Introduction

Cibotium barometz (L.) J. Sm. is a tree fern listed in Appendix II of the Convention for International Trade on Endangered Species of Wild Fauna and Flora (CITES 2015). C. barometz is the only Dicksoniaceae family member found in mainland China, it is considered as an endangered species (Zhang et al. 2002). However, the species has varied utilities in China. It is not only an ornamental plant because of its attractive fibrous trunk covered with special golden hairs, but it is also a medicinal plant in traditional Chinese medicine as the rhizome shows anti-inflammation, pain relief and anti-osteoporosis activities (Cuong et al. 2009; Wu and Yang 2009; Zhao et al. 2011). Unfortunately, due to its overexploitation and the destruction of its natural habitats in recent years, the species is under severe threat and requires urgent protection (Zhang et al. 2002). Therefore, there is an urgent need to produce high quality plantlets for horticultural, medicinal, and conservation uses without resourcing to harvest from the remaining wild population.

Conventional propagation of ferns from spores is generally a time-consuming procedure (Liao and Wu 2011). At the same time, the low multiplication rate makes the vegetative propagation also inefficient (Thakur et al. 1998). However, a variety of in vitro vegetative propagation methods have been successfully practiced in many fern species through shoot organogenesis from rhizomes (Fernández et al. 1997; Winarto and Teixeira da Silva 2012), juvenile leaves (Camloha et al. 1994), homogenized sporophytic tissues (Teng and Teng 1997), bud scales (Ambrožič-Dolinšek and Camloh 1997), callus (Hegde et al. 2006), somatic embryos (Mikuła et al. 2015a, b) or green globular bodies (GGBs) (Amaki and Higuchi 1991). Among these methods, in vitro propagation through GGBs is particularly efficient (Higuchi et al. 1987; Higuchi and Amaki 1989).

Higuchi et al. (1987) first named rapidly proliferating tissues as GGBs, which were inducted from the runner tips of Nephrolepis cordifolia cultured on medium containing N6-benzylaminopurine (BA), and converted to plantlets easily when cultured on plant growth regulator (PGR)-free medium. GGBs were successfully inducted in Asplenium nidus (Higuchi and Amaki 1989), Pteris ensiformis, Rumohra adiantiformis, Adiantum raddianum, (Amaki and Higuchi 1991), Blechnum spicant (Fernández et al. 1996) and Polypodium cambricum (Bertrand et al. 1999) using different sporophytic explants. Cytokinin is generally regarded as an important PGR for in vitro propagation through GGBs, since it promotes GGB formation while also inhibiting plantlet regeneration from GGBs. This allows for tight organogenesis control during in vitro propagation via the regulation of cytokinin content (Amaki and Higuchi 1991). GGBs are different from callus, but similar to protocorm-like bodies (PLBs) in orchids (Higuchi et al. 1987). Interestingly, GGBs from different species show distinctive external morphological features such as scale- or hair-like structures (Amaki and Higuchi 1991; Higuchi et al. 1987). However, all GGBs possess multiple meristems inside or on their periphery. Some meristems come to the surface and develop into new GGBs, which results in a high multiplication rate (Liao and Wu 2011). Thus, GGBs is considered as an important material for in vitro vegetative propagation of ferns. In addition, GGBs can also serve as plant materials for cryopreservation (Li et al. 2013; Pence 2015) or the research of somatic embryogenesis (Li et al. 2015).

There are only few studies on in vitro propagation of C. barometz and limited descriptions of fern GGBs. Goller and Rybczyński (2007) reported in vitro culture of C. glaucum and C. schiedei using spores as explants, but sporophytes were only obtained in C. schiedei. Until now, GGB induction or other in vitro vegetative propagation methods for genus Cibotium have never been reported, because of the difficulty in obtaining proper explants from tree ferns. Here, we describe an in vitro propagation protocol for C. barometz through formation of GGBs using in vitro juvenile sporophytes as explants. Furthermore, we describe the morphology and histology of C. barometz GGBs. Our new protocol is set to largely improve the propagation of C. barometz.

