Photosensitizers are common in nature and play diverse roles as defense compounds and pathogenicity determinants and as important molecules in many biological processes. Toxoflavin, a photosensitizer produced by Burkholderia glumae, has been implicated as an essential virulence factor causing bacterial rice grain rot. Toxoflavin produces superoxide and H2O2 during redox cycles under oxygen and light, and these reactive oxygen species cause phytotoxic effects. To utilize toxoflavin as a selection agent in plant transformation, we identified a gene, tflA, which encodes a toxoflavin-degrading enzyme in the Paenibacillus polymyxa JH2 strain. TflA was estimated as 24.56 kDa in size based on the amino acid sequence and is similar to a ring-cleavage extradiol dioxygenase in the Exiguobacterium sp. 255-15; however, unlike other extradiol dioxygenases, Mn2+and dithiothreitol were required for toxoflavin degradation by TflA. Here, our results suggested toxoflavin is a photosensitizer and its degradation by TflA serves as a light-dependent selection marker system in diverse plant species. We examined the efficiencies of two different plant selection systems, toxoflavin/tflA and hygromycin/hygromycin phosphotransferase (hpt) in both rice and Arabidopsis. The toxoflavin/tflA selection was more remarkable than hygromycin/hpt selection in the high-density screening of transgenic Arabidopsis seeds. Based on these results, we propose the toxoflavin/tflA selection system, which is based on the degradation of the photosensitizer, provides a new robust nonantibiotic selection marker system for diverse plants.
The generation of transgenic plants with agronomically improved traits using a recombinant DNA and plant transformation technology revolutionized the field of molecular farming and improved agricultural production and economy. Transgenic plants are often distinguished by the presence of selectable marker genes, especially when transformation efficiency is very low. Since the early 1980s, more than 50 selectable marker genes have been isolated from various sources (Hare and Chua, 2002; Erikson et al., 2004; Miki and McHugh, 2004). The three major groups of selection markers currently used for generating transgenic plants confer resistance to antibiotics, herbicides or metabolic inhibitors (Miki and McHugh, 2004). Among these, the four most widely used selection systems include: (i) kanamycin (Km) and neomycin phosphotransferase II (nptII); (ii) hygromycin B (Hyg) and hygromycin phosphotransferase (hpt); (iii) glufocinate and phosphinothricin N-acetyltransferase and (iv) glyphosate and 5-enolpyruvylshikimate-3-phosphate synthase (Miki and McHugh, 2004). Although many selection systems have been reported, nptII was most frequently used in the United States to produce transgenic crop plants (soybean, cotton, canola and corn) in 2001–2002 (Miki and McHugh, 2004). Genes of 5-enolpyruvylshikimate-3-phosphate synthase and phosphinothricin N-acetyltransferase were used one-half and one-third as frequently as nptII, respectively (Miki and McHugh, 2004). Although these antibiotic and herbicide resistance selection systems are useful in both research and commercial applications, use of toxic selection agents in the field and expression of selection marker genes in transgenic plants have created public concern (Kuiper et al., 2001). As a result, efforts to develop environmentally sound selection marker systems that use natural, nontoxic compounds are underway. One example is betain aldehyde dehydrogenase (BADH), which is the last enzyme of the glycine betain (GB) biosynthesis pathway in plants. A BADH gene from spinach was used as a nontoxic, nonantibiotic selection marker for chloroplast transformation by converting toxic betain aldehyde to nontoxic GB, which serves as an osmotic protectant (Daniell et al., 2001). BADH was shown to enhance salt tolerance in spinach and sugar beet (McCue and Hanson, 1992; Nakamura et al., 1997). Recently, dao1 (D-amino acid oxidase), with various D-amino acids (Erikson et al., 2004), and sweet pepper ferredoxin-like protein (pflp), with natural bacterial pathogen Erwinia carotovora as a selection agent, have been reported (You et al., 2003). As an alternative, selection marker–free constructs can be introduced into plants, but plant transformation without selection markers is tedious and is applicable only to reasonably efficient transformation systems (Hare and Chua, 2002). Gene stacking or pyramiding through re-transformation to introduce multiple traits into crops also requires versatile selection markers (Halpin, 2005). Hence, it is crucial to continue developing alternative selection marker systems.
