Induction and proliferation of axillary shoots from in vitro culture of Paeonia lactiflora Pall. mature zygotic embryos

Abstract

Zygotic embryos of three herbaceous peony (Paeonia lactiflora Pall.) cultivars (‘Fen Yu Nu’, ‘Zhong Sheng Fen’ and ‘Zhu Sha Pan’) were used to study embryo germination and optimum plant growth regulator (PGR) combinations for in vitro propagation and root initiation. Mature zygotic embryos (>90 days after flowering [DAF]) gave better germination and survival than immature zygotic embryos (50–70 DAF). A protocol for initiating shoot growth and axillary shoot proliferation was established using mature embryos. The best results were obtained when using excised zygotic embryos (EZEs) obtained by removing the testa and endosperm from the seed. The best medium for EZE germination (an embryo with cotyledons) was half-strength Murashige and Skoog (MS) medium (with double-strength CaCl₂) supplemented with 0.5 mg L⁻¹ 6-benzyladenine (BA) and 0.5 mg L⁻¹ gibberellic acid (GA₃). The best medium for axillary shoot proliferation was half-strength MS medium (including double-strength CaCl₂) supplemented with 1 mg L⁻¹ BA and 1 mg L⁻¹ GA₃. PGR-free half-strength MS was the best medium for promoting root development on seedlings (a germinated EZE with tender leaves) and for robust in vitro seedling establishment. The acclimatization of herbaceous peony remains the most challenging step of the in vitro protocol.

Keywords:

• embryo rescue
• herbaceous peony
• micropropagation
• tissue culture
• zygotic embryo

Introduction

Herbaceous peony (Paeonia lactiflora Pall.) is a long-lived perennial of the Paeoniaceae that has been cultivated in China for more than 3900 years (Wang & Zhang 2005). Chinese people favour herbaceous peony because of its diverse forms, colours and applications, from garden plant, to potted plant, or dry or cut flower use. Plant division and seed breeding are the main propagation techniques (Kamenetsy & Dole 2012), although these methods require a considerable amount of propagation material and can only be carried out during the vegetative growth season (Qin 2004). When propagating by plant division, in which the tuberous root is divided into several small plants each of which contains three to five dormant vegetative buds (Shannon & Kamp 1959), the coefficient of propagation is low and the ornamental value of stock plants declines because of changes in shape of the tuberous root and because of the decline in the vigour of underground buds. Seed germination is limited by epicotyl dormancy and is strongly dependent on season; peony has a slow reproductive cycle of 5–10 years (Krekler 1962; Qin 2004). The slow propagation rates mean it is difficult to meet the increasing demand of Chinese and international markets with herbaceous peony being a focus for exports from China (Wang et al. 2011). Use of tissue culture could increase propagation rates, which would be significant for the production of popular or newly selected cultivars (Shen et al. 2012). Although various explants have been used to initiate in vitro cultures of herbaceous peony, including underground buds (Wu & Yu 2010; Wu, Yu et al. 2011; Wu, Shen et al. 2011), stems, petioles and leaves (Wu, Shen et al. 2011; Yu, Wu & Pan 2011), problems continue to be encountered with tissue culture (Yu et al. 2005; Yu, Wu et al. 2011).

Peony seeds have a complex sequential dormancy and may take up to 2 years to germinate (Krekler 1962; Griess & Meyer 1976). In peony breeding, embryo culture can be used to overcome hybrid embryo abortion, bypass seed dormancy, shorten the breeding cycle and improve germination rates. Work in embryo culture of herbaceous peony has made some advances, including the identification of basal medium, choice of plant growth regulators (PGRs) and breaking epicotyl dormancy. The seed coat (testa), endosperm extract and gibberellic acid (GA₃) influence the germination of P. ostii var. lishizhenii embryos (Wang & van Staden 2002). In their study, epicotyl dormancy was broken by soaking intact seeds in GA₃ or by growing excised zygotic embryos (EZEs) on woody plant medium (WPM; Lloyd & McCown 1980) containing 1 mg L⁻¹ GA₃. Wang and van Staden (2002) reported that 0.1 mg L⁻¹ indole-3-butyric acid (IBA) induced a strong root system in in vitro germinated seedlings. Lan (2003) overcame seed dormancy in intact seeds using Murashige and Skoog (MS) medium (Murashige & Skoog 1962) supplemented with 1 g L⁻¹ lactoalbumin hydrolysate (LH), 0.1 g L⁻¹ vitamin C (VC), 3 g L⁻¹ active charcoal (AC) and 3 mg L⁻¹ GA₃. Somatic embryos could be induced from P. anomala L. EZEs in the presence of 0.5 mg L⁻¹ 6-benzyladenine (BA), and also from callus in the presence of 1 mg L⁻¹ 1-naphthyleneacetic acid (NAA) and 1 mg L⁻¹ BA (Brukhin & Baatygina 1994). High concentrations of NAA increased callus formation, but caused browning and a decrease in leaf differentiation.

