Rapid and Efficient Optimization Method for a Genetic Transformation System of Medicinal Plants Erigeron breviscapus

Abstract

Erigeron breviscapus is an important medicinal plant with high medicinal and economic value. It is currently the best natural biological drug for the treatment of obliterative cerebrovascular disease and the sequela of cerebral hemorrhage. Therefore, to solve the contradiction between supply and demand, the study of genetic transformation of E. breviscapus is essential for targeted breeding. However, establishing an efficient genetic transformation system is a lengthy process. In this study, we established a rapid and efficient optimized protocol for genetic transformation of E. breviscapus using the hybrid orthogonal method. The effect of different concentrations of selection pressure (Hygromycin B) on callus induction and the optimal pre-culture time of 7 days were demonstrated. The optimal transformation conditions were as follows: precipitant agents MgCl₂ + PEG, target tissue distance 9 cm, helium pressure 650 psi, bombardment once, plasmid DNA concentration 1.0 μg·μL⁻¹, and chamber vacuum pressure 27 mmHg. Integration of the desired genes was verified by amplifying 1.02 kb of htp gene from the T0 transgenic line. Genetic transformation of E. breviscapus was carried out by particle bombardment under the optimized conditions, and a stable transformation efficiency of 36.7% was achieved. This method will also contribute to improving the genetic transformation rate of other medicinal plants.

1. Introduction

Erigeron breviscapus (Vant.) Hand-Mazz is a perennial herbaceous plant in the Asteraceae family. This plant is distributed in Hunan, Guangxi, Guizhou, Sichuan, Yunnan, and Tibet provinces in China. It can only be seen in some open slope grasslands and forest margins at altitudes of 1200–3500 m [1] and has been listed as a nationally protected species of traditional Chinese medicine. The chemical components in E. breviscapus mainly include flavonoids, caffeoyls, coumarins, lignans, terpenoids, and other components [2], among which scutellarin has the highest content. Studies have shown that scutellarin can reduce neuronal damage caused by traumatic brain injury, cerebral ischemia, reperfusion injury and hypoxic-ischemic brain injury [3,4,5], and is commonly used to treat acute cerebral infarction in elderly patients [6] and is also effective in patients with diabetic nephropathy [7,8]. Another active ingredient is caffeoylquinic acids (CQAs), which can dilate blood vessels and inhibit thrombosis in vivo [9]. E. breviscapus is currently the best natural biological medicine for the treatment of obliterative cerebrovascular diseases and the sequela of cerebral hemorrhage, with an efficiency of >95% and slight side effects. According to the World Health Organization, the number of deaths from cardiovascular diseases will increase to 23.6 million by 2030 [10]. Therefore, E. breviscapus is of great significance for recovery and secondary prevention after the first onset of cardiovascular and cerebrovascular diseases. In addition, E. breviscapus is a natural neuroprotective agent for Alzheimer’s disease [1]. At present, according to different doses and routes of administration of E. breviscapus extract, various dosage forms have been developed, such as tablets, capsules, oral liquids, and injections, and international research on E. breviscapus unilateral preparations has also begun. It is expected to become an internationally recognized botanical medicine after ginkgo biloba preparations.

At present, scutellarin is mainly derived from extracts of E. breviscapus, but the cost of planting E. breviscapus remains high due to the limited cultivation area, low content of active ingredients, serious pests and diseases, and the degradation of germplasm variation. A yeast cell factory for the total synthesis of scutellarin has been successfully constructed in the Saccharomyces cerevisiae chassis cells, and the total synthesis of scutellarin has been achieved, but it is still a certain distance from industrial production [11]. Therefore, urgent measures are needed to develop varieties with high content of active ingredients, high yield, and disease resistance to achieve sustainable utilization of resources.

The development of genetic manipulation, DNA technology, and genetic transformation provides a reliable route for the development of transgenic medicinal plants that are tolerant to abiotic stress, resistant to pests, and have excellent agronomic traits [12]. Direct genetic transformation has become the method of choice for basic plant research and the primary technique for generating transgenic plants [13,14]. Therefore, the development of genetic transformation technology provides an opportunity to accelerate the various improvements of E. breviscapus.