Photosensitizers are commonly found as diverse forms (flavin, chlorophyll, coumarins, porphyrins and cercosporin) in nature, and chlorophyll and porphyrins function mainly as transducers of light energy in photosynthetic organisms (Spikes, 1989; Daub and Ehrenshaft, 2000). During photosynthesis, chlorophylls absorb photons and transduce light energy. Superoxide and H2O2 are produced during this process, but low levels of these reactive oxygen species can be quenched by carotenoids in plant cells. However, endogenous chlorophylls might damage plant cells under carbon dioxide deprivation, and exposure to eosin can inhibit photosynthesis (Spikes, 1989). Common dyes such as acridine orange, eosin Y, methylene blue and safranine are also photosensitizers.
Toxoflavin was identified in 1934 as one of the phytotoxins produced by Pseudomonas cocovenenans and was proposed to produce superoxide (O2−) and H2O2 during autorecycling oxidation processes under oxygen and light (Latuasan and Berends, 1961; Nagamatsu et al., 1982). We reported previously that Burkholderia glumae produces toxoflavin as a major virulence factor causing bacterial rice grain rot and demonstrated that toxoflavin biosynthesis in B. glumae is regulated by quorum sensing (Kim et al., 2004).
Given that the light-dependent generation of superoxide and H2O2 by toxoflavin results in damage to plant cells, we hypothesized that the detoxification of toxoflavin would confer positive selection in the generation of transgenic plants. In this study, we report a gene tflA encoding toxoflavin degradation enzyme and provide experimental evidence suggesting the use of a combination of toxoflavin and tflA as a new robust selection marker system in plants.
Sensitivity of various plant leaf discs to toxoflavin
We first tested the phytotoxic effect of toxoflavin on 12 different monocot and dicot plants, including rice, barley, corn, soybean, Arabidopsis, hot pepper, tomato, tobacco, watermelon, cucumber, melon and zucchini. Toxoflavin at various concentrations (0–20 mg/L) had no effect on any of the tested plants in the dark; however, all of the plant leaves tested were sensitive to toxoflavin within 24–48 h under the light (124.5 μmol) (Figure 1). Leaf discs of corn, Arabidopsis, tobacco and cucumber were very sensitive to low concentrations of toxoflavin (1–6 mg/L) (Figure 1). The light-dependent and rapid response of diverse plants to toxoflavin suggested toxoflavin is a good candidate as a positive selection agent.
Identification of tflA as a toxoflavin-degrading enzyme
Here, we tested environmental samples including soil, plant debris and rice seeds to isolate bacteria that survive in toxoflavin solution. Among the samples tested, we isolated one bacterium that survived in Luria-Bertani (LB) broth supplemented with 40 mg/L toxoflavin. The bacterium was isolated from healthy rice seeds and identified as Paenibacillus polymyxa JH2. To identify a gene responsible for toxoflavin degradation, a genomic library of P. polymyxa JH2 was constructed in Escherichia coli HB101. As the cells of E. coli HB101 are sensitive to toxoflavin, we directly screened individual colonies from the genomic library to find colonies that could survive on LB agar plates supplemented with toxoflavin. We found one colony was able to grow on LB agar plates supplemented with 40 mg/L toxoflavin. A cosmid was isolated from the colony and its restriction map was constructed (Figure S1). The cosmid was named as pJ9 and carried an approximately 21 kb insert. Deletion and subcloning analysis of pJ9 determined the 1.2- kb HindIII fragment carried a gene responsible for toxoflavin degradation (Figure S1). The DNA sequences of the 1.2- kb HindIII fragment were determined, and one possible open reading frame called tflA (GenBank accession number GQ921834) was identified (Figure S1). The gene tflA is 666 bp in size and encodes a protein, TflA, expected to be in the size 24.56 kDa based on amino acid sequence. Analyses of amino acid sequences of TflA with the 6 previously reported members of dioxygenase superfamily exhibited 36.52% identity to a predicted ring-cleavage extradiol dioxygenase of Exiguobacterium sp. 255-15 (Figure S2). Three conserved amino acids with unknown functions are indicated with asterisks (Figure S2). A homology tree revealed that TflA is subgrouped more closely with glyoxalase/bleomycin resistance protein/dioxygenase of Exiguobacterium sp. 255-15 (gi172057369) and the hypothetical protein of Bacillus halodurans C-125 (gi15614705) than others (Figure S3).