Materials and methods

Plant material

C. barometz spores were obtained from mature fronds collected in Pingbian (Yunnan province, China). The fronds were wrapped in paper bags and dried at room temperature for 1 week to release spores, and spores were stored at 4 °C in dark until use. Spores (10 mg) were soaked with distilled water for 1 h, then surface-sterilized for 5 min with 0.5 % sodium hypochlorite. Afterward, they were rinsed five times with sterile distilled water by centrifugation (3000 rpm; 2 min) and resuspension. Finally, the spores were suspended in 10 mL of sterile distilled water and 2 mL of spore suspension was inoculated in a conical flask (100 mL volume), which contained 40 mL 1/2 MS (Murashige and Skoog 1962) medium supplemented with 3 % (w/v) sucrose and 0.7 % (w/v) plant agar, and adjusted to pH 5.8 before being autoclaved. Cultures were maintained under the light intensity of 40 μmol m−2 s−1 provided by cool-white fluorescent lights (photoperiod 16 h light/8 h dark) in a controlled environment room at 25 ± 2 °C. Gametophytes that developed from spores were periodically watered to induce the sporophyte formation. In vitro juvenile sporophytes with 2–3 fronds (about 5.0 mm in length of the first frond) were used for the study.

GGB induction from juvenile sporophytes

For GGB induction, in vitro juvenile sporophytes used as explants were cultured on 1/2 MS supplemented with PGRs. Using PGR-free 1/2 MS medium as a control, the effect of two different PGRs was evaluated together at different concentrations: thidiazuron (TDZ) (0.2, 0.5 or 1.0 mg L−1) and a-naphthaleneacetic acid (NAA) (0.1, 0.3 or 0.5 mg L−1). After 4 weeks of preculture, shoot tips of juvenile sporophytes became swollen and formed primordia in PGR treatments, but not in the control. Then primordia (diameter ≥2.0 mm) were excised from juvenile sporophytes and subcultured on the same medium. Every 2 days, GGB formation was observed to calculate the time for GGB formation from primordia. After 8 weeks of subculture, GGB induction rate was recorded. Cultures were incubated under the conditions described above.

GGB multiplication

To determine the effect of PGRs on GGB multiplication, GGBs (diameter ≥ 5.0 mm) induced on 1/2 MS supplemented with 1.0 mg L−1 TDZ and 0.3 mg L−1 NAA, were cut into segments with the diameter of 2.0 mm, which the initial fresh weight was 4.0 mg. Each segment was cultured on 1/2 MS supplemented with TDZ (0.5, 1.0 or 1.5 mg L−1) and NAA (0.1, 0.3 or 0.5 mg L−1) in combination. In all sets of experiments, PGR-free 1/2 MS medium was used as a control. Cultures were incubated under the conditions described above. After 4 weeks of culture, GGB final fresh weight and plantlet regeneration from GGBs were assessed.

Plantlet regeneration from GGBs and acclimatization

In order to regenerate plantlets from GGBs, GGBs (diameter ≥ 5.0 mm) from 1/2 MS supplemented with 1.0 mg L−1 TDZ and 0.1 mg L−1 NAA after 4 weeks of multiplication culture, were cut into segments with the diameter of 2.0 mm and cultured on PGR-free media consisting of different mineral salt concentrations (1/4 MS, 1/2 MS or MS), along with various concentrations of activated charcoal (AC) (0, 0.1 % or 0.2 %, w/v). Cultures were incubated under the conditions described above. After 8 weeks of culture, the rate of GGB development into plantlets was assessed. Subsequently, plantlets regenerated from GGBs were transferred to the same fresh medium, and plantlet height was recorded after 8 weeks of plantlet cultivation.