The two main methods of plant genetic transformation are Agrobacterium-mediated transformation and biolistic-mediated transformation [15,16]. Particle bombardment is simple to perform and is virtually unlimited in terms of plant ranges and material genotypes [17]. Various types of transformation receptors are suitable for large-sized genetic cargo [18]. Target genes can be introduced into the organelles of plant cells. Therefore, the biolistic-mediated transformation is widely used in the current transgenic research. Agrobacterium-mediated transformation is more efficient in dicots than in monocots and is limited to specific plant host ranges [15,19,20]. On the other hand, Agrobacterium-mediated transformation is generally considered to be more precise and controllable than particle bombardment; however, a report on bombardment strategies [13,21] showed little evidence of major differences in levels of transgene instability and silencing when compared in the same species as in other non-model systems. Both approaches have their advantages and limitations [22,23]. At present, regenerated plants of different explants of E. breviscapus, i.e., true leaves [24], petioles [25], and anther [26], have been successfully obtained and plant regeneration programs have been established. He et al. established an Agrobacterium-mediated transformation system and applied it to the functional verification of biosynthetic pathway genes [27]. Qiu screened suitable acceptor materials for gene gun induction [28].

In this study, the leaves of E. breviscapus were used as explants, and different particle bombardment parameters were optimized using an orthogonal design. After range analysis and variance analysis, the main influencing factors were identified, and the experimental effects were accurately and comprehensively evaluated to obtain the optimal transformation protocol. This is a successful attempt to directly transform E. breviscapus by gene transfer using the particle gun, which fully demonstrates the possibility of stable transformation of E. breviscapus. This study has long-term significance for the genetic engineering research of medicinal plants.

2. Results

2.1. Selection Pressure

In plant genetic transformation, the principle of using antibiotics is that they can effectively inhibit the growth, development, and differentiation of non-transformed cells or plants, but do not affect the normal growth of transformed cells or plants, or have little effect on transformed cells. Hygromycin B (Hyg) is widely used in many crops as an ideal screening reagent due to its high selection efficiency and small genotype differences. The induction results of different hygromycin treatments on E. breviscapus callus (

Table 1. Effects of Hyg on callus induction from leaf explants.

Antibiotic	Concentration (mg·L−1)	Callus Induction Rate (%)
Hyg	0.0	88.61 ± 5.11 *
	2.5	43.34 ± 3.67
	5.0	15.93 ± 4.90
	7.5	0
	10.0	0

* Values represent the mean (±SE) of three independent experiments.

) showed that callus induction frequency drastically decreased as the hygromycin concentration increased. After 2 weeks of culture, only very few calluses initiated from the cut edge on 2.5 mg·L⁻¹ and 5.0 mg·L⁻¹ Hyg. After 4 weeks of culture on medium containing 5 mg·L⁻¹ Hyg, explants turned brown and were completely necrotic. Under 7.5 mg·L⁻¹ Hyg, explants did not develop calluses. Therefore, 7.5 mg·L⁻¹ was the lethal dose for leaf explants. Hyg significantly delayed and inhibited callus initiation and growth as the selection concentration increased. Thus, we used stepwise increasing concentration of Hyg from 2.5 mg·L⁻¹ for 2~4 weeks post-transformation, and 5.0 mg·L⁻¹ for 1–2 months in the callus selection and plant regeneration.

2.2. Pre-Culture Period

The principle of single-variable experimentation was used to confirm the suitable pre-culture period. The results showed that pre-culture of leaf explants prior to bombardment could enhance transformation efficiency. The optimal pre-culture time was 7 d. For leaf explants, when the pre-culture time was 2–5 d, or more than 3 weeks, most of the explants decayed at the incision after particle gun bombardment, and then gradually died. When the pre-culture time exceeded 7 d, the transformation efficiency did not increase, although more resistant calluses could be produced (

Table 2. Effects of pre-culture day on transformation.

Pre-Culture Day (d)	Transformed Callus Frequency (%)
2	0
3	0
5	0
7	54.40 ± 17.40 *
14	59.67 ± 12.04
21	0

* Values represent the mean (±SE) of three independent experiments.

). Therefore, the optimal pre-culture time established in this study was 7 d.

2.3. Transformation Conditions

In this study, a hybrid orthogonal experiment was adopted to confirm the optimal transformation conditions (

Table 3. Parameters and levels involved in this study.