Biochemical characteristics of His-TflA
To determine the biochemical properties of TflA, we purified N-terminal His-tagged TflA (His-TflA) recombinant protein from E. coli BL21(DE3) carrying pH904 (Figure S4). Optimum pH and temperature for toxoflavin degradation by His-TflA were pH 6–6.5 and 25°C, respectively (Figure 2a,b). Specific activity and KM values of His-TflA were 0.0413 μmoles/min/mg and 69.72 μm, respectively (Figure 2c). As shown in Figure 3a, His-TflA required Mn2+ for enzymatic activity, which is a similar requirement of other metal-dependent extradiol dioxygenases. However, unlike other extradiol dioxygenases, dithiothreitol (DTT) was required for the degradation of toxoflavin by His-TflA, indicating that reducing conditions in the degradation reaction are essential for the activity (Figure 3a).
Substrate specificity of purified His-TflA was also examined by thin-layer chromatography (TLC) with various toxoflavin derivatives (Figures 3b and S5, S6, and Table 1). Toxoflavin, 3-methyltoxoflavin, 4,8-dihydrotoxoflavin and 3-methylreumycin at 100 μm were degraded completely by 5 μm His-TflA in 10 min at 25°C (Table 1). Reumycin and 3-methyl 4,8-dihydrotoxoflavin were slightly degraded, whereas fervenulin, 3-phenyltoxoflavin, 3-phenylreumycin and 5-deazaflavin were not degraded by His-TflA (Table 1 and Figure S6). Therefore, His-TflA failed to degrade the toxoflavin derivatives with a phenyl group on the 3rd carbon, suggesting a substrate specificity of His-TflA.
Table 1. Substrate specificity of purified His-TflA to toxoflavin and its derivatives
*Degradation activity of His-TflA was marked as follows. −, 0–20% degradation; +, 20–50% degradation; ++, 50–80% degradation; +++, 80–100% degradation.
Transformation frequency of the toxoflavin/tflA selection system in rice
To compare transformation frequency of the toxoflavin/tflA selection system with that of the Hyg/hpt selection system in rice transformation, we constructed 35S::tflA and 35S::hpt in the same binary vector pCAMLA, resulting in pJ904 (Figure 4a).
Scutellum-derived rice calli (Oryza sativa japonica cultivar Choocheongbyeo) were transformed with Agrobacterium tumefaciens GV301 carrying 35S::tflA and 35S::hpt in pJ904 (Figure 4a). Callus proliferation efficiency and transformation frequencies with the toxoflavin/tflA selection system were 46% ± 18% and 7.7% ± 0.7%, respectively, based on two independent experiments (Table 2). Similar results were obtained with the Hyg/hpt selection system (35.6% ± 9% and 9% ± 3.7%) (Table 2). No negative effects on rice development as a result of toxoflavin selection were observed in any of the transgenic lines. Similar results were obtained when the toxoflavin/tflA selection system was introduced into another japonica cultivar, Dongjinbyeo. Putative rice transformants (T2) were verified by Southern hybridization using a DNA fragment carrying tflA gene as probe (Figure S7). Independent T2 transgenic Dongjinbyeo expressed tflA at different levels as assessed by immunoblot analysis (Figure 4b). Leaf discs of the two independent T4 Dongjinbyeo transgenic lines, Dt19-5 and Dt40-6, showed strong resistance to toxoflavin (up to 20 mg/L) under visible light (Figure 4c). Wild-type Dongjinbyeo became chlorotic in 5 mg/L toxoflavin when exposed to light, whereas no chlorosis was observed in discs exposed to 20 mg/L toxoflavin in the dark (Figure 4c). These results confirmed that the expression of tflA in rice plants confers toxoflavin resistance, allowing for positive selection via the toxoflavin/tflA selection system under the light.
