Agrobacterium tumefaciens-mediated transformation of maize endosperm as a tool to study endosperm cell biology.

Developing maize (Zea mays) endosperms can be excised from the maternal tissues and undergo tissue/cell-type differentiation under in vitro conditions. We have developed a method to transform in vitro-grown endosperms using Agrobacterium tumefaciens and standard binary vectors. We show that both aleurone and starchy endosperm cells can be successfully transformed using a short cocultivation with A. tumefaciens cells. The highest transformation rates were obtained with the A. tumefaciens EHA101 strain and the pTF101.1 binary vector. The percentage of aleurone cells transformed following this method varied between 10% and 22% whereas up to the eighth layer of starchy endosperm cells underneath the aleurone layer showed transformed cells. Cultured endosperms undergo normal cell type (aleurone and starchy endosperm) differentiation and storage protein accumulation, making them suitable for cell biology and biochemical studies. In addition, transgenic cultured endosperms are able to express and accumulate epitope-tagged storage proteins that can be isolated for biochemical assays or used for immunolabeling techniques.

The endosperm is a unique plant tissue that arises from a second fertilization event between a male gamete and the central cell. Its main function is to provide nutrients to the embryo either during seed development or during germination. In cereals, the endosperm consists of three main cell types: the starchy endosperm cells, which constitute the bulk of the endosperm and accumulate large quantities of storage proteins and starch; the epidermal aleurone cells; and the transfer cells, which are in contact with the maternal vascular tissues (Olsen, 2004). The cereal endosperm is important as a model system to study plant development, cell differentiation, programmed cell death, and synthesis, trafficking, and accumulation of storage compounds. In addition, it is a major source of carbohydrate and proteins for human and animal nutrition.
In spite of its importance, cell biology studies on the cereal endosperm using modern imaging approaches such as expression of fluorescent subcellular markers are very scarce because: (1) the endosperm is deeply immersed in maternal tissues and therefore, not readily available for imaging analysis and (2) the long time required for transformation and regeneration of stable transgenic plants. Although several approaches for culturing maize (Zea mays) endosperm in vitro have been reported in the past years (Shimamoto et al., 1983), only recently a novel method developed by Odd-Arne Olsen and colleagues (Gruis et al., 2006) has proven to be successful in retaining endosperm tissue and cell type identity in in vitro conditions. Cultures derived from transgenic maize lines in which endosperm cell types are identified by the activity of specific promoters have shown that aleurone and starchy endosperm cell identity continues to be established in vitro (Gruis et al., 2006).
Although Agrobacterium tumefaciens is not a natural pathogen of most monocots (Cleene, 1985;Binns and Thomashow, 1988), it has been successfully used to transform many cereals, including maize, wheat (Triticum aestivum), Sorghum, barley (Hordeum vulgare), and rice (Oryza sativa; Grimsley et al., 1989;Gould et al., 1991;Chan et al., 1993;Ishida et al., 1996Ishida et al., , 2007Gurel et al., 2009;Harwood et al., 2009;Hensel et al., 2009). In the case of maize, stable transgenic plants can be obtained by A. tumefaciens-mediated transformation using either super-binary or standard-binary vectors (Frame et al., 2002;Mohanty et al., 2009aMohanty et al., , 2009b. However, transformation of isolated maize endosperms have been only possible using transient transformation approaches such as biolistic methods (Torrent et al., 1997;Gruis et al., 2006) and protoplast transfection (Gallie and Young, 1994). Unfortunately, these two methods are not always ideal for cell biology studies. On one hand, biolistic methods often result in high-copy number transgenic events and on the other, protoplasts are usually highly stressed cells not suitable for detailed protein localization studies. A. tumefaciens-mediated transformation methods circumvent these disadvantages by resulting in a low-copy number of transgenes in intact tissues.
We have developed a method to transform in vitrogrown endosperms using a brief incubation time with A. tumefaciens cells carrying standard binary vectors. We present here a detailed explanation of the method and quantitative information on the transformation efficiency using different A. tumefaciens strains, culture density, and incubation time. We also provide evidence that the in vitro-differentiated aleurone and starchy endosperm cells are comparable to the corresponding cell types differentiated in planta and therefore, suitable for cell biology studies. In addition, we show that transgenic cultured endosperms are able to express and accumulate epitope-tagged storage proteins that can be isolated for biochemical assays or used for immunolabeling imaging techniques.