Adventitious buds can be induced on cotyledons on MS medium containing 0.5–1.0 mg L⁻¹ GA₃ (Qiu et al. 2009). Gao (2010), using mature herbaceous peony EZEs as explants, found that GA₃ at 0–1.0 mg L⁻¹ could replace low temperature (4 °C) to break epicotyl dormancy, while 1.0 mg L⁻¹ BA with 1.5 mg L⁻¹ kinetin (Kin) promoted axillary shoot proliferation. Epicotyl dormancy was broken in herbaceous peony EZEs after culture on modified Linsmaier and Skoog (LS) (Linsmaier & Skoog 1965) medium containing 1 or 10 µM BA, and leaf number and petiole length increased in response to 1.5 mg L⁻¹ GA₃ although roots that formed on EZEs growing on 1 mg L⁻¹ BA were stunted and did not form lateral roots (Buchheim et al. 1994).

In these experiments, mature seeds of three varieties of P. lactiflora were used to study EZE germination, and to explore the optimum level and combination of PGRs necessary to proliferate and root seedlings derived from EZE culture.

Materials and methods

Plant material

Paeonia lactiflora ‘Fen Yu Nu’ (Fig. 1A), ‘Zhong Sheng Fen’ (Fig. 1B) and ‘Zhu Sha Pan’ (Fig. 1C) were used in the experiment. Open-pollinated seeds (50, 70 and >90 days after flowering [DAF]) were collected from July to September 2011 from the Society of Forestry Experimental Station (Beijing). Seeds older than 90 DAF were considered to be mature, while those at 50 and 70 DAF were considered immature.

Figure 1 Culture of mature Paeonia lactiflora Pall. zygotic embryos. A, Mature flowers of ‘Fen Yu Nu’; B, Mature flowers of ‘Zhong Sheng Fen’; C, Mature flowers of ‘Zhu Sha Pan’; D, Excised zygotic embryos (EZEs) of ‘Zhong Sheng Fen’ on BM + 0.2 mg L⁻¹ BA + 0.05 mg L⁻¹ NAA + 0.5 mg L⁻¹ GA₃ 10 days after initiation; E, Growth of mature EZE of ‘Zhong Sheng Fen’ after culture in BM + 1.0 mg L⁻¹ BA + 1.0 mg L⁻¹ GA₃ for 30 days; F, Growth of mature EZE of ‘Zhu Sha Pan’ after culture in BM + 1.0 mg L⁻¹ BA + 1.0 mg L⁻¹ GA₃ for 35 days; G, Normal seedlings of ‘Fen Yu Nu’ formed in the presence of 0.2 mg L⁻¹ BA + 0.05 mg L⁻¹ NAA + 0.5 mg L⁻¹ GA₃ after 30 days (normally expanded cotyledons); H, Axillary shoot induction of ‘Fen Yu Nu’ on BM + 0.5 mg L⁻¹ BA + 0.3 mg L⁻¹ GA₃ after 35 days; I–J, Shoot proliferation of ‘Fen Yu Nu’ on BM + 0.5 mg L⁻¹ BA + 0.5 mg L⁻¹ GA₃ after 35 days; K, Root length of ‘Fen Yu Nu’ (L > 5 cm; 2 cm ≤ L ≤ 5 cm; and (L) < 2 cm) on PGR-free BM after 35 days. Scale bars = 1 cm.