For plantlet acclimatization, regenerated plantlets higher than 2.5 cm and with well-developed roots were gently washed in 0.100–0.125 % (w/v) chlorothalonil solution and transferred to plastic boxes with covers, which contained a 5:1 (v/v) mixture of sterilized peat and perlite. After 5 weeks, covers of plastic boxes were gradually opened for several days, then plantlets were moved to plastic pots containing a 5:1 (v/v) mixture of peat and perlite and irrigated with tapwater every 4–5 days. These plantlets were maintained in the greenhouse with the shade (70 %) at 25 ± 5 °C under natural photoperiod conditions.

Morphological and histological observation of GGBs

For morphological observation of GGBs, samples were collected in GGB multiplication and differentiation stages and dissected under a stereoscope (Leica MZ16, Germany) and photographed with a camera (Leica DFC 450C, Germany).

To observe the histology of GGBs, the same samples were collected again, and immediately fixed in FAA (70 % ethanol: glacial acetic acid: formalin, 18:1:1) for 24 h, then dehydrated through a graded series of ethanol solutions, and embedded in paraffin wax. The samples were sectioned to a thickness of 8–10 μm and stained with hematoxylin (Zhou et al. 2013), then observed under the light microscope (Leica DM 6000B, Germany) and photographed with a camera (Leica DFC 450C, Germany).

Scanning electron microscopy

To observe the surface structure of GGB, samples collected in GGB multiplication and differentiation stage were fixed in FAA. After dehydration through a series of ethanol rinses, samples were critical-point dried and coated with molybdenum-palladium alloy using an ion-sputter apparatus (Hitachi E-1010, Japan), and imaged under a scanning electron microscope (Hitachi S-3000N, Japan) with an accelerating voltage of 15.0 kV.

Data analysis

Experiments were performed in a completely randomized design. Experiments of GGB induction, GGB multiplication and plantlet regeneration were repeated three times, using five replicates each containing six explants. The data were analyzed by means of ANOVA analysis of variance and least significant difference (LSD) procedure using the software SPSS V16.0 for Windows (SPSS Inc., Chicago, IL, USA). Significance was set at the 0.05 level.

Results

GGB induction from juvenile sporophytes

We tested the influence of PGRs on GGB induction using PGR-free medium as a control. GGB induction in C. barometz is highly dependent on the presence of PGRs. With the exception of the control, all tested treatments had primordium formation from juvenile sporophytes after 4 weeks of preculture (Fig. 1a), and then got GGB formation from primordia. Table 1 summarizes the effect of PGRs on GGB induction. Explants cultured on the medium supplemented with 1.0 mg L−1 TDZ and 0.3 mg L−1 NAA resulted in the highest GGB induction rate (86.67 %) and the shortest time for GGB formation from primordia (32.39 days). TDZ had a significant effect on GGB induction rate and the time for GGB formation from primordia (P < 0.001). With the increase of TDZ concentration, GGB induction rate increased, and the time for GGB formation from primordia decreased. At the highest TDZ concentration (1.0 mg L−1), GGB induction rate was above 80 %, and the time for GGB formation from primordia was between 32 and 35 days. On the contrary, NAA had no significant effect on GGB induction (P > 0.05). Moreover, there was no statistically significant interaction between the concentration of TDZ and NAA (P > 0.05).

Fig. 1
figure 1

In vitro propagation of Cibotium barometz through formation of GGBs. a Arrow indicates primordium formation from an in vitro grown juvenile sporophyte after 4 weeks of preculture on 1/2 MS medium containing 1.0 mg L−1 TDZ and 0.3 mg L−1 NAA, bar = 1 mm. b GGB multiplication on 1/2 MS medium containing 1.0 mg L−1 TDZ and 0.1 mg L−1NAA after 4 weeks of culture, bar = 1 cm. c GGB differentiation on 1/4 MS medium containing 0.1 % (w/v) active charcoal (AC), bar = 1 cm. d Sporophyte plantlets developed from GGBs on 1/2 MS medium containing 0.1 % (w/v) AC, bar = 1 cm. e Formation of aposporous gametophyte on the root of a plantlet cultured on 1/2 MS medium containing 3 % (w/v) sucrose without AC, arrow indicates aposporous gametophyte, bar = 1 mm. f Acclimatized plant in the greenhouse after 5 months, bar 10 cm. g A crown covered with golden hairs of acclimatized plant, bar = 5 mm