RESULTS AND DISCUSSION
In Vitro-Cultured Endosperms Show Normal Cell Structural Features and Storage Protein Accumulation Patterns Endosperms from maize inbred lines with good culture growth properties, such as A636, are able to proliferate very well under in vitro conditions. Moreover, endosperms grown according to the endosperm in vitro culture system (EICS) developed by Gruis et al. Figure 1. Structural features of aleurone and starchy endosperm cells. A to D, Overviews and cross sections of cultured endosperms at different stages of development. Developing endosperms were excised at 6 DAP and kept in culture for 1 (A), 4 (B), 6 (C), and 8 (D) d. Note the differentiating aleurone (Al) and starchy endosperm (St E) cells in 6 + 6 (6 DAP + 6 DIC) and 6 + 8 endosperms. E, Longitudinal section of an A636 maize kernel at 23 DAP and the corresponding endosperm cross section showing the aleurone and starchy endosperm cells. E, Embryo; End, endosperm. F, Accumulation of endogenous 22-kD a, 15-kD b, and 27-kD g zeins in developing kernels (K) at 6,9,12,16,23,and 30 DAP and in cultured endosperms (E) at the corresponding developmental stages (6 + 0, 6 + 3, 6 + 6, 6 + 10, 6 + 17, 6 + 24). G to L, Ultrastructural features of aleurone and starchy endosperm cells from in vitro-and in planta-grown endosperms. Protein storage vacuoles (PSV) and lipid bodies (LB) developed in aleurone cells of both systems (G and J). Based on the results from immunolabeling experiments with specific antibodies, the deposition patterns of the 22-kD a (H and K) and 27-kD g zeins (I and L) in protein bodies of starchy endosperm cells are also identical in both systems. Bars = 0.5 mm in A, B, C, D, and E; 50 mm in A', B', C', D', and E'; 1 mm in G and J; and 200 nm in H, I, K, and L.
(2006) are able not only to proliferate but also to establish aleurone and starchy endosperm cell identity. This has been shown by monitoring the activation of starchy endosperm-and aleurone-specific promoters. Endosperms are excised from the kernels 6 d after pollination (DAP) and kept on agar medium containing Murashige and Skoog basal medium, vitamins, amino acids, cytokinin, and 15% Suc. In vitrogrown endosperms do not reach the same final size as endosperms developed in planta (Fig. 1); in addition, tissue differentiation is accelerated in in vitro-compared to in planta-grown endosperms. We observed that the epidermal layer started to acquire structural features of aleurone cells around 6 d in culture (DIC), that is 6 DAP plus 6 DIC (or 6 + 6 endosperms; Fig In addition, we also compared the steady-state levels of the 22-kD a, 15-kD b, and 27-kD g zeins between in vitro-and in planta-grown endosperms (Fig. 1F). Developing kernels at 6, 9, 12, 16, 23, and 30 DAP and cultured endosperms at the equivalent time points (6 DAP + 0 DIC or 6 + 0, 6 + 3, 6 + 6, 6 + 10, 6 + 17, 6 + 24) were collected and assayed for storage protein expression. Western-blot analysis of the 27-kD g zein showed that the earliest detectable expression starts at 9 DAP (6 + 3 stage for cultured endosperms) with steadily increasing expression through 30 DAP (6 + 24 stage for cultured endosperms) both in cultured and in planta-grown endosperms. Twenty-seven kildalton g zein accumulation appears higher in in vitrogrown endosperm compared to developing kernels at the equivalent developmental stage, which is consistent with the zein transcript profiles reported previously by Gruis et al. (2006) and our own observations that cultured endosperm show accelerated differentiation (Fig. 1,D and E [D' and E']). Comparable expression patterns were also observed for other endosperm storage proteins, such as 22-kD a and 15-kD b zeins (Fig. 1F). Thus, the endogenous storage proteins follow similar expression patterns in cultured and in planta-developed endosperms.
At the ultrastructural level, aleurone and starchy endosperm cells from in vitro-and in planta-developed endosperms were similar. Protein storage vacuoles and lipid bodies formed in aleurone cells (Fig. 1, G and J) and protein bodies and starch granules filled the starchy endosperm cells. Using specific antibodies (Woo et al., 2001;Holding et al., 2007), we also checked the deposition patterns of the 22-kD a and the 27-kD g zeins in protein bodies of starchy endosperm cells from 6 + 8 endosperms and 23-DAP kernels and found that they were identical, with 27-kD g zein located in the outer zone and 22-kD a zein occupying the central core of the protein bodies ( Fig. 1, H, I, K, and L).
Previous analysis on cell-type-specific promoter activity and transcript profiles (Gruis et al., 2006) together with our structural and immunolabeling analysis indicate that EICS is a suitable system for studying cell biology aspects of endosperm development and differentiation in maize.