Symbol	Parameters	Levels
		1	2	3
A	Precipitation agents	CaCl2 + Spd	MgCl2 + PEG	-
B	Target tissue distance (cm)	3	6	9
C	Helium pressure (psi)	650	1350	1100
D	Number of bombardments	1	2	3
E	Plasmid DNA concentration (μg·μL−1)	1	1.5	2
F	Chamber vacuum pressure (mmHg)	26	27	28

). The particle gun transformation conditions were optimized using one two-level factor (precipitation agents) and five three-level factors (target tissue distance, helium pressure, bombardment times, plasmid DNA concentration, and chamber vacuum pressure). The experimental procedure is shown in Figure 1. The number of transformed adventitious buds was used to assessed transformation efficiency.

The results showed the differences of each parameter at different levels (Figure 2). Among them, the peak of the curve showed a significant level. Specifically: precipitation agent: MgCl₂ + PEG (level 2), target tissue distance: 9 cm (level 3), helium pressure: 650 psi (level 1), number of bombardments: 3 times (level 3), plasmid DNA concentration: 1 μg·μL⁻¹ (level 3), chamber vacuum pressure: 27 mmHg (level 2).

The range value R’ is calculated by the formula (Materials and Methods 4.7). The value of R’ is used to determine the degree of influence of various factors on the conversion efficiency. The influence order of each parameter was as follows (

Table 4. Range analysis of different parameters.

Parameters	K1 1	K2 2	K3 3	R 4	R’ 5
Precipitation agents	62	96	-	34	59.13
Target tissue distance (cm)	50	24	84	60	54.04
Helium pressure (psi)	67	47	44	23	20.72
Number of bombardments	58	37	63	26	23.42
Plasmid DNA concentration (μg·μL−1)	72	39	47	33	29.72
Chamber vacuum pressure (mmHg)	42	60	56	18	16.21

¹ K1, average of all values at level 1; ² K2, average of all values at level 2; ³ K3, average of all values at level 3; ⁴ R, the range value of each level; ⁵ R’, adjust the range R and compare it with the adjusted R’.

): precipitation agents > target tissue distance > plasmid DNA concentration > number of bombardments > helium pressure > chamber vacuum pressure.

Although the results of the range analysis are more intuitive, they cannot be used to distinguish the fluctuations caused by experimental conditions or experimental errors. However, analysis of variance can be used to compensate for the shortcomings of range analysis. The significance of the parameters was determined by the F-test. To confirm the significant differences between the six parameters, the variance analysis was performed (

Table 5. Variance analysis of different parameters.

Parameters	df a	SS b	MS c	Fd	p
Precipitation agents	1	64.22	64.22	9.53	0.021 *
Target tissue distance (cm)	2	301.78	150.89	22.38	0.002 **
Helium pressure (psi)	2	52.11	26.06	3.87	0.083
Number of bombardments	2	63.44	31.72	4.71	0.059
Plasmid DNA concentration (μg·μL−1)	2	98.78	49.39	7.33	0.025 *
Chamber vacuum pressure (mmHg)	2	29.78	14.89	2.21	0.191
Error	6	40.44	6.74
Total	17	650.56	38.27

^a df, Degree of Freedom; ^b SS, Sums of Squares; ^c MS, Adjusted Mean Sums of Squares; ^d F, F-test statistic. * 0.01 ≤ p < 0.05, ** p < 0.01.

). The results showed that precipitant (p < 0.05), target tissue distance (p < 0.01), and plasmid DNA concentration (p < 0.05) had a significant effect on transformation efficiency. However, helium pressure, number of bombardments, and chamber vacuum pressure were not significant. Therefore, the optimal transformation conditions were determined as follows: target tissue distance 9 cm, precipitant MgCl₂ + PEG, and plasmid DNA concentration 1.0 μg·μL⁻¹. The remaining transformation conditions could be considered from the perspective of economy and operation, with one bombardment, helium pressure 650 psi, and chamber vacuum pressure 27 mmHg.

