treatments were investigated. Seed germination occurred after 30 DAP (Fig. 2). The seed germination percentage reached 9.9% at 45 DAP, the highest germination percentage through the whole experiment. The germination percentage then markedly declined to approximately 2% at 60 DAP. This low germination percentage lasted to 90 DAP. Almost no seed germination was observed after 120 DAP up to 300 DAP (data not shown).

Effect of seed pretreatments on germination
Pretreatment with sodium hypochlorite solutions remarkably breaks the situation of no germination of V. planifolia mature seeds under the asymbiotic germination treatments. Seeds collected at 135 DAP were soaked with a 2% sodium hypochlorite solution over 60 min before the asymbiotic germination cultures. The pretreatment with sodium hypochlorite solutions resulted in over 5% seed germination (Fig. 3). The effect of sodium hypochlorite soaking treatment on V. planifolia seed germination increased with higher strength of sodium hypochlorite and longer soaking time. However, the effect of soaking time reached a maximum at 75 minutes in all of the different concentrations of sodium hypochlorite solutions (Fig. 3). In addition, V. planifolia mature seeds treated with a 4% sodium hypochlorite solution for 75 minutes can reach over 10% seed germination. A higher concentration of sodium hypochlorite solution treatment is not recommended because a high concentration of sodium hypochlorite may destroy cell membranes and cause seed death.

Development of pod, embryo and seed coat
To understand the possible causes for the effect of pretreatment of sodium hypochlorite, morphological, histological, and histochemical changes of seed development were investigated. The main structural changes occurring within developing pods from anthesis until maturity are summarized in Table 1.
Detailed characteristics are described in follows.  4). By 240 DAP, the pod matured and became yellow, then turned black after 300 DAP (Fig. 1D).
Embryos in a pod were developed from zygotes that were combined between sperm cells in pollens and eggs in embryo sacs. The megaspore mother cells develop just after pollination in early-mid May. Numerous mature embryo sacs could be observed in a pod before fertilization. At 30 DAP, zygotes and proembryos were present within developing pods, and no endosperm was observed (Figs. 5B and C). At this stage, the seeds were white and moist (Fig. 6A). At 45 DAP, additional cell divisions occurred within the inner tiers and the surface layer (Figs. 5D and E), thus resulting in the growth of the embryo proper. Some seeds had turned light to dark brown (Fig. 6B). At 60 DAP, more cell divisions occurred within the embryo proper, resulting in an early globular embryo (Fig. 5F). At this stage, most seeds had turned black (Fig. 6C). This species lacked a structurally de ned suspensor during embryo development (Figs. 5 and 7). As the embryo developed to the globular stage at 75 DAP, the embryo proper had lled the cavity of the embryo sac (Fig. 7A). After this stage, nearly all seeds were black (Figs. 6D, E, and F). By 105 DAP, the mature embryo was about 11 cells long and 7 cells wide without the formation of shoot apical meristem and cotyledon (Fig. 7B). At this stage, the cytoplasm of the embryo proper cell was lled with a large number of storage products (e.g., protein bodies and lipid bodies), and starch grains had disappeared. By 300 DAP, the pod had fully matured and desiccated. Then, the pod split, and the mature seeds were released (Fig. 1D).
Regarding the development of seed coat, the inner and outer seed coats derived from the inner and outer integuments of the ovule, respectively (Fig. 5A), and these two distinct layers of seed coat surrounded the embryo in the mature seed of V. planifolia (Fig. 7B). The inner seed coat was two cells thick, and the cell wall of the inner seed coat remained primary in nature during the early stage of seed development (Figs. 5B and C; Fig. 8A). As the seeds approached maturity, the inner seed coat gradually compressed (Figs. 5D, E and F; Fig. 8B), and eventually became a thin layer at maturity (Fig. 8C). However, the outer seed coat consisted of three to four cell layers (Figs. 5B and C). Before fertilization, the outer seed coat was still growing and had not enclosed the embryo sac completely (Fig. 8A). At this stage, the cell wall of the outer seed coat was relatively thin (Fig. 8A), and the cell wall of the outermost layer of the outer seed coat became thickened after fertilization (Fig. 5B). By 60 DAP, the thickened wall of the outermost layer of the outer seed coat became scleri ed (Figs. 5F and 8B). As the seed approached maturity, the inner layers of the outer seed coat gradually compressed and attached to the scleri ed outermost layer of the outer seed coat (Figs. 8C and D). Using TBO staining, the cell wall of the outermost layer of the outer seed coat stained greenish-blue, indicating the presence of phenolic compounds in the cell wall (Figs. 8B and C). In addition, using Nile red staining, the surface wall of the embryo proper and the innermost and outermost walls of the inner seed coat reacted positively, which suggested the possible accumulation of a cuticular substance in the wall of these two layers (Figs. 7C and D).

