Five orchid species (Cymbidium, Phalaenopsis, Paphiopedilum, Dendrobium, and Arundina) were selected for this study. These orchids exhibited distinct morphological characteristics; for instance, the leaves of Phalaenopsis orchids were fleshy, whereas those of Cymbidium orchids were hard leather-like, and the leaves of Dendrobium and Paphiopedilum orchids were in between these. These orchid plants were obtained from the orchid breeding base of the Environmental Horticulture Research Institute, Guangdong Academy of Agricultural Sciences, China. They were grown in plastic pots (20 cm × 20 cm) in greenhouses under the exact growing conditions necessary for each species, as previously described (Yang et al., 2019). Maize and rice were also included, with maize and rice seeds obtained from the Rice Research Institute and Crops Research Institute of Guangdong Academy of Agricultural Sciences, respectively. The maize and rice seedlings were grown in plant growth rooms at 27°C under a 16 h light/8 h dark cycle and 60% humidity.

Protoplast isolation from the various plant species was conducted based on protocols established by Yoo et al. (2007) and Lin et al. (2018b) with some modifications. Various source materials collected from Cymbidium orchids, including flower pedicels, young leaves, leaf base tissues, and young root tips were tested for tissue- and organ-specific protoplast isolation. Additionally, leaf base tissues of Phalaenopsis, Paphiopedilum, Dendrobium, and Arundina orchids, and rice and maize plants were tested for protoplast isolation. Maize and rice plants were grown for 7 days after germination, and then the leaf bases of these seedlings were collected for protoplast isolation, whereas their young leaves served as controls.

The tissues were surface sterilized for 3 min in 75% ethyl alcohol (volume percentage, v/v), followed by five washes in sterilized water. The sterilized tissues were cut into 0.5–1.0 mm strips or slices using fresh surgical blades on sterile filter paper and were immediately transferred to a freshly prepared enzyme solution in a 100 mL sterile flask. The enzyme solution was prepared as follows: 20 mM KCl (Sigma, St. Louis, MO, United States), 20 mM 2-(N-morpholino)ethanesulfonic acid (pH 5.7) (Sigma) with different concentrations of D-mannitol (0.2, 0.3, 0.4, 0.5, 0.6, and 0.7 M) (Sigma), cellulose R-10 (0.6, 1.2 and 2.4, w/v), and macerozyme R-10 (0.3, 0.6, and 1.2%, w/v) (Yakult Pharmaceutical Industry Ltd., Nishinomiya, Japan). The solution was then warmed up to 55°C for 10 min, and cooled to room temperature (∼25°C). Then, 10 mM CaCl₂ (Sigma), and 0.1% (w/v) bovine serum albumin (Sigma) were added to different concentrations of 2-mercaptoethanol (0, 1, 2, 5, and 10 mM) (Sigma). For each treatment, 10 mL enzyme solution was prepared for approximately 0.5 g fresh tissues. After incubation at 28°C in the dark with a rotation of 30 rpm for different periods (2, 4, 6, and 8 h), the released protoplasts were harvested. Generally, the enzyme mixture containing protoplasts was diluted with an equal volume of wash solution (W5) that contained 154 mM NaCl (Sigma), 125 mM CaCl₂, 5 mM KCl, and 4 mM MES (pH = 5.7). The protoplast-containing solution was filtered through a 150 μm nylon mesh into a 50 mL round-bottomed centrifuge tube and centrifuged at 100 × g for 2 min to pellet the protoplasts (at room temperature ∼25°C). Subsequently, the supernatant was removed carefully as much as possible with a sterile syringe. The protoplast pellet was re-suspended in 20 mL of W5 solution (Negrutiu et al., 1986), and then the yield and viability of protoplasts were estimated.