A. tumefaciens-Mediated Endosperm Transformation Using Standard Binary Vectors
To analyze membrane dynamics or any other trafficking process it is often necessary to introduce transgenes, such as subcellular fluorescent markers, into the cell/tissues under study. We developed a protocol to transform in vitro-grown maize endosperms using A. tumefaciens-mediated transformation. This method allows for the incorporation of a low-copy number of transgenes in intact endosperm tissues. We used the A. tumefaciens strain EHA101 (Hood et al., 1986) and a pTF101.1 binary vector (Frame et al., 2002) containing GFP coding sequence fused to the endoplasmic reticulum (ER) retention signal KDEL under the control of the rice actin 1 promoter, OsAct1 prom (McElroy et al., 1991). This promoter has been shown to be active in both aleurone and starchy endosperm cells of cereals (Cho et al., 2002). For transformation, 6 DAP excised developing endosperms were cocultivated for 3 min with different culture densities of A. tumefaciens and placed on solid EICS culture medium supplemented with 500 mg/mL carbenicillin.

EICS Transformation Efficiency Is Highly Dependent on A. tumefaciens Culture Density
Transformation efficiency was evaluated using two parameters: (1) the percentage of transformed epidermal/aleurone cells and (2) the number of cell layers underneath the epidermis/aleurone layer containing transformed cells (Fig. 2). Since autofluorescence is a common phenomenon in the maize endosperm, only those cells that showed a fluorescence ER pattern and emission spectra corresponding to GFP (measured with the meta detection system of the Zeiss 510 LSM) were scored as transformed cells (Fig. 2). The number of transformed cells was analyzed at 4, 6, and 8 d after cocultivation (Fig. 3A).
We found that the transformation efficiency in the epidermal/aleurone layer was directly correlated to the density of A. tumefaciens cultures used to transform the endosperms. With the lowest A. tumefaciens concentration we tested (optical density at 600 nm [OD 600 ] = 0.2) the percentage of transformed epidermal/aleurone cells varied between 5% and 15%, whereas with the highest concentration (OD 600 = 0.8), the transformation efficiency in the epidermal/aleurone layer varied between 10% and 22% (Fig. 3A). No major changes were found in the percentage of transformed cells at the different time points checked in this study (Fig. 3A).
We also analyzed the number of transformed starchy endosperm cells in endosperm cross sections. Transformed cells were detected from the first layer (the starchy endosperm cells right beneath the aleurone layer) up to the eighth layer of starchy endosperm cells (Table I).
Active periclinal (parallel to the surface of the epidermal layer) divisions have been reported to occur in in vitro-grown endosperms (Gruis et al., 2006). Therefore, at least some transformed starchy endosperm cells found in deeper areas of the endosperms are likely to be derived from transformed epidermal/ aleurone cells.