2.4. Selection and Plant Regeneration

Transformed adventitious shoots were randomly selected 48 h after bombardment, and transient eGFP expression was detected by Laser Scanning Confocal Microscope (LSCM) (Figure 3a–c). The transformants were then transferred to callus induction medium (CIM; Murashige and Skoog (MS) + 1.0 mg·L⁻¹ 6-benzylaminopurine (6-BA) + 0.1 mg·L⁻¹ 1-naphthylacetic acid (NAA) + 3% sucrose + 0.4% phytagel) for 2 weeks and detected by LSCM. The presence of eGFP in the transgenic cells was confirmed by green fluorescent spots (Figure 3d,e). After callus formation, transformants were transferred to selection medium (CIM) containing 2.5 mg·L⁻¹ Hyg for the first round of selection for 2 weeks (Figure 3g). The transformants were then transferred to a regeneration medium (RM; MS + 2.0 mg·L⁻¹ 6-BA + 0.2 mg·L⁻¹ NAA + 3% sucrose + 0.4% phytagel, pH 5.8) containing 2.5 mg·L⁻¹ Hyg for a second round of selection for 4 weeks (Figure 3h). For efficient selection, regenerated transformants were transferred to RM containing 5.0 mg·L⁻¹ Hyg for a third round of selection for 6 weeks (Figure 3i). When selection was performed on RM containing Hyg, the non-transgenic tissues gradually turned brown while the presumed transformed tissues remained green and grew slowly (Figure 3g–i). The transformation efficiency was 36.7%. The transformation efficiency was calculated as the number of positive shoots that survived after the third round of selection relative to the total number of explants bombarded. After three subculture cycles, the transformed buds were transferred to a shoot elongation medium (SEM; 1/2MS + 2.0 mg·L⁻¹ 6-BA + 0.2 mg·L⁻¹ NAA + 3% sucrose + 0.4% phytagel, pH 5.8) for shoot elongation (Figure 3j,k). Finally, transformed rootless seedlings were transferred to a root induction medium (RIM; 1/2MS + 0.3 mg·L⁻¹ NAA + 0.5 mg·L⁻¹ indole-3-butyric acid (IBA) + 3% sucrose + 0.4% phytagel, pH 5.8) for further root formation for about 3 weeks (Figure 3l).

2.5. Identification of Transgenic E. breviscapus Plants

Putative transgenic lines were obtained by multiple rounds of selection following bombardment of E. breviscapus leaf explants. These putative transgenic lines were identified by PCR amplification using the specific primers (Hyg-F/R) and all the products were sequenced by Sanger sequencing to verify the htp gene. The results showed the expected bands of 1026 bp fragment, confirming that the htp gene had been integrated into the genome of E. breviscapus, while the wild-type (WT) plant showed no PCR amplification bands (Figure 4).

3. Discussion

One of the most effective methods for direct gene transfer is the particle bombardment method. Bacteria are not required during the transformation process. It is widely used and efficient in DNA transfer in mammalian cells, microbes, and monocots [29]. So far, particle bombardment has been used in many medicinal plants, i.e., Catharanthus roseus [30], Hypericum perforatum [31], Centella asiatica [32], Tripterygium wilfordii [33], Momordica charantia [34], Scoparia dulcis [35], etc. In this study, a simple and effective biolistic transformation method was established in E. breviscapus with a transformation efficiency of 36.7%.

The use of orthogonal design can reduce the number of experiments and the complexity of experimental analysis methods, overcome the blindness in condition optimization, and improve work efficiency and experimental accuracy. The method of biolistic transformation of E. breviscapus was optimized by the hybrid orthogonal design. The main influencing factors were precipitation agents, target tissue distance, and plasmid DNA concentration. The optimum transformation conditions were precipitant agents MgCl₂ + PEG, target tissue distance 9 cm, helium pressure 650 psi, bombardment once, plasmid DNA concentration 1.0 μg·μL⁻¹, and chamber vacuum pressure 27 mmHg. At present, a universal optimization strategy (Figure 1) has been developed and applied to a variety of medicinal plants, including Erigeron breviscapus, Salvia miltiorrhiza, Tripterygium wilfordii [33], and Aconitum carmichaelii. The explants used include leaves, stems, suspension cells, and calluses.

By optimizing the pre-incubation time, we found that 7–14 days of pre-cultivation was more conducive to the regeneration of explants after bombardment. However, shorter (2–5 d) or longer (21 d) pre-culture times were unfavorable to explants’ regeneration. The same conclusion was reached in the biolistic transformation of wheat-microspore-derived calluses and microspores. Pre-culture for 3–8 days could improve the GUS expression in microspores [36]. The results of optimization experiments of Fistulifera solaris showed that the highest conversion rate was achieved by pre-culturing for 48 h [37]. The above results indicated that the optimal pre-cultivation time varied with individual differences. Therefore, it is necessary to optimize the pre-culture time.