Full size image
Table 1 Influence of plant growth regulators on GGB induction in Cibotium barometz
Full size table

GGB multiplication

With the multiplication, GGBs increased in size and fresh weight, and became loose and easy to be separated from each other (Fig. 1b). Table 2 reports the effect of TDZ and NAA concentrations on GGB multiplication. GGBs on PGR-free medium produced the highest GGB final fresh weight (72.14 mg), but also result in the highest rate of plantlet regeneration from GGBs (67.78 %). As a result, GGB morphology changed, which was not ideal for GGB multiplication. TDZ had a significant effect on GGB final fresh weight and plantlet regeneration from GGBs (P < 0.001). With the increase of TDZ concentration, GGB final fresh weight decreased. Furthermore, the rate of plantlet regeneration from GGBs also decreased, which was helpful to maintain GGB morphology. However, even at the highest TDZ concentration (1.5 mg L−1), plantlet regeneration from GGBs could not be inhibited completely. In consideration of GGB final fresh weight as well as inhibition of plantlet regeneration from GGBs, 1.0 mg L−1 TDZ were the optimal concentration for GGB multiplication, with GGB final fresh weight above 42 mg and a low rate of plantlet regeneration from GGBs. We observed no significant effect of NAA concentration on GGB final fresh weight and plantlet regeneration from GGBs (P > 0.05), and there was no statistically significant interaction between the concentrations of TDZ and NAA (P > 0.05).

Table 2 Influence of plant growth regulators on GGB multiplication in Cibotium barometz
Full size table

Plantlet regeneration from GGBs and acclimatization

In order to obtain plantlets, GGBs were cultivated on PGR-free medium. The effect of mineral salt concentration and AC on GGB development and plantlet growth was investigated (Table 3). Plantlet formation was observed from the fourth week of culture in most of treatments (Fig. 1c). After 8 weeks of culture, the rate of GGB development into plantlets was above 65 % for the all treatments, and it was significantly affected by mineral salt concentration (P < 0.001) and AC (P < 0.05). Compared with other treatments, 1/4 MS media supplemented with 0.1 or 0.2 % (w/v) AC produced a higher rate of GGB development into plantlets, above 90 %. Moreover, during plantlet cultivation, mineral salt concentration (P < 0.001) and AC (P < 0.05) also significantly affected plantlet height. Plantlets on 1/2 MS media supplemented with 0.1 or 0.2 % (w/v) AC were higher than those on other treatments, with plantlet height above 3.2 cm after 8 weeks of plantlet cultivation (Fig. 1d). However, there was no statistically significant interaction between mineral salt concentration and AC on the rate of GGB development into plantlets and plantlet height (P > 0.05). In addition, aposporous gametophytes were observed on the roots of 4 plantlets cultured on 1/2 MS containing 3 % (w/v) sucrose without AC during plantlet cultivation (Fig. 1e).

Table 3 Influence of mineral salt concentration and AC on GGB differentiation and plantlet growth in Cibotium barometz
Full size table

Plantlets from all treatments had roots. Regenerated plantlets were successfully acclimatized (80 %) in plastic pots containing a 5:1 (v/v) mixture of peat and perlite under greenhouse conditions. Surviving plants showed a normal phenotype (Fig. 1f). Particularly, a small crown covered with golden hairs appeared at the apex after 5 months of acclimatization (Fig. 1g).