Transformation Efficiency Is Higher with the A. tumefaciens Strain EHA101
To evaluate the ability of a different A. tumefaciens strain to transform maize endosperms, the same binary vector containing the GFP reporter was introduced into A. tumefaciens GV3101 cells. We used both A. tumefaciens EHA101 and GV3101 strains for side-byside endosperm transformation at a concentration of 0.8 OD 600 and during 3-or 10-min cocultivation time (Fig. 3B). We analyzed the transformation efficiency in the aleurone layer 8 d after cocultivation. We found that the transformation efficiency was much higher when the EHA101 strain was used. In fact, no transformed cells were detected after the 3-min incubation treatment with the GV3101 strain (Fig. 3B).  0.2 3 6 0 2 6 1 3 6 1 0.4 3 6 1 3 6 0 3 6 0 0.6 3 6 1 4 6 1 4 6 1 0.8 5 6 1 4 6 2 5 6 3

Transgenic in Vitro-Cultured Endosperms Express and Accumulate Tagged Storage Proteins
To assess the accumulation levels of proteins encoded by transgenes transformed into in vitro-grown endosperms, we expressed the 22-kD a, 15-kD b, and the 27-kD g zeins fused to the HSV, FLAG, and HA tags, respectively. For driving high expression in the starchy endosperm, we used the Floury2 promoter (FL2 prom ; World Intellectual Property Organization patent publication no. 9802563) for the expression of 22-kD a zein-HSV and the maize 27-kD g zein promoter (27kDZein prom ; Ueda and Messing, 1991;Russell and Fromm, 1997) for the expression of 15-kD b zein-FLAG and 27-kD g zein-HA. The CZ19B1 prom :DsRed transgene was built in the same epitope-tagged zein con-  structs to monitor endosperm transformation since 19-kD b zein1 promoter (CZ19B1 prom ; U.S. patent 6225529) shows specific expression in the starchy endosperm. Sectors with strong red fluorescence were observed in the transformed cultured endosperms (Fig. 4, A-C [A'-C']) whereas no fluorescence was detected in control cultures (Agrobacterium incubation omitted), confirming successful endosperm transformation.
Fifteen days after transformation, endosperms showing red fluorescence were collected and ground for protein extraction. The protein extracts were subjected to SDS-PAGE followed by immunoblot analysis using antibodies against the epitope tags as well antibodies against the corresponding endogenous storage zein proteins (Woo et al., 2001; Fig. 4D). Immunoblot detection with antibodies against the HSV, FLAG, and HA tags confirmed the expression and accumulation of 22-kD a zein-HSV, 15-kD b zein-FLAG, and 27-kD g zein-HA in transgenic cultured endosperms, indicating that in vitro-grown endosperms can be used for expression and isolation of transgenic endosperm storage proteins or for immunolabeling imaging. CONCLUSION We have developed a protocol to transform endosperm tissue grown in vitro. Since in vitro-cultured endosperms undergo normal differentiation of aleurone and starchy endosperm cells, the possibility to introduce transgenes such as fluorescent subcellular markers, allows for easy imaging access to different endosperm cell types that are usually deeply immersed within other tissues.
The ability to introduce transgenes into in vitrocultured developing endosperms provides biochemical and molecular means to study cereal endosperm cell fate differentiation and early endosperm development. The possibility of imaging fluorescent subcellular markers in both aleurone and starchy endosperm cells following a very simple and short protocol (Fig. 5) represents an important technical advance in cell biology studies of differentiating cereal endosperm cells. In addition, this system offers the possibility of expressing and isolating tagged/modified endosperm storage proteins that cannot be successfully expressed in other tissues/systems.

Isolation and in Vitro Growth of Endosperms
Maize (Zea mays; inbred line A636) were grown in a greenhouse under a 14-h light/10-h dark photoperiod, supplemental lighting (700 mmol m 22 s 21 ), and average temperature of 28°C during the day and 21°C during the night. Endosperms were isolated and cultured in vitro as described by Gruis et al. (2006). Briefly 6 DAP ears were harvested and surface sterilized. Using a scalpel the kernels were dissected, the maternal tissue removed, and the fertilized embryo sacs isolated to obtain clean endosperms. Developing endosperms were immediately placed on liquid EICS culture medium until cocultivation with Agrobacterium tumefaciens. After the cocultivation, the isolated endosperms were placed on solid EICS culture medium (4.3 g/L Murashige and Skoog media; 0.5% v/v Murashige Skoog vitamins stock solution; 5 mg/L thiamine HCl; 400 mg/L Asn; 10 mg/L 6-benzylaminopurine; 15% Suc; and 3 g/L Gelrite, pH to 5.8) supplemented with 500 mg/mL carbenicillin and kept on dark at 25°C. Isolated endosperm not exposed to A. tumefaciens cultures were used as a control.

