Materials and methods

Plant materials. Twenty C. lancifolius samples were collected from the Jishou City of Xiangxi Tujia and Miao Autonomous Prefecture in Hunan Province of China. After collection, fifteen plants were transplanted in the greenhouse of the Yunnan University of Chinese Medicine where these plants were cultivated a year. Then, stems with one or two nodes and leaves from healthy individuals were chosen as explants for propagation in vitro. The stems with one or two nodes and leaves were immersed into 5 % (w/v) mild detergent for 5 min, washed under running tap water for 30 min before shifted on a sterile operating platform. Furthermore, these materials were immersed into 75 % (v/v) ethyl alcohol for 10 s and surface sterilized using 0.1 % HgCl₂ (w/v) for 8 min (stem segments with nodes) and 3 min (leaves) successively and followed by treatment with sterile distilled water eight times that was more than 3 min every time.

Single factor experiment. The sterilized leaves and stems with one or two nodes were cultured in the MS medium containing different type and concentration plant growth regulators (PGRs) which included 2,4-dichlorophenoxy-acetic acid (2, 4-D) (0.05, 0.1, 0.5 and 1.0 mg·L^-1), 6-benzylaminopurine (6-BA) (0.1, 0.5, 1.0 and 2.0 mg·L^-1), kinetin (KT) (0.1, 0.5, 1.0 and 2.0 mg·L^-1) and α-naphthaleneacetic acid (NAA) (0.1, 0.5, 1.0 and 2.0 mg·L^-1). Every group of single factor experiment had 5 culture bottles with 2 leaves or stems. The growth conditions were recorded and the types and concentration ranges of PGRs suitable for the growth of C. lancifolius were selected.

Callus induction and regeneration. Based on the results of chapter “Single factor experiment”, L₉ (3⁴) orthogonal experiment was design via KT (A: 0.01, 0.05 and 0.1 mg·L^-1), 6-BA (B: 0.5, 1.0 and 1.5 mg·L^-1) and NAA (C: 0.1, 0.5 and 1.0 mg·L^-1) as factors. Some representative groups were selected from the comprehensive test and the results equivalent to the comprehensive tests were achieved with the minimum number tests (Table 1). Every group of callus induction and regeneration contained 5 culture bottles with 2 explants (aseptic leaves or stems obtained from single factor experiment). The rate of callus induction was calculated after 30 days, and the callus was evaluated by its color and texture.

Embryonic callus induction and bud differentiation. Although both explants (leaves and stems with one or two nodes) could induce callus in the L₉ (3⁴) orthogonal experiment, the proliferation coefficient of cluster buds failed to obtain satisfactory results in the subsequent experiment. Therefore, the MS medium containing different concentration TDZ (0.05, 0.1, 0.2 and 0.3 mg·L^-1) were used for inducting the embryonic callus. Every group of embryonic callus induction and bud differentiation contained 6 culture bottles with 5 explants (1×1 cm callus induced from stems with one or two nodes). The incidence of embryonic callus was calculated after 30 days.

Rooting and transplantation of the plantlets. In light of the results of PGRs selection, 3-4 cm individual bud excised from adventitious buds was cultured on the MS medium with the complete combination of NAA (0.5, 1.0 and 1.5 mg·L^-1) and 6-BA (0.05, 0.1 and 0.5 mg·L^-1). Every group of the rooting culture contained 6 culture bottles with 5 explants (3-4 cm individual bud excised from adventitious buds). The rooting rate was determined after 30 days.

The rooting plantlets of 5-6 cm long were removed from the culture room to adapt to the environment for 3 days and the agar on the roots was gently cleaned with water. Then, they were transplanted into peat matrix and controlled the temperature at 20-25 ºC and 75-95 % humidity. The survival rate was determined after 30 days.

Growth condition. All the media in the experiment contained 3 % sucrose and 0.47 % agar, and all the experiments were repeated three times.

The cultivation temperature was maintained between 21 and 23 ºC and 12 h photoperiod with 40-60 µmol m^-2 s^-1 irradiance provided by cool-white fluorescence tubes.

Statistical analysis. The calculation method of the data in each stage was as follows:

Rate of callus induction (%) = (the number of explants that produced callus / the number of explants) × 100 %. Occurrence rate of adventitious shoot (%) = (the number of callus that produced adventitious buds / the number of callus inoculations) × 100 %. Callus proliferation coefficient = the number of callus in subculture / the number of initial inoculations.