Reagents

All chemicals and reagents (tissue culture grade) were purchased from Beijing Biodee Biotechnology Co Ltd, Beijing, China.

Surface sterilization

Before initiating in vitro cultures, the explants—mature seeds whose testa had been either removed (i.e. EZEs) or retained—were washed in tap water for 30 min. Explants were dipped in 75% ethanol for 30 s, immediately followed by 8 min sterilization with a dilute solution of HgCl₂ (0.1% w/v), then rinsed five times (5 min each time) in autoclaved distilled water. Following sterilization, the endosperms were removed from some EZEs to leave isolated embryos that were cultured on agar-solidified medium containing different PGR combinations. Mature surface-sterilized seeds with testa and endosperm intact, seeds with testa removed but with endosperm attached, and EZEs were inoculated on to the basal medium (BM) supplemented with 1.0 mg L⁻¹ BA and 1.0 mg L⁻¹ GA₃ (initiation medium), based on trials conducted and explained below.

Culture medium and growing conditions

The BM for all experiments was 1/2 MS supplemented with double-strength calcium chloride (Ca²⁺), 30 g L⁻¹ sucrose and 7 g L⁻¹ agar (≥98% purity) (Wu, Yu et al. 2011). Different combinations and concentrations of PGRs, including BA, GA₃, NAA and Kin, were added to this BM. The explants were cultured for 30 days when effects of experimental treatments were assessed and then plants were transferred to fresh medium as required.

The pH of all media was adjusted to 5.8 prior to autoclaving at 118 °C for 18 min. The culture vessels comprised 100 mL Erlenmeyer flasks with 40 mL of medium. Plants were grown in a tissue culture room at 25 ± 2 °C and a 16 h photoperiod with 50 μmol m⁻² s⁻¹ photosynthetic photon flux density (PPFD) supplied using cool white fluorescent tubes. All data (contamination percentage, germination percentage [GP] and percentage seedling conversion to plantlets, in which cotyledons expanded normally) were recorded after 30 days of culture.

Effect of embryonic age on young embryo culture

EZEs of ‘Fen Yu Nu’, ‘Zhong Sheng Fen’ and ‘Zhu Sha Pan’ of different maturities (i.e. 50, 70, 90 DAF) were inoculated on to initiation medium. After 30 days, the level of contamination (%), germination (%) and seedling conversion to plantlets (the number of seedlings/the number of explants inoculated) (%) were calculated.

Initiation of zygotic embryo culture

EZEs of ‘Fen Yu Nu’ were used for this experiment. The effects of BA (0.2, 0.5, 1.0 mg L⁻¹) in combination with NAA (0, 0.05, 0.1 mg L⁻¹) and GA₃ (0, 0.5, 1.0 mg L⁻¹) on induction of growth was studied. No zero value of BA was included because a previous study (Wu, Yu et al. 2011) had already proved that BA had an effect on EZE induction. Thus, the objective here was to confirm and investigate a wider range of effective BA concentrations. This experiment was designed as an orthogonal test (L₉ [3³] in which explants were cultured for 30 days.

Axillary shoot induction and proliferation

Two experiments were carried out in chronological order. In the first experiment, uncontaminated seedlings obtained from the initiation of zygotic embryo culture and following germination of EZEs in vitro were transferred to BM supplemented with 0.5 mg L⁻¹ BA and 0.3 mg L⁻¹ GA₃ for axillary shoot growth; this medium had previously been optimized for axillary shoot outgrowth by Wu (2011). In a subsequent experiment, the effect of different treatments on axillary shoot proliferation was tested. Shoots obtained from axillary shoot growth were transferred to BM supplemented with (screened from media reported in the literature; Shen et al. 2012): 1.0 mg L⁻¹ BA and 1.0 mg L⁻¹ GA₃, or BM supplemented with 1.0 mg L⁻¹ BA and 1.5 mg L⁻¹ Kin, or BM supplemented with 1.0 mg L⁻¹ BA, 1.0 mg L⁻¹ GA₃ and 1.5 mg L⁻¹ Kin (following the results of Shen et al. 2012). The appearance of new leaves (number of leaves/number of leaves before transfer to medium) was recorded after 30 days. Explants were subcultured on to fresh medium every 30 days, and there were no more than five transfers for any seedling as seedling vigour declined after this.

