A Rapid, Highly Efficient and Economical Method of Agrobacterium-Mediated In planta Transient Transformation in Living Onion Epidermis

Transient transformation is simpler, more efficient and economical in analyzing protein subcellular localization than stable transformation. Fluorescent fusion proteins were often used in transient transformation to follow the in vivo behavior of proteins. Onion epidermis, which has large, living and transparent cells in a monolayer, is suitable to visualize fluorescent fusion proteins. The often used transient transformation methods included particle bombardment, protoplast transfection and Agrobacterium-mediated transformation. Particle bombardment in onion epidermis was successfully established, however, it was expensive, biolistic equipment dependent and with low transformation efficiency. We developed a highly efficient in planta transient transformation method in onion epidermis by using a special agroinfiltration method, which could be fulfilled within 5 days from the pretreatment of onion bulb to the best time-point for analyzing gene expression. The transformation conditions were optimized to achieve 43.87% transformation efficiency in living onion epidermis. The developed method has advantages in cost, time-consuming, equipment dependency and transformation efficiency in contrast with those methods of particle bombardment in onion epidermal cells, protoplast transfection and Agrobacterium-mediated transient transformation in leaf epidermal cells of other plants. It will facilitate the analysis of protein subcellular localization on a large scale.


Introduction
Onion (Allium cepa L.), one kind of biennial herb Liliaceae plant, has been used as classical experimental materials in analyzing structure of plant cells, distribution location of DNA and RNA, reducing sugar of plant tissues [1], plasmolysis and recovery of plant cells [2,3], karyotype [4], protein subcellular localization and interaction [5][6][7].
Imaging subcellular localization of proteins in living cells has become an important tool for defining protein function. Fluorescent fusion proteins are ideal marker non-enzymatic protein systems for imaging protein subcellular localization in living cells, which have many apparent advantages, such as stable fluorescence properties, easy observation, visualization in living cells, non-toxic to cells, non-specifity for species, without interference to false positives and no substrate etc. In addition, expression of fluorescent fusion proteins had also been used to investigate protein interaction, trafficking, turnover, movement and inheritance in living cells [8].
However, most transient transformation methods have certain disadvantages, such as the lower transformation efficiency, equipment dependency and auxiliary material needed for particle bombardment, complex preparation procedures required for protoplasts transfection. In addition, for transient transformation in plants having complex outline of epidermal cells and the bushy epidermal hairs, laser scanning confocal microscope was usually needed to get ideal micro-images, which increased the reliability on expensive equipments. To avoid these disadvantages, we developed an in planta transient transformation method in living onion epidermal cells by using Agrobacterium-mediated infiltration. In this protocol, infiltration liquid of Agrobacterium carrying constructed vectors were injected into the interface between adaxial epidermis and mesophyll of onion bulb scales, which played an important role in yielding high transformation efficiency, and kept in the living onion bulb for about three days.
With this simple method a higher frequency of transformation was achieved without expensive equipments in comparison with other transient transformation methods like protoplast transfection of Spinacia oleracea [33], A. thaliana [34] and Populus euphratica [35], agroinfiltration of Arabidopsis epidermal cells and particle bombardment of onion cells in vitro. In addition, following this method it took about 3 days for agroinfiltration to get ideal transient transformation efficiency in living onion epidermis. Therefore, the developed method was rapid, highly efficient, equipment independent and low-cost, which will benefit to analyze protein subcellular localization in a large-scale manner.

Special agroinfiltration method could benefit transient transformation
We first modified the agroinfiltration method previously used in Nicotiana tabacum [36,37] in order to increase the efficiency of transient transformation. About 200 ml agroinfiltration liquid was slowly injected into the interfaces between adaxial epidermis and mesophyll of onion bulb scales by using a plastic syringe with needle, which resulted in an agroinfiltration bubble at the injection spot filled with infiltration liquid occupying about 1 cm 2 area of epidermis ( Figure 1). The formation of agroinfiltration bubble was demonstrated in the schematic diagram ( Figure 2). We found that 1 cm 2 agroinfiltration area with forming bubble gave higher transformation efficiency than a larger agroinfiltration area without forming bubble. The reason could be that forming agroinfiltration bubble filled with agroinfiltration liquid gave the infiltrated epidermal cells more chance to be infected by agrobacteria because the ratio of agrobacteria to infiltrated epidermal cells was higher. The special injection method could contribute to the high transformation efficiency of the developed transient transformation method. In contrast with other plant material, the adaxial epidermis and mesophyll of onion bulb scales are more flexible to be separated to form an interface bubble, which can take more agroinfiltration liquid.