Morphological and histological observation of GGBs

Top view showed that GGBs during multiplication was a yellow-green globular structure, which was composed of the single GGB (Fig. 2a). From the bottom view, single GGBs were connected to each other, and the brown root-like structures were gathered in the center of GGBs (Fig. 2b). To some degree, a single GGB had two discrete bipolar structures: shoot meristem on the top, as well as root meristem on the bottom that was always attached to the brown root-like structures (Fig. 2c). During differentiation, shoot meristem on the top of the single GGB developed into young fronds (Fig. 2d). At the same time, root meristem on the bottom of the single GGB formed more roots or root-like structures (Fig. 2d). As a result, one single GGB developed into one plantlet (Fig. 2d). Protrusions were present on the surface of GGBs during differentiation, due to the formation of young fronds (Fig. 2e). Finally, one cluster of GGBs developed into one cluster of plantlets (Fig. 2f).

Fig. 2
figure 2

Morphological observation of Cibotium barometz GGBs in the multiplication and differentiation stage. a–c GGBs or a single GGB in the multiplication stage. a GGB morphology from top view, arrow indicates a single GGB. b GGB morphology from bottom view, arrow indicates brown root-like tissues. c Single GGB morphology, arrow indicates brown root-like tissue. d–f GGBs or single GGB development into sporophyte plantlet. d One single GGB development into one plantlet, arrow indicates young fronds. e Young frond formation on GGBs. f GGBs development into a cluster of plantlets. bar 1 mm

Full size image

Histological analysis of GGB during multiplication revealed that there were meristems on the top of the single GGB (Fig. 3a). Compared with other cells, meristem cells were cytoplasmically denser (Fig. 3b), and the epidermal cells of meristem zone were bigger in size and nucleus (Fig. 3b). As the epidermal cells divided, dome-shaped protrusions formed on the surface of the single mother GGB (Fig. 3b, c), then developed into the single daughter GGB (Fig. 3c). Interestingly, the single daughter GGB, as well as the single mother GGB produced new single GGBs at the same time (Fig. 3c). As a result, GGBs tended to multiply fast. The single GGB during multiplication had flat or dome-shape top surface (Fig. 3a), while the single GGB during differentiation extended outward (Fig. 3d), which indicated that the single GGB turned into shoot organogenesis.

Fig. 3
figure 3

Histological analysis of Cibotium barometz GGBs in the multiplication and differentiation stage. a–c Single GGBs in the multiplication stage. a Arrow indicates meristems with dense cytoplasm on the top of the single GGB, bar 100 μm. b Arrow indicates protrusion formation from epidermal cells, that gives rise to the new single GGB, bar 100 μm. c Single mother GGB (M) and single daughter GGB (D) produce a new single GGB at the same time, arrow indicates the formation of a new single GGB, bar 50 μm. d Single GGB in the differentiation stage, arrow indicates that the top surface of a single GGB extends outward, bar 50 μm

Full size image

Scanning electron microscopy

We evaluated GGB surface structure during multiplication and differentiation using scanning electron microscopy imaging. The single GGB was covered with hair-like structures (Fig. 4a), but few hair-like structures were observed on the protrusions (Fig. 4b) and new single GGBs (Fig. 4c). Protrusions were observed on the surface of GGBs during multiplication (Fig. 4b). These protrusions gave rise to new single GGBs (Fig. 4c). During differentiation, sheet-like structures, likely early embryonic leaves, appeared on the top of the single GGB (Fig. 4d, e). In addition, new single GGBs were also found in the differentiation stage, but their top surface was slightly pointier than new single GGBs in the multiplication stage (Fig. 4f), indicating that new single GGBs formed in the differentiation stage probably differentiated directly instead of multiplication.