Root initiation and seedling establishment

In this experiment, seedlings showing normal growth were transferred to seedling medium in an attempt to improve acclimatization success. The media tested comprised BM, BM supplemented with 2 mg L⁻¹ BA and 0.2% AC, and BM supplemented with 2 mg L⁻¹ BA, 2 mg L⁻¹ IBA and 0.2% AC (these media were selected on the basis of results from Shen et al. 2012). The numbers of seedlings with defined root lengths (L < 2 cm, 2 cm ≤ L ≤ 5 cm, or L < 2 cm) were determined after 35 days.

Acclimatization and transplant of rooted plantlets

Parafilm, which covered the flasks, was half opened for 7 days, and then completely opened for 7 days prior to deflasking. At deflasking, the seedling-derived plantlets (roots having already emerged) were removed from the flask using forceps and rinsed with warm water to detach agar from the roots. Then they were dipped in 0.3% (w/v) potassium permanganate solution medium to provide some protection against fungi. Finally they were transferred to an autoclaved soil-less medium comprising perlite and vermiculite (1:1) and watered with 1/2 MS nutrient solution every day. The greenhouse conditions were identical to in vitro conditions: 25 ± 2 °C in a 16 h photoperiod with 50 μmol m⁻² s⁻¹ PPFD supplied by cool white fluorescent tubes.

Experimental design and statistical analyses

Experiments were laid out in a completely randomized design. Means were separated by one-way analysis of variance and significant differences were assessed using Duncan’s multiple range test at P = 0.05 using SPSS (Statistical Product and Service Solutions; USA) software version 13.0. Each experiment included at least three replications of 10 explants per treatment.

Results

When mature seeds of ‘Fen Yu Nu’ with or without the testa were inoculated on to initiation medium, the contamination level was high (37.67% and 62.91%, respectively) (Table 1) and GP and seedling conversion to plantlets were close to zero. By comparison, EZEs showed lowest contamination (6.67%) while GP and seedling conversion to plantlets were highest (89.43% and 61.33%, respectively) (Table 1). Moreover, about 5 days after transfer to initiation medium, cotyledons expanded. Seeds with the testa (i.e. intact seeds) only germinated after 20 days. Thus, EZEs were the most suitable explants for culture initiation.

Table 1 Assessment of three parameters using explants derived from zygotic embryos of P. lactiflora ‘Fen Yu Nu’ (n = 30).

Treatments	Contamination (%)	Germination (%)	Seedling conversion to plantlets (%)
Seeds with testa retained	37.67 ± 2.36^b	0.00 ± 0.00^c	0.00 ± 0.00^b
Seeds with testa peeled off	62.91 ± 1.96^a	6.67 ± 2.68^b	0.00 ± 0.00^b
ESEs	6.67 ± 1.88^c	89.43 ± 2.18^a	61.33 ± 1.35^a

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). Explants were inoculated on to BM supplemented with 1.0 mg L⁻¹ BA and 1.0 mg L⁻¹ GA₃.

GP and seedling conversion to plantlets were significantly higher in ‘Fen Yu Nu’ EZEs harvested 90 DAF (90.54% and 71.29%, respectively; Table 2) than 50- and 70-day-old EZEs. Similar trends were observed for ‘Zhong Sheng Fen’ and ‘Zhu Sha Pan’. GP and seedling conversion to plantlets were higher (95.33% and 81.29%, respectively) (Table 3) in 90-day-old ‘Zhong Sheng Fen’ EZEs than in ‘Zhu Sha Pan’ EZEs (95.00% and 66.67%, respectively; Table 4) (Figs 1D, 1E). In summary, mature (i.e. 90-day-old) EZEs plated on to initiation medium had better and more effective germination and development into seedlings than less mature EZEs (i.e. 50 or 70 days old).

Table 2 Effect of age on culture of young embryos of ‘Fen Yu Nu’ (n = 30).