Optimization of optical density (OD) of bacteria and agroinfiltration duration
The OD of bacteria and agroinfiltration duration were optimized. It showed that among the different agroinfiltration durations (24 h, 48 h, 72 h and 96 h), three-day agroinfiltration gave the significantly higher transformation efficiency (Table 1). While among three different concentrations of bacteria, agroinfiltration liquid with OD 600 of 0.10 resulted in significantly higher transformation efficiency (Table 1). Considering both elements, the conditions of agroinfiltration liquid with OD 600 of 0.10 and three-day infiltration were adopted to achieve the highest transformation efficiency of 43.87% (Table 1).
Agroinfiltration liquid with special combination of components contributed to the high efficiency of transient transformation The components of infiltration liquid and transfection conditions were referenced to relevant reports [13,[38][39][40][41][42] and with special modification. Based on agroinfiltration liquid for floral dip stable transformation and other transient transformation assays, the modified agroinfiltration liquid included 6-benzylaminopurine (BAP), Silwet L-77, D-glucose, acetosyringone (AS), magnesium chloride hexahydrate (MgCl 2 ), calcium chloride dehydrate (CaCl 2 ) and MES-KOH. To our knowledge, adding plant growth regulator BAP into the agroinfiltration liquid for transient transformation was seldom, although it had previously been applied in floral dip stable transformation. As an analogue of cytokinin, BAP plays roles in promoting cell division and prevent cell senescence, adding of low concentration of BAP in the agroinfiltration liquid could benefit to keep onion epidermal cells vigorous during agroinfiltration. Surfactant, glucose, osmotic buffer, calcium and magnesium ions, and acetosyringone were also supplemented to benefit the transient transformation.
In order to confirm the necessary roles of all the components, the transformation efficiencies of different infiltration liquids without one of the components were investigated ( Table 2). The results indicated that all the components of agroinfiltration liquid were necessary and significantly contributed to the high efficiency of Agrobaterium-mediated transient transformation in onion epidermal cells.
The developed method is advantageous to particle bombardment and onion is more suitable for the developed method than tobacco and Arabidopsis Through analyzing the side by side experiments of transforming pLPGM413 into onion epidermal cells with the developed method and particle bombardment, it indicated that the developed method resulted in significantly high transformation efficiency and showed advantages on cost and equipment needed ( Table 3).
The side by side experiments of Agrobacterium mediated transient transformation in onion, tobacco and Arabidopsis [14] with the same agroinfiltration liquid showed that the developed transient transformation method of onion epidermis gave significantly high transformation efficiency than tobacco and Arabidopsis ( Table 4). The developed method of transformation in onion epidermis took only 5 days from the pretreatment of onion bulb to the best timepoint for visualizing FFP signals, which is shorter than those in tobacco and Arabidopsis if considering the preparation time of plants [14]. In addition, the developed method is easier to get ideal images because onion epidermis has transparent cells arranged in a monolayer. Therefore, the developed method of onion has advantage to those of tobacco and Arabidopsis.

Pretreatment of onion bulb benefited the Agrobacterium mediated transient transformation in onion epidermis
The effects of different pretreatment times of onion bulb before injection of agroinfiltration liquid were evaluated (Table 5). It indicated that with the increase of pretreatment time the transformation efficiencies were increased (Table 5). However, 48-h and 72-h pretreatments resulted in similar transformation efficiencies, which are significantly higher than those resulted from 0-h and 24-h pretreatments (Table 5). It implied that 48-h pretreatment could be the suitable pretreatment time, which helped to achieve the highest transformation efficiency with the least time ( Table 5).

Monitoring of transformation and subcellular localization confirmation
The process of transient transformation was monitored by using four different plasmids carrying report genes mediated by A. tumefaciens strain GV3101, which constructed as the following expression binary vectors, pLPGM413 carrying RcSERK1 (somatic embryogenesis receptor-like kinase 1 of Rosa canina) [43] and GFP (green fluorescent proteins), pLPGM413 modified from pSAT6-GFP-N1 [44][45][46][47] by adding T-DNA border region of pCAMBIA1303, pLPGM113 modified from pEZS-NL-GFP [48] by adding T-DNA border region of pCAMBIA1303 and pCM1205-RFP (red fluorescent protein) carrying RFP [49,50]. The Agrobacterium cells harboring above vectors were grown in liquid yeast extract and beef extract medium (YEB) until stationary phase and then resuspended in infiltration liquid with above special components. Two to three days after injection of infiltration liquid, obvious GFP/RFP signals corresponding to different report genes could be observed in epidermal cells (Figure 1). It indicated that the developed agroinfiltration method can efficiently mediate in vivo transient expression of genes in living onion epidermal cells. For pCM1205-RFP, pLPGM202 and pLPGM113, the proteins encoded by their harboring genes were localized throughout the cells (Figure 1), and the fusion protein encoded by RcSERK1-GFP in pLPGM413 was localized in cell membrane ( Figure 1). The results were the same as those of previous studies [43][44][45][46][47][48][49][50], suggesting the authenticity of the developed method.
The fluorescence signals of the nuclei in the cells transformed by pCM1205-RFP, pLPGM202 and pLPGM113 overlapped with those nuclei stained by 49,6-diamidino-2-phenylindole (DAPI) (Figure 1), while no fluorescence signals were observed in nuclei of the cells transformed by pLPGM413. It further confirmed the credibility of the developed transient transformation method. It indicated that the developed transient transformation method in onion epidermal cells could be used to determine subcellular localization of the target proteins with the consistence to other transient transformation systems.