Protocorm and seedling growth
For the development of protocorms, most embryos were still enveloped by the seed coat at 30 days of culture on 1/2 MS medium. Only a few embryos had emerged from the seed coat, which was considered seed germination (Fig. 9A). By 60 days of culture, young protocorms had turned pale green, and numerous rhizoids appeared at the basal protocorms (Fig. 9B). By 90 days of culture, the protocorm had enlarged and turned dark green to differentiate shoot apical meristem (Fig. 9C). Subsequently, the developing protocorm with the differentiation of the rst root was observed (Fig. 9D). By 150 days of culture, the protocorm had further elongated with prominent root formation (Fig. 9E). After transferring onto the seedling growth medium for additional 90 days, the seedlings grew into vines with several healthy roots that could be taken out of asks (Fig. 9F).

Discussion
The mature seed of V. planifolia is black and hard, which is distinct from the seeds of many orchids (Arditti and Ghani 2000). In the present study, based on a de ned time frame, we investigated the morphological and structural changes of seeds (Table 1; Figs. 5, 6 and 7). In many orchids, the outer seed coat develops from two cell layers (Yam et al. 2002), whereas in V. planifolia, the outer seed coat is derived from four cell layers (Figs. 5 and 8). The histological and histochemical results indicated that the scleri cation primarily occurred in the outermost cell layer of the outer seed coat (Figs. 5, 7 and 8). Similar hard seeds can be found in a few vanilloid species, e.g., Galeola septentrionalis (Suetsugu et al. 2015) and G. javanica (Yang and Lee 2014). Their seed coats are multi-layer, and the walls are heavily thickened with lignin polymers as seeds mature. The cell wall of the outermost cell layer of the outer seed coat stained greenish-blue by TBO indicated the presence of phenolic compounds (Fig. 8). In the seed coat of many plants, phenolic compounds, including caffeic acid, sinapic acid, coumarin, chlorogenic acid, and ferulic acid, are esteri ed to the wall structure (Gubler and Ashford 1985;Pan et al. 2002). The accumulation of phenolic compounds in the seed coat inhibits seed germination (Bewley and Black 1994). In V. planifolia, further analysis by nuclear magnetic resonance spectroscopy revealed the heavy deposition of catechyl units during ligni cation of the cell wall of the seed coat (Chen et al. 2012).
Also, the presence of a cuticular substance or suberin has been detected in the cell wall of the inner seed coat (also known as a carapace) and/or outer seed coat of some orchids, such as Cephalanthera, Cymbidium, and Cypripedium (Carlson 1940;Yeung et al. 1996;Lee et al. 2005;Yamazaki and Miyoshi 2006). In V. planifolia, after Nile red staining, a thin uorescent layer was observed in the innermost layer of the inner seed coat cells and the surface wall of the globular embryo, which suggests the accumulation of cuticular substance in walls. However, the uorescence pattern was absent in the walls of the outer seed coat (Figs. 7C and D). In orchids, the deposition of cuticular substance on the surface of the embryo proper and in the innermost layer of the seed coat may provide physical protection and ensure moisture retention to developing embryos.
During the same time, we investigated the effect of seed maturity and seed pretreatments on asymbiotic germination. The optimal germination was at 45 DAP ( Fig. 3). At this stage, the early globular embryo appeared, and the outermost cell layer of the outer seed coat had become thickened but had not been heavily ligni ed (Fig. 5D). The germination decreased sharply by 60 DAP, which agreed with the beginning of seed coat ligni cation (Fig. 3). Because the outermost cell layer of the outer seedcoat had been ligni ed, a stronger hydrophobic nature was observed during the operation of seed sowing. As the seed matured, the outer seed coat gradually compressed into a thick and heavily ligni ed layer surrounding the embryo (Fig. 8). The low permeable seed coat restricted the water uptake and solute diffusion, which resulted in poor germination. Although V. planifolia is an epiphytic orchid species and pantropical in distribution, this pattern of seed germination is similar to those of terrestrial orchids from temperate regions, where immature seeds are easier to germinate in vitro than are mature seeds (Linden 1980;Zhang et al. 2013).
The poor germination of mature seeds in terrestrial orchids from temperate regions may be attributed to the accumulation of chemical inhibitors to germination, such as abscisic acid in Dactylorhyza and Epipactis (van der Kinderen 1987), Calanthe (Lee et al. 2007) and Cypripedium (Lee et al. 2015) and phenolics in Cymbidium (Kako 1976), and/or the formation of an impermeable container in the seed coat that may make an embryo di cult to obtain water and nutrients for germination (Miyoshi and Mii 1988;Lee et al. 2005). Seed pretreatments may release dormancy and enhance germination by changing the physical integrity of seed coverings, allowing the embryo to absorb water and nutrients and to uptake oxygen (Taylor et al. 1998). Seed pretreatments using bleaching solutions, such as calcium hypochlorite and sodium hypochlorite, can greatly stimulate seed germination of several European terrestrial orchids (Linden 1980;Rasmussen 1995;Steele 1996;Van Waes and Debergh 1986). In this study, the germination of mature seeds was enhanced with soaking in 2% and 4% sodium hypochlorite solutions from 75 to 90 min. However, no germination was recorded in fully mature seeds (Fig. 4). Common bleaching agents such as sodium hypochlorite have been widely used for removing residual lignin from the wood pulp (Holik 2006). From the scanning electron microscopic observations, the seed coat was scarred with hypochlorite oxidation, so the seed coat was relatively more permeable to water (Lee 2011). Of note, as compared with previous reports of seed pretreatments using hypochlorite solutions, we used a stronger concentration (4% sodium hypochlorite) and a longer duration (90 min) to pretreat mature seeds. However, the optimal germination of 12.66 ± 1.14% was relatively low. This low germination may re ect the hard seed coat of V. planifolia with its extremely impermeable nature. The relative low germination of pretreated seeds may be attributed to the accumulation of inhibitory substances, such as abscisic acid inside the embryo, which still cannot be leached out by hypochlorite solutions. The orchid seed with a heavily ligni ed, hard seed coat is adapted to the seed dispersal strategy involved in endozoochory by animals (Suetsugu 2018). In Cyrtosia and Vanilla species, as their fruits ripen, they usually have vivid colors and/or a heady fragrance to attract animals, such as rats, bats, or birds (Soto Arenas and Dressler