The protoplasts were counted and photographed using a Leica DM2500 microscope (Leica, Wetzlar, Germany) with a hemocytometer to calculate the estimated yield of protoplasts. The protoplast yield was calculated as the total number of protoplasts released in the enzyme mixture divided by the FW of tissues used for protoplast isolation (protoplasts/g FW). Fluorescein diacetate (FDA) (Sigma) staining was applied to determine the viability of protoplasts as follows: 9 μL of the protoplast-containing solution was placed onto a glass slide, followed by the addition of 1 μL of 0.2% FDA solution [dissolved in acetone (Solarbio Science and Technology Co., Ltd. Beijing, China)] and incubated at room temperature for 1–2 min. Viable protoplasts with green fluorescence were visualized and photographed using an LSM 710 confocal laser microscope (Carl Zeiss, Inc., Jena, Germany) with blue excitation block. Protoplast viability was measured as the number of protoplasts with green fluorescence divided by the total number of protoplasts ×100%. Three images were selected for the yield and viability measurements of each sample. Each experiment was repeated three times.

To test the feasibility of orchid leaf base protoplasts for protein subcellular localization and gene regulation analysis, vector plasmids expressing green fluorescent protein (GFP), GFP-protein fusions, proteins, and gene-specific target sequence were prepared for PEG-mediated protoplast transfection (Supplementary Figure 1A). The empty vector pAN580-GFP expressing GFP was used as a control for the protein subcellular localization. Plant organelle markers, including pGreenII62-SK-AtWAK2-GFP and pGreenII62-SK-AtPIP2A-GFP, were used in the present study. The two AtWAK2 and AtPIP2A signal peptides isolated from Arabidopsis thaliana were specifically accumulated in the endoplasmic reticulum and plasma membrane, respectively (Gomord et al., 1997; Cutler et al., 2000).

The recombinant plasmid pAN580-CsDELLA-GFP was obtained by inserting the full-length coding sequence (CDS) of CsDELLA (GenBank accession number: MK282635.1) without a termination codon between the dual cauliflower mosaic virus 35S promoter and the GFP gene of the vector pAN580-GFP. Meanwhile, the full-length CDS of CsDELLA was cloned into vector pAN580-GFP, and replaced with the GFP gene, resulting in the pAN580-CsDELLA plasmid (Supplementary Figure 1B). For the gene silencing vector pTCK303-CsDELLA-RNAi, the 243 base pair sequence of CsDELLA was cloned into the vector pTCK303-RNAi with an Ubiquitin promoter (Supplementary Figure 1C). Briefly, total RNA was extracted from freshly collected young leaves of Cymbidium sinense using an RNA Simple Total RNA Kit (Tiangen, Beijing, China). The DNA-free RNA was used for first-strand cDNA synthesis with oligo (dT) primers and a PrimeScript^TM 1st strand cDNA Synthesis Kit (Takara, Dalian, China), following the manufacturer’s instructions. Specific primers with overlapping homologous ends were designed using the Primer Premier 5.0 software (Premier, Palo Alto, CA, United States) according to the full-length CDS and gene-specific target sequence of CsDELLA (Supplementary Table 1). Subsequently, the fragment of CsDELLA was amplified using PrimerSTAR Max DNA Polymerase (Takara) with the cDNA and specific primers. The polymerase chain reaction (PCR) products were purified and cloned into certain vectors using a Seamless Assembly Cloning Kit (CloneSmarter, Houston, TX, United States) following the manufacturer’s instructions.

Vectors including pGreenII62-SK-AtWAK2-GFP, pGreenII62-SK-AtPIP2A-GFP, pAN580-GFP, pAN580-CsDELLA-GFP, pAN580-CsDELLA, and pTCK303-CsDELLA-RNAi were transformed into Escherichia coli DH5α competent cells (Tiangen) according to the manufacturer’s instructions, respectively. Each single E. coli colony was picked and inoculated in 50 mL Luria-Bertani broth [Tryptone (1.0%, w/v), Yeast extract (0.5%, w/v), and NaCl (1.0%, w/v), pH = 7.0] containing the appropriate antibiotic. Bacterial cells were pelleted by centrifuging cultures at 3,000 × g for 10 min following culturing at 37°C and 200 rpm shaking for 12–16 h. Following mass replication of the bacterium, plasmid DNA was extracted using an Endo-Free Plasmid Maxi Kit (Omega Bio-tek, Norcross, GA, United States). The concentrated plasmid DNA was prepared at different concentrations (up to 2.0 μg/μL), and was transformed into protoplasts.