Table 2. Comparison of efficiency of rice transformations
Selectable marker genes
Independent experiment (n = 4)
No. of calli inoculated (A)
No. of resistant calli (B)*
Calli proliferation (%)†
No. of putative regenerated plants (C)
PCR-positive plants (D)‡
Transformation frequency (%)§
*Number of resistant calli from the 2nd 2N6 selection media (B).
†Calli proliferation (%); (B)/(A).
‡PCR was performed to amplify either hpt or tflA gene in putative regenerated plants.
§Transformation frequency (%); (D)/(A).
Comparison of the toxoflavin/tflA and Hyg/hpt selection systems in Arabidopsis thaliana
The transformation frequency of the Hyg/hpt and the toxoflavin/tflA selection systems were further compared in Arabidopsis transformed with A. tumefaciens GV301 carrying 35S::tflA and 35S::hpt in p1301-tflA in the same lot of T1 transgenic seeds (Figure 4a,d). Putative transformants on the medium supplemented with Hyg were differentiated from wild-type seedlings based on their long root phenotype (Figure 4d). Selection was visually more distinct using the toxoflavin/tflA selection system. Only transformed T1Arabidopsis seeds germinated, whereas nontransformed wild-type seeds did not germinate at all on the 0.5× Murashige and Skoog (MS) medium supplemented with 10 mg/L toxoflavin (Figure 4d). We screened 4682 and 2474 T1 seeds for the toxoflavin/tflA and Hyg/hpt selection systems, respectively, resulting in transformation frequencies of 0.6% ± 0.2% and 0.8%± 0.5%. Among selected T1 transformants, 50% and 43% of the putative tflA and hpt transformants, respectively, exhibited positive responses suggesting true transformants as assessed by histochemical ß-glucuronidase (GUS) staining (data not shown). Five randomly selected T1 seedlings showed strong level of GUS activity, suggesting the selection efficiency of toxoflavin was unambiguous compared to that of Hyg. Roots of transgenic Arabidopsis lines possessing tflA were initially shorter than those of the wild type but they developed fully when the plants were transferred to 0.5× MS medium without toxoflavin. We did not observe any abnormality in plant growth and development of selected transgenic plants. Taken together, our comparison data indicated that the toxoflavin/tflA selection system is a robust and efficient selection system for high-density screening of transgenic Arabidopsis.
In modern agriculture, the development of major crop plants transformed with disease, insect, or herbicide resistance genes provided new ways of overcoming biotic stresses, thereby achieving high productivity. Clear separation of transgenic plants from the wild type is critical to improving plant transformation. Hence, the development of a more efficient and safe selection marker system is always desired. While the current antibiotic and herbicide resistance selection markers are useful for generating transgenic plants, their use can be limited in some plants by poor selection efficiency and a high frequency of false positives. For example, a Km and nptII selection system is useful in Arabidopsis but has low selection efficiency for rice transformation. Therefore, the identification of a versatile selection marker that can be used for many photosynthetic organisms would be beneficial.
In this study, we based our search for a novel selection marker system on the fact that under light, toxoflavin produces superoxide and H2O2 that damage plant cells (Nagamatsu et al., 1982). Previous studies reported that photodynamic action is not limited to nucleic acids, but also occurs with lipids, amino acids and proteins, resulting in photosensitizing viruses, cells and muticellular organisms (Spikes, 1989). Among known photosensitizers, cercosporin is one of the photoactivated perylenequinones produced by plant pathogenic fungi such as Alternaria, Cercospora and Cladosporium (Spikes, 1989). Therefore, attempts were made to identify genes responsible for the detoxification of cercosporin for developing resistant cultivars against cercosporin-producing fungi. However, the detoxification of cercosporin is complicated and multi-genes may be involved because of its large and complex structure (Spikes, 1989).