Sampling time	Germination (%)	Seedling conversion to plantlets (%)
50 d after flowering	51.67^c	41.67^c
70 d after flowering	72.00^b	59.33^b
90 d after flowering	90.54^a	71.29^a

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). Explants were inoculated on to BM supplemented with 1.0 mg L⁻¹ BA and 1.0 mg L⁻¹ GA₃.

Table 3 Effect of age on culture of young embryos of ‘Zhong Sheng Fen’ (n = 30).

Sampling time	Germination (%)	Seedling conversion to plantlets (%)
50 d after flowering	61.67^b	49.73^c
70 d after flowering	86.67^a	63.33^b
90 d after flowering	95.33^a	81.29^a

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). Explants were inoculated on to BM supplemented with 1.0 mg L⁻¹ BA and 1.0 mg L⁻¹ GA₃.

Table 4 Effect of age on culture of young embryos of ‘Zhu Sha Pan’ (n = 30).

Sampling time	Germination (%)	Seedling conversion to plantlets (%)
50 d after flowering	76.79^c	54.72^c
70 d after flowering	87.88^b	58.79^b
90 d after flowering	95.00^a	66.67^a

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). Explants were inoculated on to BM supplemented with 1.0 mg L⁻¹ BA and 1.0 mg L⁻¹ GA₃.

The highest GP of EZEs (70%) (Table 5) was observed on BM supplemented with 0.2 mg L⁻¹ BA, 0.05 mg L⁻¹ NAA and 0.5 mg L⁻¹ GA₃ (Fig. 1F). The highest rate seedling conversion to plantlets (Fig. 1G), 40% (Table 5), was observed from treatment 4 (Table 5). The highest rate of root initiation (23.33%) was on initiation medium (Table 5). The presence of GA₃ was the most important factor ensuring germination of EZEs and seedling formation, while NAA was the most important factor for root initiation. As the concentration of NAA increased, the percentage of rooting decreased, and there was increased callus formation on the cotyledons. As GA₃ concentration increased, the frequency of deformities in seedlings increased, particularly to the epicotyls, which showed symptoms of hyperhydricity. Based on the average seedling rate (Table 6), the most suitable medium for initiation of in vitro cultures from EZEs was BM supplemented with 0.5 mg L⁻¹ BA and 0.5 mg L⁻¹ GA₃.

Table 5 The influence of different PGRs on seedling production of P. lactiflora ‘Fen Yu Nu’ EZEs (L₉ (3³)) (n = 30).

Treatments	BA (mg L^−1)	NAA (mg L^−1)	GA3 (mg L^−1)	Germination (%)	Seedling conversion to plantlets (%)
1	1 (0.2)	1 (0)	1 (0)	40.00 ± 1.06^d	0.00 ± 0.00^d
2	1 (0.2)	2 (0.05)	2 (0.5)	70.00 ± 3.87^a	30.68 ± 2.38^b
3	1 (0.2)	3 (0.1)	3 (1.0)	46.67 ± 0.00^c	0.00 ± 0.00^d
4	2 (0.5)	1 (0)	2 (0.5)	60.00 ± 10.00^b	40.00 ± 10.00^a
5	2 (0.5)	2 (0.05)	3 (1.0)	26.67 ± 1.89^e	10.00 ± 0.00^c
6	2 (0.5)	3 (0.1)	1 (0)	13.33 ± 0.00^g	0.00 ± 0.00^d
7	3 (1.0)	1 (0)	3 (1.0)	23.56 ± 1.76^e,f	5.56 ± 5.09^c,d
8	3 (1.0)	2 (0.05)	1 (0)	17.72 ± 1.00^f,g	3.33 ± 5.77^c,d
9	3 (1.0)	3 (0.1)	2 (0.5)	30.00 ± 0.00^e	0.00 ± 0.00^d

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). EZEs were inoculated on to BM supplemented with different concentrations of PGRs. Values in parenthesis = level of concentration of the different factors. The PGR combination in treatments 1, 6 and 9 were unsuitable for tender/true leaves, so seedlings were unable to convert into plantlets even though they developed roots. Seedling conversion to plantlets is different from germination. This is because in the process of embryo growth, many cotyledons are unable to expand normally and cannot form healthy seedlings, so there is some difference between both processes, hence the use of two terms.