Cluster bud proliferation coefficient= the number of cluster buds in subculture / the number of initial inoculations. The rooting rate (%) = (the number of rooted seedlings / the number of seedlings inoculations) × 100 %. The survival rate (%) = (the number of survival seedlings / the number of transplantations) × 100 %

The results obtained from every section were subjected to analysis of variance (ANOVA) using SPSS software (IBM Corp, Armonk, USA). The minor difference among every treatment method was determined using Least Significant Differences Test at 5 % probability (P ≤ 0.05), the mean values were further separated using analysis of range.

Results

Single factor experiment. In the MS medium containing 6-BA or 2, 4-D, all explants (stem segments with nodes and leaves) could induce callus, but most of the texture was loose and fragile. Among them, the most excellent callus was observed in MS medium with 0.5 mg·L^-1 2, 4-D (Figure 1A-B), whereas in concentration of 2, 4-D higher than 0.5 mg·L^-1, all the explants died. Besides, in the 6-BA groups, axillary buds began to germinate after 12 days (Figure 1C), and then some of them were continuously elongated (Figure 1D). Adventitious buds from the leaves occasionally differentiated and the probability of recurrence was very low (Figure 1E-F). Regarding to the KT groups, no callus was found on the stem segments with nodes, while the germinated axillary buds from the nodes showed vitrification (Figure 1G). The redifferentiation time of callus from leaves was long. Two kinds of explants in NAA treatment group had no callus, among which adventitious roots were formed in the nodes. Especially, in the MS medium with 1.0 mg·L^-1 NAA, the rooting rate reached 36.67 %, but the roots were thin and easily broken (Figure 1H).

Callus induction. According to the results of a single factor experiment, 6-BA, NAA and KT were selected as the factors of optimization experiments of callus induction to design L₉ (3⁴) orthogonal experiment. Two types of explants were cultured after 30 days, the callus induction results and Range (R) analysis were shown in Table 1, and the Variance analysis was shown in Table 2 and 3.

As could be seen from Table 1, R of different PGRs sorted in descending order was 6-BA (24.44) > KT (10.00) > NAA (4.44) > Blank (2.22). It was indicated that 3 kinds of hormone contributed to callus induction equally with reliable ability. Based on the results of variance analysis in Table 2, B (6-BA) had an extremely significant effect to callus induction (P < 0.01), A (KT) also had a significant effect (P < 0.05), while C (NAA) had no significant effect (P > 0.05). According to Duncan test in Table 3, the level 1 (0.5 mg·L^-1) of 6-BA had the vast influence on callus induction, which was significantly different with the level 3 (1.5 mg·L^-1). Moreover, level 2 of KT (0.05 mg·L^-1) had obviously effect on callus induction, which was significantly different with level 1 (0.01 mg·L^-1). By the mean value (K) analysis, the optimal hormone combination for callus induction was A2B1C2 (C04: MS + 0.05 mg·L^-1 KT + 0.5 mg·L^-1 6-BA + 0.5 mg·L^-1 NAA) with 100 % callus induction rate. The statistical standard of Table 1 was the tight texture callus (Figure 2A).

Stem segments with nodes were cultured in the medium of C04 group, axillary buds showed signs of germination after 8 days and then the base formed green callus with tight texture after 15 days (Figure 2B). After 30 days, the expanded callus at the base was removed from stem segments with nodes and divided into 1cm×1cm for transfer to the fresh medium (Figure 2C). After 10 days of subculture, although the divided callus surface occasionally had bud differentiation, the growth of buds was not normal (the bud was small, thin and vitrification). Meanwhile, the callus gradually became loose and lost green and finally withered (Figure 2D). When the callus were cultured in the subsequent experiment of adventitious bud differentiation, the differentiation rate was very low and the growth was slow even after changing the combination of different hormone types and concentrations. Therefore, we considered the induction of embryonic callus with TDZ of high cytokinin activity. In addition, we also found that in the callus induction process, the induced callus from leaves was looser than that from stem segments with nodes and was not suitable for adventitious bud differentiation.

Differentiation and proliferation of adventitious shoot. Tight callus from stem segments with nodes was transferred into the MS medium containing different concentrations of TDZ. Surprisingly, there was not embryogenic callus and signs of somatic embryogenesis in all experimental groups. On the contrary, a large number of adventitious buds differentiated with callus proliferation (Table 4). Proliferation coefficient of callus and adventitious shoot increased with increasing the concentration of exogenous hormone TDZ in a certain range. When the concentration was 0.2 mg·L^-1, proliferation coefficient of callus and adventitious shoot reached the highest, which was 4.0 and 12.0 respectively with 100 % the occurrence rate of adventitious shoot. When the TDZ concentration was more than 0.2 mg·L^-1, the proliferation coefficient decreased rapidly, which indicated that the lower concentration of TDZ was beneficial for the callus proliferation and adventitious shoot differentiation (Table 4, Figure 3).