The orientation of the plant tissue explants placed on the medium is another important factor affecting the efficiency of plant transformation and regeneration [38]. The abaxial orientation of the leaf implies that the lower surface of the leaf is in contact with the medium, while the adaxial position means that the upper surface of the leaf is in contact with the medium [39]. In our pretest study, we found that more adventitious buds were produced during the regeneration stage when the abaxial side of the leaf was exposed to the medium than the adaxial side. Mazumdar et al. reported that the regeneration efficiency of the abaxial end of explants exposed to the medium was twice as high as that of the adaxial end [40]. Similarly, tomatoes of two different genotypes exhibited higher regeneration efficiency and higher shoot numbers per explant when placed abaxially than when exposed to medium in the adaxial direction [41], as confirmed in earlier studies [42,43].

The PEG/Mg²⁺ coating protocol in this study showed better stabilization of the transformation than the standard Spd/Ca²⁺ method. Due to the hygroscopicity, oxidizable nature, and deamination over time of spermidine (Spd) solutions, frozen aliquots should be replaced for fresh at least monthly [44]. Stock solutions of CaCl₂ and Spd must be used separately, whereas PEG and MgCl₂ can be prepared conveniently as a single stock solution that is stable for many years when stored at −20 °C. The PEG/Mg²⁺ procedure has been successfully applied to wheat for stable transformation [45].

The distance from the microcarrier to the target tissues can affect the velocity of the microparticles and thus the frequency of transformation [46]. As in previous studies, a target tissue distance of 9 cm was reported as the optimal propagation distance for bananas [47], wheat [48], and cumin [49]. We found that a target tissue distance of 9 cm was a significant factor in improving transformation frequency. This distance can reduce damage to a great extent and ensure the distribution of DNA microcarriers on target tissues [50].

Similar to Jähne et al., we found that changing the helium pressure in the range of 1100–1350 psi had no significant effect on the number of positive shoots [51]. In contrast to the results of this study, Jähne et al. found that lower (450–900 psi) or higher (2000–2200 psi) pressures reduced the frequency of transformation. When Harwood et al. used barley microspores pre-incubated for 1–4 days, they found that a lower pressure of 450 psi increased the number of GUS-positive microspores [52]. We also found that lower bombardment pressures increased the number of positive shoots. When Mentewab et al. compared the effect of using 650 psi or 1100 psi pressure, transient expression of microspores in culture for 1 day was observed only after using low pressure, while multicellular structures of 8 days were only observed when high pressure was applied [53]. It was demonstrated that the optimal helium burst pressure may depend on several factors related to the cell wall properties and the damaging effect of the treatment.

Comparing the number of surviving shoots, an overall optimal parameter was observed, i.e., precipitant agents MgCl₂ + PEG, target tissue distance 9 cm, helium pressure 650 psi, bombardment once, plasmid DNA concentration 1.0 μg·μL⁻¹, and chamber vacuum pressure 27 mmHg. In this study, the frequency of transformation obtained by biolistic transformation was 36.7%. The transformation efficiency of E. breviscapus was not reported for either Agrobacterium-mediated or particle bombardment transformation (three adventitious shoots were obtained). The higher transformation efficiency of the particle gun bombardment may be due to the attempted use of the hybrid orthogonal methods. Optimization strategies are highly efficient compared to the previous single-variable method, thus potentially facilitating genetic modification for improved traits.

4. Materials and Methods

4.1. Plant Material and Culture Conditions

Seeds of E. breviscapus were manually rubbed to remove their pappus, and the plump and undamaged seeds were selected for the experiment. Before inoculation, seeds were soaked for 4 h, dried for 3 h, and then sterilized. Under sterile conditions, seeds were soaked in 75% (v/v) ethanol for 8~10 s and washed 3~4 times with sterile water. After that, they were sterilized with 2.5% (v/v) NaClO for 8~10 min and rinsed with sterile water. Seeds were inoculated on Murashige and Skoog (MS) medium [54] hormone-free medium and cultured at 25 °C, 16 h/d light conditions of 4000 lx to obtain sterile seedlings.