Following yield and viability measurements, protoplasts were further purified before transfection and transient expression. The 20 mL protoplast-containing W5 solution was re-centrifuged at 100 × g for 2 min (at room temperature). Following resuspension in 20 mL W5 solution, protoplasts were placed on ice for 30 min. Viable protoplasts settled at the bottom of the tube by gravity. Finally, the supernatant was carefully removed as much as possible, and the purified protoplasts were adjusted to a density of 1 × 10⁵ – 1 × 10⁶/mL with the pre-chilled MMG solution (Negrutiu et al., 1987). The MMG solution comprised 0.5 M D-mannitol, 15 mM MgCl₂, and 4 mM MES (pH = 5.7).

Protoplast transfection was conducted following a modified PEG-mediated protocol (Yoo et al., 2007). For each transfection, 100 μL of MMG-protoplast mixture was gently mixed with 10 μL of pre-chilled plasmid DNA in 2 mL round-bottomed centrifuge tubes. Then, an equal volume (110 μL) of freshly prepared PEG-CaCl₂ solution was added and immediately mixed by gentle inversion. The PEG-CaCl₂ solution was composed of 100 mM CaCl₂, 0.2 M D-mannitol, and 25% PEG4000 (final concentration, w/v) (Sigma). The mixture was incubated at room temperature in the dark for 10 min. The transfection was stopped by adding two volumes of W5 solution (440 μL) followed by centrifugation at 100 × g for 2 min. The transfected protoplasts were then washed with W5 solution and re-suspended in 1 mL WI solution for each of the 6-well tissue culture plates. The WI solution was composed of 0.4 M D-mannitol, 20 mM KCl, and 4 mM MES (pH = 5.7). For transient expression of genes/proteins, the protoplast mixture was incubated in the dark at 23°C for 12–36 h.

After incubation in the dark at 23°C for 12–24 h, the fluorescence of GFP or GFP-protein fusions was viewed under an LSM710 confocal laser scanning microscope (Carl Zeiss, Inc.). Transformation efficiency of the protoplasts was determined based on the GFP-reporter expression using the transient expression vector pAN580-GFP. GFP fluorescence was observed and 3–5 images were taken randomly under an LSM780 fluorescent microscope (Carl Zeiss, Inc.) or LSM710 confocal laser scanning microscope. The transformation efficiency was measured as a bright green fluorescent protoplast number in view/total protoplast number in view (%). At least three photographs were taken for each sample, and these experiments were independently conducted at least three times. For subcellular localization analysis, plant organelle markers pGreenII62-SK-AtWAK2-GFP and pGreenII62-SK-AtPIP2A-GFP, and the control vector pGreenII62-SK-GFP were transfected into orchid leaf base protoplasts. Red chlorophyll fluorescence was used to indicate the intercellular location of chloroplasts.

Protoplast treatments were undertaken in W5 solution supplemented with different concentrations of gibberellin (GA3) (1, 10, and 100 μM). Approximately 10⁷ protoplasts in 1 mL W5 solution were used for each treatment. W5 solution supplemented with water was used as the control treatment at each time point. Protoplasts were cultivated in 6-well plates in a growth chamber at 23°C in the dark for 6–24 h.

Protoplasts treated with GA3 were harvested at 6, 12 and 24 hours post treatment (hpt), and transfected protoplasts were harvested at 12, 24 and 36 hours post transfection (hpt). First-strand cDNA was obtained from the total RNA of the harvested protoplasts. Gene-specific primers were designed according to the CDSs of CsDELLA, SUPPRESSOR OF OVEREXPRESSION OF CONSTANS1 (CsSOC1, MF474250.1), FLOWERING LOCUS T (CsFT, HM803115.1), SHORT VEGETATIVE PHASE3 (CsSVP3, MF462098.1), CsLFY (KC138806.1), and APETALA1 (CsAP1, JQ326260.1) using Primer Premier 5.0 software (Premier, Palo Alto, CA, United States). CsUbiquitin in Cymbidium (referred to as CsUBQ, AY907703) was used as an internal reference control to normalize the total amount of cDNA in each reaction (Supplementary Table 1).