The formed callus in C04 medium were cultured after 30 days and then transferred into the MS medium containing 0.2 mg·L^-1 TDZ for callus proliferation and adventitious buds differentiation. After 10 days of culture, as the axillary buds on the nodes grow rapidly, callus was also continuously proliferating and differentiating a large number of conical buds on the surface (Figure 4A). Next, these conical buds gradually formed adventitious bud clusters after 20 days (Figure 4B). After 25 days, the growth of the main buds was accompanied by the differentiation and growth of cluster buds at the base (Figure 4C). Finally, the proliferation coefficient of adventitious cluster buds reached more than 12.0 after 30 days (Figure 4D). However, in the subsequent subculture, it was found that the callus proliferation was slow, the base gradually turned yellow and started to lose the redifferentiation ability (Figure 4E), accompanied by a sharp decline in the proliferation coefficient of adventitious shoot. After repeated experiments, the seeding from the MS medium with 0.2 mg·L^-1 TDZ were transferred into blank MS medium, the proliferation coefficient of callus and clusters buds did not change and the growth of cluster bud was well-grown (Figure 4F). Thus, MS medium with 0.2 mg·L^-1 TDZ and MS medium were used alternately as the culture method for the differentiation and proliferation of adventitious shoot in this study.

Rooting culture. Main seedlings (3-4 cm) from the MS medium with 0.2 mg·L^-1 TDZ and blank MS medium were excised for complete combination experiment. As can be seen from Table 5, the rooting rate of adventitious shoots from different differentiation medium was significantly different. The differentiated adventitious shoots by blank MS medium had higher rooting rate. When 1.0 mg·L^-1 NAA and 0.1 mg·L^-1 6-BA were added to the MS medium, the rooting rate reached 100 %. In these medium, adventitious roots were differentiated from the base of stem segment after 15 days (Figure 5A-B) and the seedlings grew healthily and had developed roots system after 30 days (Figure 5C-D). In contrast, differentiated adventitious seedlings directly from TDZ not only had a lower rooting rate, but also produced a small number of calluses, from which adventitious roots were formed.

The seedlings with 5-6 cm high and 3-4 cm root were transplanted into the outdoor with natural light for 7 days. Then culture remnants on the root were carefully washed and transplant it into peat soil matrix, and the survival rate was 95 % after 30 days (Temperature: 20~25 ºC, Humidity: 75 - 95 %) (Figure 5E-F).

Discussion

In general, perennial herbs with strong woody stems mostly achieve the purpose of proliferation through axillary bud germination. In these species, although the explants have strong dedifferentiation ability, their redifferentiation ability is very weak (^{Cao et al. 2017}, ^{Wang et al. 2018}). Hence, the method of direct organogenesis was considered in this experiment. However, in the single factor experiment, axillary buds rarely germinated from stem segments with nodes and were vitrified (Figure 1C and G). As reported by ^{Frank et al. (1995)}, the vitrified shoots have a high activity of superoxide dismutase which leads to hydrogen peroxide accumulation. Meanwhile, the occurrence rate of adventitious bud was still very low after many combination experiments. On the other hand, the somatic embryo is one of the most ideal and commonly used receptors in genetic engineering, because of the strong regeneration ability. The somatic embryo is also a good way of plant regeneration. At present, many plants have been reported to induce somatic embryos in vitro with different explants (^{Choffe et al. 2000}, ^{Tanaka et al. 2000}, ^{Sagare et al. 2000}). Therefore, the method of inducing embryonic callus was considered in this study to establish a plant regeneration system of C. lancifolius. Interestingly, when the callus was inoculated on MS medium with TDZ, no embryonic callus was observed, only a large number of adventitious buds were differentiated. This suggested that TDZ played a decisive role in the redifferentiation of herbs with woody stems in addition to its obvious effect in inducing embryonic callus, which has also been reported in Vaccinium vitis-idaea L. (^{Debnath 2005}).

The dedifferentiation ability of different tissues and organs of the Campanulaceae plants is different, so the choice of explant is also different. Similarly, there were differences in the calluses between leaves and stem segments with nodes. According to ^{Peng et al. (2010)} and ^{Tang et al. (2010)}, the induced callus from stem segments of Codonopsis lanceolata (Sieb. et Zucc.) Trautv. did not grow as well as that from leaves. On the contrary, as reported by ^{Liu (2004)}, stem segments of Platycodon grandiflorus (Jacq.) A. DC. were more beneficial to callus induction, and its texture was more compact than induced callus from the leaves, which was like the results of our study. At the same culture conditions, the induced calluses from various explant not only have differences in morphology (callus from stem segments with nodes had a compact texture and bright green color, while callus from leaf was loose and yellow-green) but also have significant differences in redifferentiation ability. In this study, the incidence of adventitious buds from stem segments with nodes was higher than that from the leaves. Whereas, the induced callus from leaves produced adventitious buds belong to the accidental phenomenon. It was speculated that the morphology and texture of callus had some effect on the differentiation of adventitious buds. The indirect effect of the selection of explants on the differentiation efficiency of adventitious buds was further explained (^{Dayal et al. 2003}), so stem segments with nodes were the optimal explant for rapid propagation of C. lancifolius in vitro.