4.2. Optimization of Explant Pre-Culture Time

Ten leaf explants were placed on each plate containing callus induction medium (CIM; MS + 1.0 mg·L⁻¹ 6-BA + 0.1 mg·L⁻¹ NAA + 3% sucrose + 0.4% phytagel, pH 5.8). A total of six conditions of 2 d, 3 d, 5 d, 7 d, 14 d, and 21 d were established in order to confirm the most suitable pre-culture time and the explants were incubated in the dark at 25 °C. After 4 h of osmotic treatment, the leaf explants were bombarded with the following parameters: 6 cm target distance, 27 mmHg, with 1100 psi rupture disks. The optimal pre-incubation time was evaluated by the frequency of transformed shoots. The methods of bombardment are described in 4.4 and 4.5.

4.3. Selection Pressure of Leaf Explants to Hygromycin B

The concentrations of Hyg were set at five levels of 0, 2.5, 5.0, 7.5, and 10.0 mg·L⁻¹, and the medium was CIM. Leaf explants were placed abaxially and inoculated with each treatment and replicated three times. All the explants were cultured in light at 25 °C for 4 weeks and sub-cultured every 2 weeks. The critical concentration of Hyg was confirmed by counting the induction rate of resistant callus.

4.4. Preparation of Gold Microprojectile

Plasmid PBI-1300 (provided by Pro. Meng Wang, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences) was isolated using the plasmid maxi kit (Omega, United States) according to the manufacturer’s protocol. Transformation conditions were confirmed by plasmid PBI-1300, which harbors the eGFP reporter gene and the hpt selectable gene, both driven by the cauliflower mosaic virus (CaMV) 35S promoter.

The gold particles must be sterilized before being used for DNA coating. An amount of 30 mg of 1.0 μm gold particles were added with 1 mL of 100% ice-cold ethanol and sonicated for 15 s, and were centrifuged at 3000 rpm for 60 s. The supernatant was carefully discarded and the gold particles were resuspended with 1 mL ice-cold ddH₂O. The gold particles were centrifuged at 3000 rpm for 60 s and supernatant was removed. The washing step of ddH₂O was repeated twice. Finally, the gold particles were suspended with 500 µL of 50% (v/v) sterile glycerin with a final concentration of 60 mg·mL⁻¹. The above microcarriers were stored at −20 °C before use.

The DNA/gold-coating method consists of two levels of optimization (Table 1). The first method is based on Bio-Rad protocol. An amount of 5 μL DNA (1 μg·μL⁻¹), 50 μL CaCl₂ (2.5 M), and 20μL Spd (0.1 M) were added to 50 uL microparticle solution, and the mixture was vortexed for 2~3 min and left to stand for 1 min. The supernatant was discarded after centrifugation. The particles were washed with 140 μL of 70% ethanol and absolute ethanol. Finally, the pelleted DNA was suspended with 48 μL of absolute ethanol. The second method corresponded to the previous method [45]. An amount of 50 μL of gold particles was coated with 10 μL of DNA (1 μg·μL⁻¹) and supplemented with 10 µL of PM solution (42% PEG 2000 and 560 mM MgCl₂) under vortexing. The mixture was vortexed for 1 min and incubated for 20 min at room temperature. The suspension was centrifuged for 1~5 min. Then, the pelleted DNA was washed with 100 µL of 100% ethanol and was resuspended in 60 µL of 100% ethanol.

4.5. Microprojectile Bombardment

The bombardments were performed with a particle gun (PDS 1000/He, Bio-Rad). The transformation conditions of the particle gun were optimized using one 2-level factor (precipitation agents) and five 3-level factors (target tissue distance, helium pressure, bombardment times, plasmid DNA concentration, and chamber vacuum pressure). These parameters and the levels of the variables studied are shown in Table 3. A hybrid orthogonal table L₁₈(2¹ × 3⁵) was designed according to the above six parameters and different levels of each parameter [55], as shown in

Table 6. L₁₈(2¹ × 3⁵) hybrid orthogonal table.