Plasmids
A DNA fragment containing OsActin1 prom :GFP-KDEL was cloned into the pTF101.1 vector using a two-step cloning strategy. The GFP-KDEL sequence was amplified from the CD3-955 vector (Nelson et al., 2007) using the forward primer 5#-TCTAGAATGAAGGTACAGGAGGGT-3# and the reverse primer 5#-CCCGGGTTACAGCTCGTCATG-3#, containing XbaI and XmaI restriction sites (underlined), respectively, and cloned into pTF101.1. The OsActin1 promoter was amplified from the pDM302 vector using the following primers: forward 5#-AAGCTTGAAGAGAGTCGGGATAGTC-3# containing a HindIII restriction site and reverse 5#-TCTAGACAGAAATATATAAAAATATAAAC-CAT-3# containing an XbaI restriction site. The vector pDM302 carrying the OsActin1 promoter was kindly donated by Ajay Garg (Cornell University) and the GFP coding region fused to the ER retention signal KDEL (CD3-955; Nelson et al., 2007) was obtained from the Arabidopsis Biological Resource Center at Ohio State University. The resulting plasmid was introduced into competent A. tumefaciens cells by freeze-thaw transformation (Chen et al., 1994).
Additional binary vectors used in this study contained CZ19B1 prom (U.S. patent 6225529) fused to DsRed (CLONTECH), FL2 prom :22kD a zein fused to the HSV tag (QPELAPEDPED), 27kDZein prom :15kD b zein gene fused to FLAG (DYKDDDDK), and the 27kDZein prom :27kD g zein gene fused to HA (YPYDVP-DYA). Sequence data for the zein genes used in this article can be found in the GenBank/EMBL data libraries under accession numbers AF371261, AF371264, and AF371274.

A. tumefaciens-Mediated Transformation
A. tumefaciens EHA101 or GV3101 strains carrying the different constructs were grown at 25°C for 2 d in Luria-Bertani medium supplemented with the appropriate antibiotics for strain (100 mg/mL kanamycin for EHA101 or 100 mg/mL gentamicin plus 10 mg/mL rifampicin for GV3101) and plasmid selection (100 mg/mL spectinomycin for pTF101.1). After 2 d of growth, the cultures were centrifuged and washed twice with the infiltration media (EICS culture media supplemented with 100 mM acetosyringone). Finally the bacterial suspension was diluted with infiltration media to adjust the inoculum concentration to the final OD 600 value.
Transformation was performed by cocultivating the isolated endosperms with the bacterial suspensions under gentle agitation. After cocultivation with A. tumefaciens cultures, the endosperms were washed three times with EICS culture medium supplemented with carbenicillin, plated, and kept in the dark.

Confocal Imaging of Fluorescent Proteins
Cross or paradermal sections of the transformed endosperms were imaged using a 510 Zeiss laser-scanning confocal microscope. The transformed tissues were excited with 488 nm and the GFP emission was detected using a 500 to 530 band-pass filter. The GFP emission spectrum was collected for every image using the spectral meta detector. Only those cells that showed a positive signal for GFP were scored as transformed cells.
The percentage of transformed cells was calculated by determining the number of transformed cells over the total of cells per field. The number of total cells per field ranged from 45 to 160 in the different images obtained. At least three fields were analyzed for each section and a total of three independent in vitro-grown endosperms were analyzed.
For determining the number of starchy endosperm layers (layers underneath the epidermal/aleurone layer) containing transformed cells, we used transverse sections of the endosperms.
The images were analyzed using the LSM image browser (www.zeiss. com/lsm) and edited using Adobe Photoshop CS4.