The quantitative real-time polymerase chain reaction (qRT-PCR) was performed in a 20 μL reaction volume comprising 2.0 μL of 5× diluted first-strand cDNA (approximately 20 ng), 0.8 μL of each primer (10.0 μM), 10.0 μL of 2× SYBR Green I Master Mix (Takara), and 6.4 μL of sterile distilled H₂O. All reactions were performed in 96-well reaction plates using a Bio-Rad CFX-96 Real-time PCR System (Bio-Rad, Hercules, Canada) with three technical replicates. The following PCR conditions were used: 95°C for 5 min, followed by 40 cycles at 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s, and then at 68°C for 5 min. The expression of candidate genes was quantified using the relative quantification (2^–ΔΔCT) method. Each sample was collected independently with three biological replicates.

Statistical analysis was performed using one-way analysis of variance with SPSS version 18.0 software (SPSS Inc., Chicago, IL, United States). Data are presented as the mean ± standard error from three independent experiments. The least significant differences among treatments were determined at the 5% level of probability with Duncan’s multiple range tests.

To establish an efficient protoplast isolation protocol for orchids, various source materials, including flower pedicels, young leaves, leaf base tissues, and young root tips of Cymbidium orchids were tested (Figure 2A). Viable protoplasts were successfully isolated from various tissues or organs of Cymbidium plants following the protocol described by Yoo et al. (2007) (Figure 2B). However, the yield and viability of released protoplasts were relatively low, and abundant calcium oxalate crystal raphides were observed in the protoplast-containing solution (Supplementary Figure 2).

To reduce the protoplast rupture or collapse during enzymatic digestion, gradient concentrations of D-mannitol in the enzyme solution (0.2, 0.3, 0.4, 0.5, 0.6, and 0.7 M) were included for different tissues. We found that the D-mannitol concentration significantly affected the efficiency of protoplast isolation. Generally, increasing D-mannitol concentrations resulted in increased protoplast yields, whereas exceeding osmotic pressures resulted in decreased yields. The optimal D-mannitol concentrations for various tissues gave the highest yield and viability simultaneously. For flower pedicels, leaf base, young leaves, and root tips of Cymbidium orchids, the optimal D-mannitol concentrations were determined as 0.6, 0.6, 0.4, and 0.5 M, respectively. With the optimal D-mannitol concentrations, the orchid leaf base gave the highest yield of viable protoplasts (∼6.67 × 10⁶/g FW), followed by the flower pedicels (∼3.67 × 10⁶/g FW), young leaves (∼1.07 × 10⁶/g FW), and root tips (∼2.97 × 10⁵/g FW) (Figure 2C). In addition, protoplasts isolated from young leaves and root tips remained intact for more than 24 h in the W5 solution (Supplementary Figure 2), whereas that of flower pedicels ruptured within 6 h post-isolation. Most protoplasts isolated from the leaf base tissues (>70%) remained intact for up to 3 days in the WI solution (Nagy and Maliga, 1976), which was stronger and longer-term healthy than that of the other tissues/organs. Therefore, the leaf base was a better source material for protoplast isolation from Cymbidium orchids.

More factors affecting protoplast isolation were investigated to increase the yield and viability of protoplasts isolated from the leaf base of Cymbidium orchids. An increase in total enzyme concentrations increased the protoplast yield, whereas excess enzymes resulted in a decrease in protoplast yield and viability. The optimal combination of enzymes was 1.2% (weight/volume, w/v) of cellulose and 0.6% (w/v) of macerozyme, which gave the highest yield (6.97 × 10⁶/g FW) and viability (∼88.49%) of leaf base protoplasts (Figure 3A). Moreover, insufficient digestion time led to incomplete release of protoplasts, whereas excess incubation resulted in the rupture of protoplasts. With optimal D-mannitol concentration and enzyme combination, the highest protoplast yield (6.40 × 10⁶/g FW) and viability (∼92.76%) were obtained after 6 h of enzymatic digestion (Figure 3B). Further optimization revealed that the addition of 2-mercaptoethanol contributed to decreased calcium oxalate raphides and increased yield and viability of protoplasts. Given the combination of optimal D-mannitol concentration (0.4–0.6 M), enzyme mixture (1.2% cellulose and 0.6% macerozyme), digestion time (6 h), and 2-mercaptoethanol (5 μM), the highest Cymbidium protoplast yield (∼2.50 × 10⁷/g FW) and viability (∼92.09%) were achieved (Figure 3C).