Group	A	B	C	D	E	F	Vacant Column
Group 1	1	1	1	1	1	1	1
Group 2	1	1	2	2	2	2	2
Group 3	1	1	3	3	3	3	3
Group 4	1	2	1	1	2	2	3
Group 5	1	2	2	2	3	3	1
Group 6	1	2	3	3	1	1	2
Group 7	1	3	1	2	1	3	2
Group 8	1	3	2	3	2	1	3
Group 9	1	3	3	1	3	2	1
Group 10	2	1	1	3	3	2	2
Group 11	2	1	2	1	1	3	3
Group 12	2	1	3	2	2	1	1
Group 13	2	2	1	2	3	1	3
Group 14	2	2	2	3	1	2	1
Group 15	2	2	3	1	2	3	2
Group 16	2	3	1	3	2	3	1
Group 17	2	3	2	1	3	1	2
Group 18	2	3	3	2	1	2	3

, where A to F represent precipitation agents, target tissue distance, helium pressure, number of bombardments, plasmid DNA concentration, and chamber vacuum pressure. Each factor was repeated three times, and the whole set of experiments was repeated twice.

4.6. Selection and Regeneration of Transformed Plants

The bombarded leaf explants were incubated in the dark at 25 °C for 10~16 h and then transferred to CIM. After 2 days, the transformed leaves were observed for expression of eGFP using Laser Scanning Confocal Microscope (LSM880NLO, Zeiss, German) under an excitation wavelength of 488 nm. The explants were then cut into small pieces of about 1 cm², placed abaxially on fresh CIM, and cultured at 25 °C, 16 h/d light. After 2 weeks, the transformed cells were observed for the expression of eGFP through LSCM under an excitation wavelength of 488 nm. The explants were then placed abaxially on CIM containing 2.5 mg·L⁻¹ Hyg. After three rounds of selection, the calluses were split into 3~5 pieces and transferred to the regeneration medium (RM; MS + 2.0 mg·L⁻¹ 6-BA + 0.2 mg·L⁻¹ NAA + 3% sucrose + 0.4% phytagel, pH 5.8) and incubated at 25 °C under 16 h/d light conditions of 4000 lx. After 2 weeks, green shoots were transferred to shoot elongation medium (SEM; 1/2MS + 2.0 mg·L⁻¹ 6-BA + 0.2 mg·L⁻¹ NAA + 3% sucrose + 0.4% phytagel, pH 5.8). For rooting formation, adventitious buds were transferred to root induction medium (RIM; 1/2MS + 0.3 mg·L⁻¹ NAA + 0.5 mg·L⁻¹ IBA + 3% sucrose + 0.4% phytagel, pH 5.8).

4.7. PCR Analysis

The genomic DNA of transformed plants was isolated using CTAB method [56]. PCR analysis was performed with hpt (selection maker gene)-specific primers (F: 5′-ATGAAAAAGCCTGAACTCACCG-3′; R: 5′-CTATTTCTTTGCCCTCGGACG-3′) to confirm the transformed strain. The denaturation temperatures used were 98 °C for 30 s, then 35 cycles for amplification with denaturation at 98 °C for 10 s, annealing at 60 °C for 30 s, extension at 72 °C for 20 s, and final extension at 72 °C for 7 min. Then, all the PCR products were sequenced by Sanger sequencing.

4.8. Statistical Analysis

The optimization experiment was repeated three times. The different levels of each experimental factor and the contribution rate of each factor were determined by range analysis and variance analysis. The statistical analysis of all data was calculated by formula. In the range analysis, the values of R and means of Ki were calculated:

$$\mathrm{R}=\max (\overline{{\mathrm{K}}_{\mathrm{i}}})-\min (\overline{{\mathrm{K}}_{\mathrm{i}}})$$

The range analysis of the hybrid orthogonal experiment was used to adjust the range R value and compared with the adjusted R’ value, where r is the number of replicates for each level of parameters, and d is the conversion coefficient, which is related to the parameter level:

$${\mathrm{R}}^{\prime }=\mathrm{dR}\sqrt{\mathrm{r}}$$

In the analysis of variance, the sum of squared deviations of the factors was calculated as follows:

$$SSj=\frac{1}{r}{\displaystyle \sum }_{i=1}^m{K}_{ij}^2-\frac{({{\displaystyle \sum }}_{i=1}^{n}Xi)2}{n}\left(j=1,2,\dots ,k\right)$$

In the analysis of variance, if the level of the factor is 2, the sum of squared deviations of the factors was calculated as follows:

$$\mathrm{SSj}=\frac{1}{\mathrm{n}}({\mathrm{K}}_{1\mathrm{j}}-{\mathrm{K}}_{2\mathrm{j}}{)}^2 (\mathrm{j}=1,2, ..., \mathrm{k})$$

The significance of difference was calculated by F-test (* p < 0.05; ** p < 0.01).