Table 6 Analysis of average rate of seedling conversion to plantlets.

Treatments	BA (mg L^−1)	NAA (mg L^−1)	GA3 (mg L^−1)
R (seedling)	13.71	15.19	22.45
x̄1 (seedling)	10.23	15.19	1.11
x̄2 (seedling)	16.67	14.67	23.56
x̄3 (seedling)	2.96	0	5.19

Range = the detection of that concentration of PGR that was most suitable for the experiment. X₁, X₂ and X₃ are the average values of each concentration of PGR for three tests whose maximum minus minimum represents the range, R. For example, for BA: X1: 10.23 = (0.00 + 30.68 + 0.00)/3; X2: 16.67 = (40.00 + 10.00 + 0.00)/3; X3: 2.96 = (5.56 + 3.33 + 0.00)/3; R: 13.71 = 16.67−2.96.

In the transitional stage from germination to axillary branching, the cotyledons had fully expanded, the taproot extended normally, and seedlings were growing well after 30 days on 1/2 MS medium supplemented with 0.5 mg L⁻¹ BA and 0.3 mg L⁻¹ GA₃ (Fig. 1H). Significantly more axillary shoots grew on initiation medium than on the two other media with a concomitantly high number of expanded leaves produced (Table 7). Thus, the optimal medium for axillary shoot proliferation was initiation medium (Figs 1I, 1J).

Table 7 Impact of three PGR treatments on axillary shoot proliferation and growth of P. lactiflora ‘Fen Yu Nu’ seedlings (n = 30).

Treatment (mg L^−1)	Leaf expansion (number of times that leaves expanded)	Number of axillary shoots
1.0 BA + 1.0 GA3	3.76 ± 0.28^a	10.98 ± 0.85^a
1.0 BA + 1.5 Kin	2.69 ± 0.09^b	7.43 ± 1.15^b
1.0 BA + 1.0 GA3 + 1.5 Kin	2.72 ± 0.16^b	7.38 ± 1.29^b

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). EZEs were inoculated on to BM supplemented with different concentrations of PGRs.

During establishment, most seedlings developed a taproot that was slender and weak. In an attempt to form a stronger root system and seedling—thought essential for effective acclimatization—shoots were transferred to media expected to promote root initiation. The highest rooting percentage (100%) was observed on PGR-free BM, on which the frequency of seedlings with roots longer than 5 cm was 41.9% which was higher than that observed on the other two media (BM + 2 mg L⁻¹ BA + 0.2% AC and BM + 2 mg L⁻¹ BA + 2 mg L⁻¹ IBA + 0.2% AC), at 30.79% and 37.23%, respectively (Table 8). The frequency of roots with length between 2 cm and 5 cm was highest (47.53%) on BM supplemented with 2 mg L⁻¹ BA and 0.2% AC (Fig. 1K). Thus PGR-free BM was the best medium for seedling growth and establishment (Table 8).

Table 8 Treatments used to increase lateral root induction and strengthen seedlings (n = 10).

		Seedlings with defined root length (L) (%)
Treatments (mg L^−1)	Rooting (%)	L < 2 cm	2 cm ≤ L ≤ 5 cm	L > 5 cm
1/2 MS (control)	100.00 ± 0.00^a	25.26 ± 1.72^a	32.82 ± 2.52^c	41.92 ± 0.73^a
2.0 BA + 0.2% AC	98.89 ± 1.92^a	21.68 ± 2.05^b	47.53 ± 1.48^a	30.79 ± 1.51^c
2.0 BA + 2.0 IBA + 0.2% AC	96.67 ± 0.00^b	21.41 ± 2.38^b	41.36 ± 2.77^b	37.23 ± 0.84^b

Different letters within a column indicate significant differences according to Duncan’s multiple range test (P ≤ 0.05). BM supplemented with different concentrations of PGRs.

Plantlet acclimatization was generally poor. Approximately 7 days after deflasking, most weak seedlings gradually yellowed, wilted and perished, while some perished from bacterial or fungal contamination. Even though the strong seedlings persisted for 3 weeks, and leaf area and plant height increased, acclimatized plants did not produce new leaves. These plants also eventually died.