Structural Characterization and Gold Immunolabeling
In vitro-grown endosperms at different stages of development (6 + 1, 6 + 3, 6 + 6, and 6 + 8) and thin slices of endosperm tissue from developing kernels were processed by high-pressure freezing/freeze substitution. Pieces of tissue were transferred to freezing planchettes containing 0.1 M of Suc and highpressure frozen in a Baltec HPM 010 unit (Technotrade). Substitution was performed in 2% OsO 4 in anhydrous acetone at 280°C for 72 h, and followed by slow warming to room temperature over a period of 2 d. After several acetone rinses, samples were removed from the holders and infiltrated in Epon resin (Ted Pella Inc.) according to the following schedule: 5% resin in acetone (4 h), 10% resin (12 h), 25% resin (12 h), 50%, 75%, and 100% (24 h each concentration). Polymerization was carried out at 60°C. Sections were stained with 2% uranyl acetate in 70% methanol for 10 min followed by Reynold's lead citrate (2.6% lead nitrate and 3.5% sodium citrate, pH 12) and observed in a FEI CM120 electron microscope.
For gold immunolabeling, high-pressure frozen samples were substituted in 0.2% uranyl acetate (Electron Microscopy Sciences) plus 0.2% glutaraldehyde (Electron Microscopy Sciences) in acetone at 280°C for 72 h, and warmed to 250°C for 24 h. After several acetone rinses these samples were infiltrated with Lowicryl HM20 (Electron Microscopy Sciences) for 72 h and polymerized at 250°C under UV light for 48 h. Sections were mounted on formvar-coated nickel grids and blocked for 20 min with a 10% (w/v) solution of nonfat milk in phosphate-buffered saline (PBS) containing 0.1% Tween 20. The sections were incubated with the primary antibodies (1:10 in PBS-Tween 20) for 1 h, rinsed in PBS containing 0.5% Tween 20, and then transferred to the secondary antibody (anti-rabbit IgG 1:10) conjugated to 15-nm gold particles for 1 h. Controls omitted the primary antibodies. The antibodies against 22-kD a zeins and 27-kD g zeins have been characterized elsewhere (Woo et al., 2001;Holding et al., 2007).

Immunoblot Analysis
In vitro-grown endosperms were ground with a plastic pestle in extraction buffer (50 mM Tris-HCl, 5 mM EDTA, and 2% SDS; pH 8.0) containing a protease inhibitor cocktail (Sigma-Aldrich) at a 1:2 (w/v) ratio. Cellular debris was removed by centrifugation at 140,000 rpm for 10 min. Protein extracts were diluted 1:2 (v/v) in sample buffer and boiled for 5 min before separation through a 4% to 12% (w/v) SDS polyacrylamide gel (Invitrogen) and then transferred to a cellulose membrane in a submerged blotting system (Mini-Trans Blot; Invitrogen). Membranes were blocked for 1 h with Tris-buffered saline containing 5% (w/v) nonfat dry milk. Immunoblotting was performed using the corresponding epitope tag antibodies (HA, FLAG, HSV; Sigma-Aldrich), or the corresponding storage protein antibodies (22-kD a, 15-kD b, and 27-kD g zeins; Woo et al., 2001), and the anti-rabbit/anti-mouse secondary antibodies (Bio-Rad) in Tris-buffered saline containing 0.1% (v/v) Tween 20. Proteins that cross-reacted with antibodies were detected with chemiluminescent substrates (Pierce) and visualized on film.
The anti-22-kD a zein antibody was used at a dilution of 1:3,000, the anti-15-kD b zein antibody was used at 1:3,500, and the anti-27-kD g zein antibody, at 1:3,500. Antibodies against HA, FLAG, and HSV were used at dilutions of 1:8,000, 1:8,000, and 1:4,000, respectively. Sequence data for the zein genes used in this article can be found in the GenBank/EMBL data libraries under accession numbers AF371261, AF371264, and AF371274.