Conclusions
We have developed a rapid and low-cost transient transformation assay by using in-planta agroinfiltratoin in living onion epidermal cells, which transformation efficiency is higher than those of the conventional transient assays. It demonstrated that subcellular localization of proteins evaluated by using this method was same as that by other transient transformation assays and DAPI staining. Since it is rapid, efficient and low-cost, this method is suitable for large-scale analyses of protein subcelluar localization.

Plant material
Cultivated onion ''Hongtaiyang'' (A. cepa L.), which is a kind of Chinese onion cultivar with red skin, was used as experimental

GFP/RFP-based constructs
The binary expression vector pLPGM202 harboring GFP genes was modified from pSAT6-GFP-N1 provided by Prof. Tao Wang (State Key Laboratory of Agro-biotechnology, China Agricultural University) by adding T-DNA border region of pCAMBIA1303, which contains the right border, kanamycin (Kan) resistance gene for bacterial resistance selection and the left border. The steps of creating pLPGM202 was as follows: the T-DNA border region (2888-9172) of pCAMBIA1303 was amplified by using PrimeS-TARH GXL DNA Polymerase (TaKaRa, Japan) with PCR primers of Fsat-YR (NotI-GTAAACCTAAGAGAAAAGAG) and Rsat-YR (PI-PspI-TTTGCCTGTTTACACCACAAT) carrying the restriction enzyme cutting sites of NotI and PI-PspI, respectively. The PCR products were ligated into pSIMPLE-19 EcoRV/BAP vector by T4 DNA Ligase (TaKaRa, Japan) and sequenced to confirm the success of ligation. The constructed recombinant pSIMPLE-19 EcoRV/BAP with the T-DNA border region of pCAMBIA1303 was transformed into Escherichia coli DH5a for propagation. The pSAT6-GFP-N1 were digested with partial digestion method by using NotI and PI-PspI (NEB, USA), the target digested pSAT6-GFP-N1 was obtained with electrophoresis method. The T-DNA border region of pCAMBIA1303 was obtained from the recombinant pSIMPLE-19 EcoRV/BAP through complete digestion with NotI and PI-PspI, and cloned into the target partially-digested pSAT6-GFP-N1 to form pLPGM202 ( Figure S1). The successful construction of pLPGM202 was confirmed by sequencing (Sangon Biotech, China).
The binary expression vector pLPGM113 was modified from pEZS-NL-GFP provided by Dr. Liang Zhang (Department of Life Science, Henan Normal University) by adding T-DNA border region (2888-9172) of pCAMBIA1303 same as that for pLPGM202 described above. The steps of creating pLPGM113 were similar as those of pLPGM202 to be simply described as follows: The T-DNA border region of pCAMBIA1303 was amplified with the primers Fnl-YR (PstI-GTAAACCTAAGA-GAAAAGAG) and Rnl-YR (SpeI-TGTTTACACCACAATA- TATCC) carrying the restriction enzyme cutting sites of PstI and SpeI, respectively. The PCR products were ligated into pSIMPLE-19 EcoRV/BAP vector by T4 DNA Ligase (TaKaRa, Japan) and propagated in E. coli DH5a. The pEZS-NL-GFP were digested with partial digestion method by using PstI and SpeI (NEB, USA), the target digested pEZS-NL-GFP was obtained with electrophoresis method. The T-DNA border region of pCAMBIA1303 was obtained from the recombinant pSIMPLE-19 EcoRV/BAP through complete digestion with PstI and SpeI, and cloned into the target partially-digested pEZS-NL-GFP to form pLPGM113 ( Figure S2). The successful construction of pLPGM113 was confirmed by sequencing (Sangon Biotech, China). The binary expression vector pCM1205-RFP harboring RFP gene was provided by Dr. Wencai Qi (Bioengineering Department, Zhengzhou University).
The cDNA of RcSERK1 was isolated from cDNA library of R. canian PLBs and its sequence was deposited in GenBank as accession number of HM802242. Coding sequence of this gene (without stop codon) was amplified by PCR using Platinum pfx DNA polymerase (Invitrogen, USA) with the following primers: RcSERK1 (F-SKs, forward: 59-CCCAAGCCTCATGGATAG-CAGGCTT-39; R-SKs, reverse: 59-TCCCCCGGGCCTTG-GACCAGATAAC-39). The PCR products were digested with HindIII and SmaI, subsequently ligated into the HindIII/SmaI sites between the CaMV 35S promoter and GFP of pLPGM202 to construct the expression binary vector pLPGM413.