Fig. 4
figure 4

Scanning electron microscopy of Cibotium barometz GGBs in the multiplication and differentiation stage. a–c Single GGBs in the multiplication stage. a Single GGB covered with hair-like structures, bar 100 μm. b Arrow indicates protrusion on the surface of the single GGB, that gives rise to a new single GGB, bar 100 μm. c Arrow indicates numbers of new single GGBs on the surface, bar 100 μm. d–f Single GGBs in the differentiation stage. d Arrow indicates the formation of sheet-like structures, bar 100 μm. e Arrow indicates sheet-like structures on the top of the single GGB, bar 50 μm. f Morphology of a new single GGB in the differentiation stage, arrow indicates that its top surface is slightly pointy, bar 50 μm

Full size image

Discussion

In this study, we used in vitro juvenile sporophytes as explants for GGB induction in C. barometz, because of the difficulty in getting proper explants from tree fern individuals. With this method, GGB induction rate was above 45 % except the control (Table 1). Similarly, in vitro juvenile sporophytes were also used to initiate somatic embryo formation in the tree fern Cyathea delgadii (Mikuła et al. 2015a, b). Therefore, in vitro juvenile sporophytes are potential explants for in vitro vegetative propagation of tree ferns, especially those with in vitro propagation protocols from spores.

GGB induction in C. barometz was significantly dependent on TDZ, which promoted GGB induction (Table 1). Cytokinins, such as BA and TDZ, are often used to induce adventitious buds in many spermatophytes (Skała et al. 2015; Zhang and Finer 2015), and PLBs in orchids (Malabadi et al. 2004; Ng and Saleh 2011; Park et al. 2003). BA is also optimal for GGB induction in some herbaceous fern species, such as Nephrolepis cordifolia (Higuchi et al. 1987), and Asplenium nidus (Higuchi and Amaki 1989). However, some species seem to behave differently. For example, increasing the concentration of cytokinin decreased GGB formation in Platycerium bifurcatum, while the medium supplemented with 5.37 μM NAA alone displayed the maximum GGB induction (Liao and Wu 2011). In Blechnum spicant, rhizomes of juvenile sporophytes were cultivated on the medium with BA for one month preculture. Subculture of these rhizomes on the same medium with BA induced the development of aposporous gametophytes. Conversely, subculture on the medium with NAA promoted GGB formation (Fernández et al. 1996). In our study, NAA had no significant effect on GGB induction in C. barometz, similarly to what reported for Nephrolepis cordifolia and Asplenium nidus (Higuchi and Amaki 1989; Higuchi et al. 1987).

Cytokinin BA controls the organogenesis of GGBs (Higuchi et al. 1987). BA inhibited the plantlet regeneration from GGBs in Asplenium nidus (Higuchi and Amaki 1989), Adiantum raddianum, Pteris ensiformis, and Rumohra adiantiformis (Amaki and Higuchi 1991). BA also decreased GGB final fresh weight in Asplenium nidus (Higuchi and Amaki 1989). Removal of BA from medium promotes GGB differentiation (Higuchi et al. 1987). The medium with PGR alone or combinations that supports rapid growth of GGBs without plantlet regeneration is optimal for GGB multiplication (Liao and Wu 2011). Our data obtained here show that TDZ plays a similar role in GGB multiplication of C. barometz as BA in other fern species. However, even at the highest TDZ concentration (1.5 mg L−1), plantlet regeneration from GGBs was not be inhibited completely (Table 2), indicating that it is difficult to absolutely control organogenesis of C. barometz GGBs by a single cytokinin at low concentration. Considering the balance between the inhibition of plantlet regeneration and GGB final fresh weight, we found the optimal TDZ concentration for GGB multiplication in C. barometz to be 1.0 mg L−1. Our results also showed that NAA had no significant effect on GGB multiplication in C. barometz. In contrast, Liao and Wu (2011) found that the combination of NAA and cytokinin N6-(2-Isopentenyl)adenine (2iP) was more suitable than using 2iP alone for GGB multiplication in Platycerium bifurcatum.

