To facilitate the use of multi-color BiFC in plants, we adapted a previously-described series of pSAT vectors 8. The pSAT vectors 11 are built in modular fashion, with rare-cutting restriction endonuclease or homing endonuclease sites surrounding an "expression cassette". Each "set" of pSAT vectors is flanked by different rare-cutting sites, and includes a double Cauliflower Mosaic Virus (CaMV) promoter, a Tobacco Etch Virus (TEV) translational leader, a multi-cloning site either preceding or following an autofluorescent protein N- or C-terminal fragment, and a CaMV polyA addition signal. Shyu et al. 12 showed that pairing the C-terminal fragment of cyan fluorescent protein (cCFP) with either the N-terminal fragment of Cerulean (nCerulean) or the N-terminal fragment of Venus (nVenus) results in more intense blue or yellow fluorescence, respectively, than using other GFP-derived autofluorescent protein fragments. "Overlapping" the N- and C-terminal autofluorescent protein fragments additionally increased signal intensity. We therefore used nVenus or nCerulean fragments from amino acids 1–173, and cCFP from amino acids 155–238. Figures 1A–C show the final constructions. Table 1 lists the currently available multi-color BiFC vectors that we have constructed.

One of the versatile features of the pSAT series of vectors is that, by using expression cassettes flanked by different rare-cutting enzyme sites, multiple cassettes can be "loaded" into a common replicating plasmid or a T-DNA binary vector for simultaneous introduction into plant cells. Thus, we constructed each new expression cassette in the pSAT1 or pSAT1A vector series (flanked by AscI sites), pSAT4 or pSAT4A vector series (flanked by I-SceI sites), and pSAT6 (flanked by PI-PspI sites). As "recipient" vectors for these expression cassettes, we introduced a multiple rare-cutting site (RCS) sequence into pBluescript KS⁺ (pBS-RCS), pUC119 (pUC-RCS), and an altered version of the T-DNA binary vector pPZP-RCS2 13. pPZP-RCS2 was modified by placing a P_ocs-bar-Term_ocs selection marker cassette into the EcoRI site of the binary vector, near the T-DNA left border, generating ocs-bar-RCS2-2 (pE3519). Figures 2A and 2B show maps of the pUC119 and T-DNA binary vectors, respectively.

As proof of concept, we investigated the interaction of Agrobacterium tumefaciens VirE2 protein with two members of the Arabidopsis nuclear import apparatus, AtImpa-1 and AtImpa-4. VirE2 is a single-stranded DNA binding protein that is exported from A. tumefaciens to the plant cell, where it presumably binds to the single-stranded T-DNA (the T-strand). Binding serves to protect the T-strand from nucleolytic degradation within the plant 1415 and may help direct the T-strand to the nucleus 1617. Our laboratory has recently shown that VirE2 can individually interact with the Arabidopsis importin α proteins AtImpa-1 and AtImpa-4 in yeast, in tobacco BY-2 cells, and in vitro. Individually, VirE2 interacts with AtImpa-1 in the cytoplasm, whereas VirE2 interacts with AtImpa-4 in the nucleus 18. We therefore conducted multi-color BiFC to determine whether, using this technique, the sub-cellular site of VirE2 interaction with these two importin α isoforms yielded the same results as we had previously determined using individual AtImpa isoforms.

We tagged VirE2 on its C-terminus with cCFP (VirE2-cCFP) in the pSAT1-derived plasmid. Similarly, we tagged AtImpa-1 and AtImpa-4 at their C-termini. Proteins tagged with nVenus were in pSAT4-derived vectors, whereas proteins tagged with nCerulean were in pSAT6-derived vectors. Additionally, we constructed a full-length mCherry expression vector in pRTL2 19, and an expression vector containing mCherry-VirD2NLS in pSAT6. Both of these mCherry genes were placed under the control of a double CaMV 35S promoter. mCherry localizes to both the nucleus and to the cytoplasm of cells, whereas mCherry-VirD2NLS localizes predominantly to the nucleus 8. Various combinations of expression cassettes, encoding VirE2, AtImpa-1, and AtImpa-4, were inserted into the corresponding sites of pBluescript-RCS and introduced into tobacco BY-2 suspension cells by electroporation or direct DNA uptake, and into onion cells by particle bombardment. In addition, we co-transfected either a mCherry or a mCherry-VirD2NLS expression cassette to indicate which cells were transfected, and to distinguish various sub-cellular compartments. Experiments in which the VirE2 and AtImpa constructions were co-transfected in the absence of the mCherry markers indicated that expression of mCherry in cells did not alter the sub-cellular sites of localization of the BiFC interaction (data not shown).

Figure 3 shows results of the tobacco BY-2 transfection assays, as visualized both by epifluorescence and confocal microscopy. AtImpa-1-nVenus interacted with VirE2-cCFP in the cytoplasm, whereas AtImpa-4 simultaneously interacted with VirE2-nCerulean. This latter interaction occurred predominantly in the nucleus, but also weakly in the cytoplasm (see Figure 3A, fourth panel and Figures 3B and 3C, third panels). Nuclear localization of AtImpa-4 was confirmed by co-localization of the nuclear marker mCherry-VirD2NLS.