Conclusions
In this report, we provide reliable protocols for seedling production of V. planifolia through asymbiotic seed germination. For the successful culture of immature seeds, the timing of seed collection is short and critical, and optimal germination could be obtained from immature seeds collected at 45 DAP (before seeds turned black). For improving the germination of mature seeds, pretreatment with 4% sodium hypochlorite solution for 75 to 90 min is recommended. We propose that the thickened and ligni ed seed coat, forming an impermeable envelope, is responsible for the seed coat-imposed dormancy. Further work is needed to examine the changes in levels of endogenous inhibitors in embryos, such as abscisic acid, during seed development. Such studies should help to fully explain the low germination percentage of V. planifolia seeds.

Declarations
Author contributions CHY, KYC and YIL designed the study; CHY and YIL performed experiments; CHY, KYC and YIL wrote and edited the manuscript; all authors commented on the manuscript. showing a longitudinal section through a globular embryo at 60 DAP. The outer most layer of outer seed coat (OS) became ligni ed. Scale bar = 100 μm.

Figure 7
The late embryo development of V. planifolia. (A) As the seed approached maturity, a number of tiny protein bodies (arrows) appeared within the embryo proper cells after amido black 10B stain. The thickened outer seed coat (OS) became dehydrated and compressed, with the inner seed coat (IS) compressed into a thin layer. (B) Light micrograph showing a longitudinal section through a mature seed. Several tiny protein bodies (arrows) were found within the embryo proper cells. In this preparation, the lipid bodies were not preserved; the spaces (arrowheads) between the protein bodies were occupied by storage lipid bodies. (C) Nile red staining uorescence micrograph of an early globular embryo at the stage similar to Figure 3E. After Nile red staining, the surface wall of the embryo proper (arrows) reacted weakly to the stain, and the innermost (arrowhead) and outermost (double arrowhead) walls of the inner seed coat also reacted positively. (D) Nile red staining uorescence micrograph of a mature seed at the stage similar to Figure 5B. The inner seed coat compressed into a thin layer and attached the embryo tightly. The thin inner seed coat (IS, arrowheads) and the surface of the embryo proper (arrow) reacted positively to the stain. Fluorescence was never observed in the thickened outer seed coat (OS). Scale bar = 100 μm.