With this protocol, higher yields of viable protoplasts were isolated form young leaves (∼3.22 × 10⁶/g FW), flower pedicels (∼5.26 × 10⁶/g FW), and young root tips (∼7.66 × 10⁵/g FW) (Figure 3D). The isolated protoplasts ranged from 20 to 100 μm in diameter, and a large proportion of protoplasts isolated from the young leaves and flower pedicels were rich in cytoplasm and anthocyanidin. Protoplasts isolated from the leaf base and root tips were almost translucent with few chloroplast auto-fluorescent signals. Therefore, our protoplast isolation protocol is suitable for investigating intracellular processes in tissue- and organ-specific protoplasts of Cymbidium orchids.

To test the suitability of our protoplast isolation protocol for other species, leaf base tissues collected from Phalaenopsis, Paphiopedilum, Dendrobium, and Arundina orchids were tested with different D-mannitol concentrations (0.2–0.7 M) in enzyme solution (Figure 4A). High yielding protoplasts was obtained from these orchids, and optimal D-mannitol concentrations were determined to be 0.4–0.6 M (Figure 4B). The maximum yields of viable protoplasts were isolated from Phalaenopsis (1.83 × 10⁷/g FW) and Paphiopedilum (1.10 × 10⁷/g FW), which were higher than those of Dendrobium (8.21 × 10⁶/g FW) and Arundina (3.78 × 10⁶/g FW) (Figure 4B). Protoplasts isolated from Phalaenopsis and Paphiopedilum (ranged from 40 to 100 μm in diameter) were larger than those isolated from Dendrobium and Arundina (approximately 20–60 μm in diameter) (Figure 4C). Thus, our protocol exhibits broad suitability in efficient protoplast isolation for orchids exhibiting distinct morphological characteristics.

Sharing similar anatomical characteristics with orchids, economically important monocot crops maize and rice were also tested for the suitability of our protoplast isolation protocol. Leaf base tissues and young leaves were all used, and leaf base protoplast isolation method showed marked advantages over previous mesophyll protoplast isolation protocols (Table 1). High yielding viable protoplasts were released from the leaf base of maize (3.25 × 10⁷/g FW) and rice (4.31 × 10⁷/g FW), and their optimal D-mannitol concentrations were both determined to be 0.5 M (Figure 5). The leaf base protoplasts isolated from maize (ranging from 20 to 60 μm in diameter) were smaller than those of rice (20–90 μm) (Supplementary Figure 3). The majority of rice protoplasts were transparent and highly viable. The efficiency of protoplast isolation from the leaf base tissues was higher than that of the leaves (Figure 5 and Table 1). It indicated that our leaf base protoplast isolation protocol enables highly efficient protoplast isolation from many other monocot plants.

To test the feasibility of the leaf base protoplasts for gene transient expression, PEG-mediated Cymbidium protoplast transfection was performed using the vector pAN580-GFP. According to the GFP expression in the transfected protoplasts, a maximum transfection efficiency of more than 80% was obtained by optimizing factors affecting PEG-mediated transfection, including incubation time, final PEG4000 concentration, and amount of plasmid DNA (Figure 6A). In addition, most of the transfected protoplasts maintained integrity up to 36 hpt. These results indicate that leaf base protoplasts of orchids are suitable for transient gene/protein expression.