For plantlet regeneration of C. barometz, mineral salt concentration and AC had a significant effect on GGB differentiation and plantlet growth (Table 2). Mineral salt concentration is associated with plantlet regeneration in some fern species. For example, mineral salt concentration was associated with a greater frequency of abnormal regenerated sporophytes in Cyathea delgadii, 1/2 MS medium was needed to stimulate normal sporophytes from somatic embryos (Mikuła et al. 2015b). Similarly, 1/2 MS medium without sucrose favored leaf expansion of Pteris ensiformis (Fernández et al. 1999). In our study, 1/4 MS was an optimal for GGB development into plantlets in C. barometz, and 1/2 MS was best for plantlet growth. According to Thomas (2008), AC promotes a morphogenic responses in some ferns in vitro. In Platycerium bifurcatum and Matteuccia struthiopteris, AC enhanced sporophyte regeneration (Teng 1997; Thakur et al. 1998). Our results also confirmed that addition of AC was beneficial for GGB development and plantlet growth in C. barometz.

We observed aposporous gametophytes on the roots of plantlets cultured on 1/2 MS containing 3 % (w/v) sucrose without AC during plantlet cultivation (Fig. 1e). In other fern species, aposporous gametophytes were induced successfully, that influenced by various factors, such as PGRs and sucrose-limiting conditions (Ambrožič-Dolinšek et al. 2002; Fernández et al. 1996; Martin et al. 2006; Teng and Teng 1997). Here, we could not determine the reason for aposporous gametophyte formation in C. barometz, but it may be related to residual PGRs from GGB multiplication.

Higuchi et al. (1987) first reported GGB propagation system of Nephrolepis cordifolia, but with the limited morphological and histological description of GGBs. C. barometz GGBs during multiplication were composed of the single GGB with meristems on its top (Figs. 2a, 3a). New single GGBs were initiated from the epidermal cells of meristem zone (Fig. 3b), confirming previous observations in Asplenium nidus (Higuchi and Amaki 1989). Moreover, our results showed that the single daughter GGB and the single mother GGB were able to produce new single GGBs at the same time (Fig. 3c), explaining the efficient multiplication of GGBs. In general, GGBs of different fern species have distinctive features on their surface, such as scale-like structures in Nephrolepis cordifolia and Adiantum raddianum, hair-like structures in Pteris ensiformis and Rumohra adiantiformis (Amaki and Higuchi 1991). These same structures were absent in the somatic embryo of Cyathea delgadii, but trichomes were observed, which were smaller and shorter than hair-like structures (Mikuła et al. 2015a). In C. barometz, there were hair-like structures on GGB surfaces (Fig. 4a), but no scale-like structures. Additionally, sheet-like structures, likely early embryonic leaves, were found during GGB differentiation (Fig. 4d, e). With differentiation, one single GGB of C. barometz developed into one plantlet with fronds and roots (Fig. 2d). Similar observations have been obtained in the development of PLBs in orchids (Teixeira da Silva 2013; Ng and Saleh 2011), somatic embryos in Cyathea delgadii (Mikuła et al. 2015a, b) and buibils in some fern species, which are the small rounded leafless structures on the fronds (Jones 1987). Moreover, our observations showed that shoot meristem on the top of the single GGB developed into young fronds (Fig. 2d), and root meristem on the bottom of the single GGB formed roots or root-like structures (Fig. 2d), are similar to PLBs in orchids, but differ from zygotic or somatic embryogenesis that includes several development stages. To some degree, C. barometz GGBs could be regarded as a structure which is similar to PLBs in orchids.

Conclusion

This is the first report that describes an in vitro propagation protocol for the endangered tree fern C. barometz through formation of GGBs using in vitro juvenile sporophytes as explants. The results of this study demonstrate that GGB induction and multiplication in C. barometz are highly dependent on PGRs, especially on the cytokinin TDZ. GGB development and plantlet growth are affected by mineral salt concentration and AC. Moreover, our observation shows that C. barometz GGBs is a yellow-green globular structure composed of the single GGB with meristems and hair-like structures, and new single GGBs are initiated from the epidermal cells of meristem zone. This protocol described here could be used to facilitate the rapid propagation of the tree fern C. barometz for commercial and conservation purposes.