To assure that the sub-cellular sites of interaction truly reflected properties of the test proteins and not those of the autofluorescent protein tags, we "switched" the tags: AtImpa-1 was now tagged with nCerulean, and AtImpa-4 with nVenus. AtImpa-1-nCerulean continued to interact with VirE2-cCFP in the cytoplasm, whereas AtImpa-4-nVenus interacted with VirE2-cCFP predominantly in the nucleus. Thus, the sub-cellular site of interaction was a function of the test proteins and not of the autofluorescent tag.

Although A. tumefaciens can infect a wide variety of plants, including monocots 20, dicotyledonous plants such as tobacco are more "natural" hosts (i.e., plants which can develop Crown Gall disease). We therefore were interested in determining whether Agrobacterium VirE2 protein could interact with Arabidopsis importin α proteins in onion, which is not a natural Agrobacterium host. We introduced the various combinations of tagged protein-coding genes into onion epidermal cells by particle bombardment and visualized the sub-cellular sites of interaction using laser scanning confocal microscopy. Figure 4 shows that, as in tobacco BY-2 cells, VirE2 interacted with AtImpa-1 in the onion cytoplasm, and with AtImpa-4 predominantly in the onion nucleus. These experiments indicate that the sub-cellular sites of interaction of these proteins are not altered by the particular host, nor by the method of transgene delivery into these hosts.

To assure that, using these vectors, interaction of a bait protein in the presence of multiple prey proteins does not differ from interaction of a bait protein in the presence of a single prey protein, we co-expressed VirE2-cCFP with either AtImpa-1 or AtImpa4. Figures 5A and 5B show that, as described above, VirE2 interacts with AtImpa-1 in the cytoplasm and with AtImpa-4 predominantly in the nucleus of bombarded onion cells. In addition, we tested whether fluorescence complementation could occur in the absence of interacting bait and prey proteins. Figure 5C shows that when VirE2-cCFP was co-expressed with nVenus (empty-nVenus), no yellow fluorescence was observed. To assure that the onion cells had received the various BiFC constructs, we included in this latter control experiment a mCherry-VirD2NLS expression construct cloned into the same vector as the BiFC expression cassettes. Red fluorescence was detected and localized to the nucleus, indicating that the onion cell had received the various BiFC constructs. Thus, we did not detect reconstituted yellow fluorescence in the absence of appropriately tagged bait and prey proteins.

PCR amplified nVenus, nCerulean, and cCFP fragments were initially cloned into pBluescript II KS⁺. After sequences were confirmed, all three fragments were released from the vector by digestion with the appropriate restriction endonucleases and cloned into pSATN(A) vectors. Primers used for nVenus and nCerulean were: FP-1C (NcoI): TTA A

CC ATG G

AGA TCT

GG ATC C

TCT AGA

CC ATG G

AGA TCT

GG ATC C

TCT AGA

For electroporation experiments, protoplasts were isolated from five-day old Nicotiana tabacum BY-2 suspension cells. Suspension cultures were grown at 23°C with shaking (130 rpm) in a medium containing Murashige and Skoog salts 21 supplemented with 1 mg/L thiamine-HCl, 370 mg/L KH₂PO₄, 30 g/L sucrose, and 2 mg/L2, 4-dichlorophenoxyacetic acid, pH 5.7. Cells were sub-cultured once per week by adding 2.5–3 ml of inoculum to 50 ml of fresh medium in 250 ml Erlenmeyer flasks. The 50 ml of suspension culture was centrifuged at 250 × g at room temperature for 5 min in a Sorvall GLC-2 centrifuge, and the pellet (15 ml packed cell volume) was re-suspended in 50 ml of protoplast isolation solution containing 7.4 g/L CaCl₂ ·2H₂O, 1 g/L NaOAc, and 45 g/L mannitol supplemented with 1.2% cellulose R10 (Onazuka) and 0.6% Macerozyme (Duchefa), pH 5.7. Approximately 15 ml of suspension culture was transferred into three 20 × 100 mm sterile Petri dishes and incubated in the dark with a gentle shaking (40 rpm) at room temperature for 4 hours. The protoplasts were washed twice with protoplast isolation solution and the pellet was re-suspended in 50 ml of floating solution (99 mg/L myo-inositol, 2.88 g/L L-proline, 100 mg/L enzymatic casein hydrolysate, 102.6 g/L sucrose, 97.6 mg/L MES buffer, 4.3 g/L MS salts, 1 mg/L thiamine-HCl, 370 mg/L KH₂PO₄, pH 5.7). Protoplasts (floating on the top of the solution) were transferred to a new tube, washed twice, and suspended in 50 ml of electroporation solution (10 mM NaCl, 4 mM CaCl₂·2H₂O, 120 mM KCl, 10 mM HEPES, 0.6 M mannitol, pH 7.2). Cells were incubated at 42°C for 5 min and kept in ice for 10 min before electroporation. Aliquots of protoplasts containing approximately 3 × 10⁶ cells/ml were used for electroporation. ~20 μg of each plasmid DNA were added to 300 μl of protoplasts in a tube and placed on ice. The electroporation was conducted using a BioRad Gene Pulser apparatus at 0.16 kV, with the Pulse Controller set to infinity and the capacitance extender set to 960 μFD. After 10 min incubation on ice, the protoplasts were transferred into 10 ml of BY-2 culture medium supplemented with 0.4 M mannitol and incubated overnight.