The efficient transient expression system established in the present study was tested further for the feasibility of the subcellular localization of proteins. Empty control vector pAN580-GFP, the two plant intercellular organelle markers (pGreenII62-SK-AtWAK2-GFP and pGreenII62-SK-AtPIP2A-GFP), and the recombined vector pAN580-CsDELLA-GFP were transfected into Cymbidium leaf base protoplasts, respectively. A total of 12–16 h later, green fluorescence was observed in the intracellular compartments of the transfected protoplasts. Fusion proteins AtWAK2-GFP and AtPIP2A-GFP were detected specifically in the endoplasmic reticulum and plasma membrane, respectively, whereas the fusion protein CsDELLA-GFP and vector control were distributed throughout the entire cell (Figure 6B). Thus, the leaf base protoplasts are suitable for molecular and cellular characterization of genes/proteins.

To test the feasibility of the Cymbidium protoplast system in cellular and molecular studies, functional identification of CsDELLA was conducted in Cymbidium protoplast. As DELLA-mediated GA signaling is known to affect flowering (Yamaguchi et al., 2014; Bao et al., 2020), we analyzed the role of CsDELLA in the regulation of GA to flowering-related genes. To determine the optimum concentration, the Cymbidium protoplasts were treated with different concentrations of GA3 (1, 10, and 100 μM), and the expression of CsDELLA in Cymbidium protoplasts was examined by qRT-PCR. CsDELLA expression was suppressed by the addition of GA3, and 10 μM GA3 was the best concentration for protoplast treatment (Figure 7A). To determine the reliability and repeatability, the viability of protoplasts post the treatment was estimated with FDA staining, and most protoplasts remained viable up to 24 hpt (Figure 7B).

Protoplasts treated with 10 μM GA3 and H₂O control were collected at 6, 12, and 24 hpt, and the expressions of CsDELLA and flowering-related genes were analyzed by qRT-PCR. It was resulted that CsDELLA expression was significantly downregulated 3–6-fold from 12 to 24 hpt in the presence of 10 μM GA3 compared with controls at each time point (Figure 7C). However, the three key flowering-related genes CsFT, CsSOC1, and CsAP1 were all significantly upregulated 1.3–15-fold from 12 to 24 hpt in response to 10 μM GA3 (Figure 7D). Whether GA promotes the expression of flowering related genes by suppressing the expression of CsDELLA remains unclear, and the role of CsDELLA in the regulation of GA to flowering-related genes needs further investigation. These results indicate the potential of our protoplast system for investigating cellular and molecular behaviors in response to phytohormones and other inducing factors.

To analyze the regulation of CsDELLA to flowering-related genes, CsDELLA was transiently over-expressed or knocked-down in Cymbidium protoplasts. The expression of flowering-related genes was examined with qRT-PCR. The recombinant plasmid pAN580-CsDELLA was transfected into Cymbidium protoplasts, and CsDELLA was successfully overexpressed 90–350-fold from 12 to 24 hpt (Figure 8A). The expressions of CsSOC1 and LEAFY (CsLFY) were significantly downregulated 5–10-fold from 12 to 24 hpt, while those of CsFT, CsSVP3, and CsAP1 were slightly downregulated. It indicated that the transient overexpression of CsDELLA suppressed the expression of CsSOC1 and CsLFY (Figure 8B).

To identify the regulation relationships, CsDELLA was silenced with the RNA interference (RNAi) technology. The 243-bp CsDELLA-specific sequence was inserted into the RNAi vector pTCK303, and the vector pTCK303-CsDELLA-RNAi was transfected into Cymbidium protoplasts. It was resulted that the vector pTCK303-CsDELLA-RNAi (approximately 14.8 kb) was expressed in Cymbidium protoplasts, and CsDELLA was silenced to almost 70–80% (Figure 8C). The expression of CsSOC1 was significantly upregulated 3–30-fold, and those of CsFT was greatly upregulated 50–160-fold from 12 to 24 hpt. However, CsSVP, CsLFY, and CsAP1 were significantly downregulated (Figure 8D). These results indicated that silenced CsDELLA promoted the expression of CsSOC1 and CsFT, CsDELLA probably positively regulates the expression of CsSOC1 and CsFT. Our leaf base protoplast transfection system was confirmed to conveniently analyze gene expression regulation relationships, and to screen efficient vectors before genetic transformation.