Therefore, researchers have not been able to establish whether genes involved in detoxification of cercosporin can be used as a source of conferring disease resistance against cercosporin-producing fungi. Unlike cercosporin detoxification, a single gene mediates toxoflavin detoxification. Therefore, we hypothesized that a combination of toxoflavin and the tflA gene would be a good candidate as a positive selection system for plant transformation. The phytotoxicity of toxoflavin was effective on broad ranges of monocots and dicots. Because toxoflavin is a known photosensitizer, it was not surprising to observe that all of the plant species in this study were sensitive to toxoflavin when exposed to light. Based on our study results, we predict that all photosynthetic organisms are likely sensitive to various concentrations of toxoflavin.
Amino acid sequence analysis of TflA revealed it is similar to ring-cleavage extradiol dioxygenase, which opens the aromatic ring by incorporating two atoms of dioxygen in their substrates. However, a homology tree showed that TflA was separated from many known members of the dioxygenase superfamily. This implied TflA might not be a typical dioxygenase or suggests it might have a substrate specificity that requires enzymatic activities different from the other dioxygenases. Indeed, purified His-TflA required neutral pH, metal ions and DTT for its full enzyme activity, unlike other extradiol dioxygenases. It is of note that His-TflA failed to degrade toxoflavin derivatives with a phenyl group on the 3rd carbon. This suggested the status of the 3rd carbon may be critical for TflA activity, in contrast with other dioxygenases. However, the nature of toxoflavin degradation by His-TflA is not completely understood, owing to the complicated oxidation and reduction reactions that are common in photosensitizers (Spikes, 1989).
For rice transformation in general, the Hyg/hpt selection system is commonly used over Km and genecitin selection systems because of higher transformation frequency (Hiei et al., 1994). We believe our novel toxoflavin/tflA selection system is a good candidate for rice transformation when additional markers are required and further suggests two reasons our novel system could replace the current selection systems in rice. First, the toxoflavin/tflA selection system was comparable to the currently used Hyg/hpt selection system. Second, in contrast with the Hyg/hpt selection, we did not observe any negative effects on calli proliferation and regeneration of rice growing on medium infused with toxoflavin. However, escapes through the callus selection and plant regeneration stages were frequent.
We often have encountered false-positive selections from the Km/nptII selection system or the Hyg/hpt selection system in high-density screening of T1 transgenic Arabidopsis seeds. This may be because we select high-density transgenic seedlings from among all germinated seedlings 5–7 days after germination. With the toxoflavin/tflA selection system, we observed the clear separation of putative germinated transgenic plants from untransformed wild-type plants on the medium supplemented with toxoflavin compared to germinated plants grown on medium containing Hyg. Toxoflavin inhibited germination of the wild-type Arabidopsis seeds, eliminating false-positive selection of transgenic plants. Although root growth of transgenic Arabidopsis was retarded by toxoflavin, transfer to fresh culture media resulted in complete root growth recovery. Rooted transgenic Arabidopsis successfully flowered and showed no negative effects on plant growth and phenotype. Therefore, we conclude there are no negative effects associated with the toxoflavin/tflA selection system in plant development.
In conclusion, the use of naturally occurring photosensitizers such as toxoflavin as selection agents appears to give rapid and unambiguous selection results owing to their unique phytotoxic mode of action. As a nonantibiotic or nonherbicidal selection system, the toxoflavin/tflA selection system is a new and alternative all-around selection marker system that can be applicable to many plant species. When versatile selection markers are needed for gene stacking, the toxoflavin/tflA selection system is a good candidate with a unique mode of action compared to the currently available selection marker systems. In particular, the toxoflavin/tflA selection system might be useful for generating transgenic plants where high false-positive backgrounds with the current selection marker systems are problematic.
Leaf disc assays
Seeds of both monocot (rice, barley and corn) and dicot plants (soybean, tobacco, tomato, hot pepper, zucchini, cucumber, watermelon, melon and Arabidopsis) were germinated and grown in a plant growth room (23°C, 16/8h, day/night). Leaf discs of both monocot and dicot plants (25 days old) were cut and immersed in various concentrations (0–20 mg/L) of synthetic toxoflavin. Leaf discs were and incubated under visible light (124.5 μmol) or in the dark and then photographed after 44 h.