Agrobacterium infiltration
The vectors, pCM1205-RFP, pLPGM113, pLPGM202 and pLPGM413, were transformed into A. tumefaciens strain GV3101 for further transient transformation in onion epidermis. Positive Agrobacterium harboring pCM1205-RFP was selected and cultivated in YEB media supplemented with 100 mg/L rifampicin and 25 mg/L chloramphenicol. Positive Agrobacterium harboring pLPGM113, pLPGM202 and pLPGM413 were selected and cultivated in YEB media supplemented with 100 mg/L rifampicin and 100 mg/L Kan. Positive Agrobacterium cultivated overnight at 28uC were harvested at OD 600 of 1.5 to 2.0, centrifuged at 5000 rpm for 10 min and re-suspended in 50 ml of infiltration liquid, and the centrifugation and resuspension procedure was repeated three to five times. Finally, Agrobacterium cells were diluted in agroinfiltration liquid to appropriate concentration for agroinfiltration. Different agroinfiltration durations (24 h, 48 h, 72 h and 96 h) and Agrobacterium concentrations (OD 600 0.05, 0.10 and 0.15) were evaluated to determine conditions to obtain high transformation efficiency ( Table 1).
The complete infiltration liquid was made as following: 41.65 mM D-glucose, 100 mM CaCl 2 , 100 mM MES-KOH (pH 5.6) stock solution, 0.011 mM BAP, 0.01% Silwet L-77, 0.05 mM MgCl 2 and 12.5 mM AS (made with DMF, dimethylformamide) stock solution, and suitable amount of ddH 2 O to make final volume to 20 ml. About 200 ml infiltration liquid with Agrobacterium carrying constructed vectors were injected into the interface of adaxial epidermis and mesophyll of onion scales to make a bubble for agroinfiltration. In order to investigate whether all the components are necessary, different infiltration liquid without one of the components were used to evaluated the effect of different components of the infiltration liquid (Table 2).

Particle bombardment
The adaxial epidermis was obtained from onion bulb and placed on MS medium for 1-day incubation. The binary Table 3. Comparison of agroinfiltration and particle bombardment methods on transformation in onion epidermis. Note: three samples per unit of infiltration components were investigated for ten 2 mm 2 epidermal areas. The efficiency value represents the mean of transformation efficiencies of thirty replicates of 2 mm 2 epidermal areas from three samples (the percentage of positive cells in total cells per unit area) in the part of transformation efficiency, and the standard errors were calculated by using Excel. doi:10.1371/journal.pone.0083556.t003 expression vector pLPGM413 was transformed into the onion epidermal cells with particle bombardment method as described by [51].

DAPI staining
To visualize nuclei, the epidermis was stained with DAPI (5 mg/ mL, sigma, USA). Materials were soaked in the dye liquid phosphate buffer solution (PBS) (pH 7.0; DAPI: PBS (v/ v) = 1:1000) and kept in darkness for 20 min. Pieces of onion epidermis were arranged on slides to make wet mounts, the made slides were observed and photographed in dark-field of fluorescence microscope (Olympus BX 61, Japan).

Microscopic investigation
Transformation efficiency was determined by calculating the proportion of positive epidermal cells with fluorescent signals among the cells in 2 mm 2 epidermis area that was measured with micro ruler under microscope. Images of epidermal cells were taken by using a motorized fluorescence microscope with a mirror unit (U-MNU2), dichroic mirror (DM400), excitation filter (BP360) and barrier filter (BA420) for DAPI (DNA staining). Images of epidermal cells positive for GFP were taken with a mirror unit (U-MSWB2), dichroic mirror (DM500), excitation filter (BP470-490) and barrier filter (BA520IP) for GFP. Images of epidermal cells positive for RFP were taken with a mirror unit (U-MSWG2), dichroic mirror (DM570), excitation filter (BP530-550) and barrier filter (BA590).

Statistical analysis
All statistical analyses were performed with SPSS (version 10.0).  Note: three repeat samples for each treatment were investigated by observing ten 2 mm 2 epidermal areas of each sample. Each efficiency value represents the mean of transformation efficiencies of thirty replicates of 2 mm 2 epidermal areas from three samples (the percentage of positive cells in total cells per unit area), and the standard errors were calculated by using Excel. Different capital and lowercase letters within the same column exhibit significant difference at the 1% and 5% probability level according to the Duncan test of SPSS 10.0 statistic analysis. doi:10.1371/journal.pone.0083556.t005