For studies on indirect organogenesis using leaves and veins as materials, some researchers believe that induced callus from leaves originated from two types of cells, mesophyll cells, phloem parenchyma in small vascular bundles and bundle sheath cell, whereas induced callus from veins are derived from vascular bundles, sclerenchyma and parenchyma cells (^{Peng et al. 2007}). In our study, different from Amorphophallus dunnii Tutcher (^{Luo et al. 2006}) and Nicotiana tabacum L. (^{Peng et al. 2004}), which adopted leaf veins to induce callus and then differentiate and form regenerated plants, the occurrence of adventitious buds from the leaf veins of C. lancifolius was more similar to the direct organogenesis. This phenomenon has not been reported in other species and deserved further study.

PGRs types and concentrations are two kinds of external factors in tissue culture. TDZ is active substituted urea cytokinins, which is beneficial for the proliferation of adventitious buds and cluster buds in a few woody plants (^{Ledbetter & Preece 2004}, ^{Radhika et al. 2006}), its most important physiological function is to induce somatic embryogenesis (^{Gallo-Meagher et al. 2000}). In this study, the maximum proliferation coefficient of the cluster buds was 12.0 with the help of TDZ after 30 days. Next, the main bud was used as the rooting material and the rest callus with buds cultured in the MS medium with TDZ 0.2 mg·L^-1 after 30 days. The proliferation coefficient of the callus with buds was higher than the single callus, which was estimated that synthesized auxin form the bud tip during the growth process was transported to the callus at the base due to the downward transport, and then combined with exogenous cytokinins (TDZ) to promote its proliferation. This phenomenon has been reported in Gentiana rigescens Franch. ex Hemsl. (^{Xi et al. 2018}). While stem segments with nodes were cultured on this medium individually (MS + 0.2 mg·L^-1 TDZ), the rate of callus induction was lower with malformed adventitious buds (Figure 4G). Therefore, in the medium with TDZ, the callus with buds was the optimum explants for bud differentiation culture. This may be related to the fact that a low concentration of TDZ can induce the biosynthesis of cytokinin and inhibit the degradation of endogenous auxin to promote the proliferation and redifferentiation of callus (^{Chen et al. 2006}). In the subsequent subculture, the callus with buds decayed and the proliferation coefficient of adventitious bud gradually decreased. However, when the callus with buds was transferred into blank MS medium, the proliferation coefficient began to upturn, which indicated that the accumulation of endogenous hormone to a certain extent will inhibit the occurrence of buds (^{Bairu et al. 2011}). It was guessed that the residual TDZ activity in the regenerated plants was reduced to the critical value of suitable cluster bud differentiation after the MS medium culture, and then the cluster bud differentiation was resumed. The proliferation coefficient of the adventitious bud was more than 12.0 in the alternate culture of MS medium with 0.2 mg·L^-1 TDZ and the blank MS medium. Although it is unclear that the mechanism of TDZ on the endogenous hormone of C. lancifolius, the efficient shoot formation system of C. lancifolius can be established by selecting the appropriate TDZ concentration, which provides a reference for species that are difficult to redifferentiate cluster bud after the formation of callus.

In rooting culture of C. lancifolius, the differentiated seedlings from TDZ had few callus and part of the root system were differentiated from callus. Such adventitious roots were not attached to the stem by vascular bundles, and the root system was easily separated from the plant during the transplanting process. This may be due to the synergistic effect between the activity of TDZ hormone left in adventitious buds cultured by TDZ and the addition of other exogenous hormones, which promoted callus induction, making its rooting rate was not as good as differentiated seedlings from blank MS medium. Therefore, the differentiated adventitious shoots by blank MS medium were more suitable as rooting materials. Furthermore, the formed root-adventitious by the synergism of the hormones 1.0 mg·L^-1 NAA and 0.1 mg·L^-1 6-BA was better than that by NAA alone (^{Fulzele et al. 2002}, ^{Yu et al. 2006}). It indicated that the low concentration of cytokinin could promote the formation of the root-adventitious of C. lancifolius.