Discussion

The seed coat or testa is not a significant barrier to water uptake in peony seed but does provide a degree of mechanical resistance, thereby inhibiting germination (Wang & van Staden 2002). Our study confirms previous findings for P. ostii that EZEs are the best explants to ensure high germination (Wang & van Staden 2002). He et al. (2006) also used zygotic embryos (equivalent to EZEs) of different ages (20, 28, 48, 66 and 90 days after pollination) of ‘Shu Sheng Peng Mo’ and ‘Feng Dan Bai’ as explants. In their study, 90-day-old EZEs grew into seedlings 5 weeks after culture on MS medium + BA (0–2 mg L⁻¹) + IAA (0–1 mg L⁻¹) + GA₃ (0–1 mg L⁻¹) for 3 to 4 months, with a seedling percentage of 7.4% and 13.5% for ‘Shu Sheng Peng Mo’ and ‘Dan Feng Bai’, respectively. Their results showed that successful germination and seedling establishment depend on embryo maturity (>90 days for best germination), with poor germination and seedling establishment observed when immature embryos (50–70 days) were used. GA₃ is frequently used to replace low temperature to break epicotyl dormancy in herbaceous peony zygotic embryos (Zhang & Wang 2008; Gao 2010). The seed structures surrounding the embryo impose mechanical restraints on germination while the embryo itself appears to not have sufficient PGRs to promote epicotyl growth as exogenous GA₃, which functions to promote epicotyl growth (Wang & van Staden 2002). In the present study, GA₃ was the most important factor in promoting zygotic embryo germination and seedling development. NAA is also a good inducer of embryo development in tree peony (Jia et al. 2006; Zhou et al. 2009). However, in our study, as the concentration of NAA increased, so too did the amount of callus at the base of seedlings. Conversely, in the absence of added NAA, many seedlings (23.33%) grew roots directly, indicating either that BA could directly induce embryos and form taproots without any specific and separate rooting step being required, or that the level of BA was too low to interfere with root development. BA and NAA are usually combined during peony (‘Shu Sheng Peng Mo’) seedling induction from zygotic embryos (Gao 2009; Zhou et al. 2009), although both their experimental results show that leaving out NAA was most favourable for forming seedlings when the concentration of BA was 0.2 or 0.5 mg L⁻¹. However, in the present study, the combination of BA and GA₃ was best for ‘Fen Yu Nu’, supporting the notion that protocols remain genotype-dependent and conditions need to be optimized for each cultivar. Buchheim et al. (1994) noted a highly significant increase in leaf number when embryos were cultured on medium containing BA and usually treated with GA₃, which resulted in good rates of axillary shoot proliferation.

Roots developed spontaneously from EZEs, as expected. Using media with double the Ca²⁺ concentration is useful for the in vitro growth of both herbaceous peony (Guo 2001) and tree peony (Bouza et al. 1994). The use of BA, IBA and AC resulted in more roots 2–5 cm in length, but fewer roots less than 2 cm or longer than 5 cm in length, compared with the control treatment (Table 4), indicating the need for more focus on root initiation and growth in P. lactiflora in future studies. The development of a weak root system appears to remain to be the main limiting factor when deflasking herbaceous peony (Shen et al. 2012). At the end of these experiments, no seedlings were successfully acclimatized. Possible reasons for poor survival may be few lateral roots on the seedlings, the need to adjust the balance of porosity and moisture in the growing medium after deflasking, or a requirement for better environmental control, including temperature, humidity and sterility following transfer to the greenhouse. Special attention needs to be paid in the future to these factors to achieve a higher level of survival of acclimatized plantlets. Germinating peony normally results in about 70% GP but can only be done after seed maturation (5–6 months after pollination) and can only be carried out in the autumn (Xiaonan Yu, pers. obs.). We thus believe that this in vitro method can help shorten the breeding cycle.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (31400591) and Fundamental Research Funds for the Central Universities (No. YX2014-20). The authors wish to thank the valuable comments by the editor and anonymous reviewers.