Isolation of toxoflavin-degrading bacterium
A toxoflavin-degrading bacterium that survived on LB medium containing 40 mg/L toxoflavin was isolated from healthy rice seeds during initial screening. The bacterium was further purified on LB medium supplemented with 40 mg/L toxoflavin to isolate a single colony. Physiological and molecular characteristics of the isolated strain were further examined by Biolog analysis, gas chromatography of fatty acid methyl esters and 16S rDNA sequence analysis. The carbon source utilization profiles of the isolates were compared on Biolog microplates as described by the manufacturer’s recommendation (Biolog GN MicroPlate, Hayward, CA, USA). The isolate was cultured on a trypticase soy broth (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) agar for 48 h at 28°C, and the fatty acid methyl esters were extracted using a standard method (MIDI Inc., Newark, DE, USA). The fatty acids were analysed with the Sherlock Microbial Identification System Version 2.11 (MIDI Inc., Newark, DE, USA). The 16S rDNA gene of the isolate was amplified by standard PCR with the following primer combinations (27mF: 5′AGAGTTTGATCMTGGCTCAG-3′ and 1492mR : 5′-GGYTACCTTGTTACGACTT-3′). The amplified DNA was completely sequenced and analysed using the BLAST program of the National Center for Biotechnology Information.
Construction of P. polymyxa JH2 cosmid library
Genomic DNA of P. polymyxa JH2 was prepared by standard procedures and partially digested with Sau3A. The 20–30 kb Sau3A fragments were obtained from a sucrose gradient centrifugation [10 to 40% (w/v)] as described previously, followed by cloning into pLAFR3 (Kim et al., 2003). Packaging into λ bacteriophage and transfection of E. coli HB101 was performed as described by the manufacturer (Promega, Madison, WI, USA).
Identification of tflA
The genomic library of P. polymyxa JH2 was screened by isolating the surviving colony on an LB agar plate supplemented with 40 mg/L toxoflavin. A cosmid, pJ9, was isolated from the surviving colony, and a restriction enzyme map of the cosmid was constructed. Based on the restriction map and sub-clone analysis of pJ9, the 1.2- kb HindIII fragment was subcloned into pBluescript II SK (+) vector to produce pJ90. The complete DNA sequence of the 1.2- kb HindIII fragment was determined and analysed by NCBI BLAST program, MEGALIGN software (DNASTAR, Madison, WI, USA) and GENETYX-WIN software (Genetyx Co., Tokyo, Japan).
Purification of N-terminal His-TflA
To purify His-TflA from E. coli BL21(DE3), we amplified the gene tflA by PCR as a NdeI-BamHI fragment and the product was cloned into pET14b vector (Novagen, Darmstadt, Germany), resulting in pH904. His-TflA was over-expressed and purified by the Ni-NTA Spin Column (QIAGEN, Valencia, CA, USA). The eluted protein was dialysed with 50 mm sodium phosphate (pH 6.5) buffer to remove imidazole, and the purified protein concentration was measured by the method of Bradford using bovine serum albumin as the standard (Bradford, 1976).
Enzyme assay of His-TflA
Purified His-TflA was assayed in vitro to establish optimum pH, temperature, and Mn2+and DTT requirements to degrade toxoflavin by TLC. His-TflA activity was determined at pH 4 through 8 (using 50 mm sodium citrate buffer for the pH range of 4–5.5 and 50 mm sodium phosphate buffer for the pH range of 6 to 8) and a temperature range of 10–40°C. All enzymatic assays for temperature were performed in 50 mm sodium phosphate buffer (pH 6.5). His-TflA (5 μm) in an assay buffer [50 mm sodium phosphate (pH 6.5), 10 μm MnCl2 and 5 mm DTT] and toxoflavin (100 μm) were incubated for 10 min at 25 °C. The activity of TflA was then estimated by a TLC assay as described later.