For direct DNA uptake experiments, 20 ml BY-2 suspension cells were transferred to a sterile conical centrifuge tube and centrifuged at 3000 rpm for 10 min. The pelleted cells were suspended in 10–20 ml protoplast digestion enzyme solution (1.2% Cellulase Onozuka RS [Duchefa] and 0.6% Macerozyme R-10 [Duchefa] in 10 mM CaCl₂·2H₂O, 12 mM NaOAc, 11% mannitol, pH 5.7) and incubated in the dark with shaking (40 rpm) for 3–4 hr at room temperature. The protoplasts were filtered through 40 μm nylon mesh and centrifuged in a 50 ml conical tube at 250 × g for 5 min. The protoplasts were collected and suspended in 10 ml protoplast floating solution (per liter: 99 mg myo-inositol, 2.88 g L-proline, 100 mg enzymatic casein hydrolysate, 102.6 g sucrose, 97.6 mg MES buffer, 4.3 g MS salts, 1 mg Vitamin B1, 370 mg KH₂PO₄, pH 5.7) and centrifuged at 250 × g for 10 min. Protoplasts floating in this solution were removed and 10 ml W5 solution (154 mM NaCl, 125 mM CaCl₂, 5 mM KCl, 2 mM MES, pH 5.7) was added. The solution was centrifuged at 250 × g for 5 min, and the protoplasts pelleted. Protoplast concentration was adjusted to 1 × 10⁶/ml and the solution incubated on ice for 30 min. The protoplasts were again centrifuged at 250 × g for 5 min and suspended at a density of 1 × 10⁶ cells/ml in MMg solution (0.6 M mannitol, 15 mM MgCl₂, 4 mM MES, pH 5.7). DNA (10 μg in 10 μl) was added to 100 μl protoplasts, followed by addition of 110 μl PEG solution (per ml: 0.4 g PEG 4000 [Fluca], 0.6 ml 1 M mannitol, 100 μl 1 M CaCl₂), and the protoplasts incubated at room temperature for 30 min. After addition of 1 ml W5 solution, the protoplast suspension was centrifuged at 250 × g for 5 min. The protoplast pellet was suspended in 1 ml incubation solution (per liter: MS salts and vitamins, 2 mg/L 2,4-D, 3% sucrose, 0.4 ~ 0.6 M mannitol) and incubated in the dark for 16 hr at 26°C.

Particle bombardment of onion epidermal peel layers was carried out using a Biolistic Particle Delivery System (Bio-Rad) PDS-1000. Samples (whole onion from which the dry outer layer was removed) were sterilized in ~300 ml 2% NaOCl and 2–3 drops Tween-20 for 15 min. The tissue was washed with sterilize H₂O at least five times. The upper epidermal layer of the onion was removed, cut into 2 × 2 cm squares, and placed on a plate containing 1/2 MS medium. 5 μg of each plasmid DNA was used in all experiments. Gold particles size was 1.6 μm (INBIO GOLD). 0.15–0.2 mg particles/per shot were used with a chamber vacuum of 27 in Hg. Particles were accelerated with a pressure of 1100 psi. The distance between the projectile source and the samples was 6 cm.

Transfected cells were imaged using a Nikon Eclipse E600 fluorescence microscope, or a Zeiss LSM510 Meta confocal microscope. For confocal microscopy, the objective lens was a C-Apochromat 63×/1.2 W corr. Channel specifications were as follows:

Multi Track: Channel 1 (Venus track): Argon laser; Excitation: Line active 488 nm; Transmission 8%; Main Beam Splitter 1: 488/543/633; Beam Splitter 2: 545; BP 500–530IR; Detector Gain: 620; Amplifier Offset: -0.1.

Multi Track: Channel 2 (mCherry track): HeNe laser; Excitation: Line active 543 nm; Transmission 100%; Main Beam Splitter 1: 488/543/633; Beam Splitter 2: 545; BP 565–615IR; Detector Gain: 550; Amplifier Offset: -0.1.

Multi Track: Channel 3 (Cerulean track): Argon laser; Excitation: Line active 458 nm; Transmission 20%; Main Beam Splitter 1: 458; Beam Splitter 2: mirror; BP 480–520IR; Detector Gain: 550; Amplifier Offset: -0.1.

DIC channel: HeNe laser 543 nm; Detector Gain: 198; Amplifier Offset: -0.1