In order to determine the requirement of Mn2+and DTT, 5 μm His-TflA protein and 100 μm toxoflavin in assay buffer [50 mm sodium phosphate (pH 6.5)] were incubated for 10 min at 25°C in the presence and/or absence of 10 μm MnCl2 and 5 mm DTT. The enzyme reaction was stopped by the addition of 400 μL of chloroform. The chloroform layer was dried completely and then dissolved in 10 μL of 100% methanol. The reaction product was then spotted on a TLC plate and developed in a solvent mixture of chloroform:methanol (95:5, v:v) at room temperature. The degradation of toxoflavin and its derivatives were detected on the TLC plate under UV (254 or 365 nm). The experiment was replicated three times.
Vector construction for plant transformation
The BamHI-XbaI fragment of tflA was cloned into pCAMLA (Lee et al., 2005) to create pJ904 for rice transformation. The EcoRI-HindIII fragment of 35S::tflA from pJ904 was subcloned into a pCAMBIA1301 (Cambia, Canberra, Australia) carrying the hygromycin resistance gene and GUS reporter gene under the 35S promoter, resulting in p1301-tflA for Arabidopsis transformation.
Callus induction and transformation of rice
Callus induction and rice transformation were conducted according to standard protocols (Hiei et al., 1994). The standard protocols were modified for the toxoflavin/tflA selection as follows. Co-cultivated rice calli with pJ904 were initially incubated in the dark for 2 week on 2N6 media with 250 mg/L cefotaxime. These rice calli were then transferred to 2N6 media supplemented with 19 mg/L toxoflavin and 250 mg/L cefotaxime and incubated under continuous illumination for an additional 2 weeks (120–150 μmole). Selected toxoflavin-resistant calli were transferred onto the first regeneration media supplemented with 10 mg/L toxoflavin and 250 mg/L cefotaxime at 28 °C under continuous illumination for another 3 weeks. Selection on the second regeneration media supplemented with 10 mg/L toxoflavin under continuous illumination for 3 weeks resulted in shoot development. Final regeneration produced vigorous growth of toxoflavin-resistant transgenic plants.
Agrobacterium-mediated transformation of Arabidopsis thaliana
Arabidopsis thaliana (Col-0) was grown under a 16/8 h day/night cycle for 5 weeks at 22°C and then subjected to transformation using the floral dip method with A. tumefaciens GV3101 carrying p1301-tflA (Clough and Bent, 1998). Transgenic T1 seeds were surface-sterilized and germinated on each selection growth medium: 0.5 × MS medium (Murashige and Skoog, 1962) containing 1% (w/v) sucrose and 0.3% (w/v) MB phyto agar (MBcell, Seoul, Korea) supplemented with either 25 mg/L hygromycin B (Duchefa, MO, USA) or 7 mg/L toxoflavin. Various concentrations of toxoflavin (4–10 mg/L) were used for the wild-type A. thaliana Col-0 to determine the optimum selection concentration. Transformants were selected in the 7–10 mg/L range of toxoflavin.
Southern and immunoblot blot analyses
All transgenic Dongjinbyeo cultivars (T2) were confirmed by Southern hybridization (Sambrook and Russell, 2001) using a DNA fragment carrying the tflA gene as a probe. For immunoblot detection, a mouse polyclonal anti-TflA antibody was raised against a synthetic peptide (SDSPVKPPESKFSAEIRRC) and used as the primary antibody. Alkaline phosphatase–conjugated goat anti-mouse immunoglobulin G (Pierce Biotechnology, Rockford, IL, USA) was used as the secondary antibody, and positive signals were detected using One-Step NBT/BCIP solutions (Pierce Biotechnology, Rockford, IL, USA).
Histochemical GUS staining
Histochemical staining for GUS activity of transgenic Arabidopsis was performed as described previously (Jefferson et al., 1987).
This work was supported by the Crop Functional Genomics Center of the 21st Century Frontier Research Programs, which is funded by the Ministry of Education, Science and Technology of the Republic of Korea.