We constructed a double fluorescent protein-expressing Ac/Ds transposon vector in which the Ds element carrying a GOI is tandemly linked to Ac transposase (AcTPase), hygromycin phosphotransferase (HPT), GFP, and mCherry (pBDL11; Figure 1A). Compared with a single fluorescent protein-expressing Ac/Ds transposon vector (pLJ26; Gao et al., 2015), pBDL11 carries an mCherry fluorescent protein-expression cassette in addition to the GFP cassette.

The working pattern of the double fluorescent protein-expressing Ac/Ds vector is shown in Figure 1B. T-DNA-harboring T1 progeny can be identified by GFP and mCherry fluorescence assays, and non-fluorescent T1 progeny can be screened by PCR assays for Ds and the GOI to obtain marker-free transgenic plants.

In the Ds element, a Pi21-RNAi expression cassette that confers resistance to rice blast disease (Magnaporthe oryzae) was inserted into the AscI site as a GOI to construct pBDL23 (Figure 1C). pi21 is a recessive rice gene conferring broad spectrum and durable resistance to rice blast. RNAi knockdown of the dominant Pi21 gene was shown to confer rice blast resistance (Fukuoka et al., 2009; Zhang et al., 2016). For comparison and parallel control experiments, the Pi21-RNAi cassette was cloned into the single fluorescent protein-expressing Ac/Ds vector pLJ26 (Gao et al., 2015) to construct pBDL22 (Figure 1C).

pBDL23 and pBDL22 were transformed into rice cultivar Nipponbare via Agrobacterium. Rice calli were co-cultivated with Agrobacterium strains for 3 days and then cultured on hygromycin medium for 19 days for first-round selection. The rice calli expressing GFP and mCherry fluorescence were counted. The transformation efficiencies for pBDL23 and pBDL22 were 67.26% and 59.78%, respectively (Table 1). The hygromycin-resistant calli were subjected to a second-round selection culture for 20 days (Figure 2). The pBDL23-transformed calli expressing GFP and mCherry fluorescence and the pBDL22 calli expressing GFP were selected and cultured on plant regeneration medium to obtain rice primary transformants (T0 plants) with regeneration rates of 55.17% and 48.58%, respectively (Table 1). Subsequently, healthy double-fluorescent pBDL23-transformed plants and GFP-fluorescent pBDL22 plants were transplanted to soil until maturity, after which T1 rice seeds could be harvested.

To test the transposition activity of the Ds elements, transposon empty donor sites (EDS) in pBDL23- and pBDL22-transformed rice plants were amplified using specific PCR primers (Figure 3). EDS sequences were amplified in 60% and 40% of the pBDL23- and pBDL22-transformed T0 plants, respectively, suggesting the transposability of the Ds elements in rice.

T1 seeds of pBDL23-transformed rice lines were germinated and assayed for mCherry and GFP fluorescence. GFP and mCherry segregation ratios of the T1 lines were calculated, and chi-square fitting analyses for Mendelian genetic ratios (3:1, 15:1 and 63:1) were carried out for each rice line (Table 2). The T1 seedlings of pBDL22 rice lines were assayed for GFP fluorescence, and the GFP segregation ratios of T1 lines were then analyzed (Table 3). Among 42 pBDL23-transformed rice lines, 31 (73.8%) lines from the T1 generation showed Mendelian segregation ratios of 3:1, 15:1, or 63:1 for both GFP and mCherry; in five rice lines, neither GFP nor mCherry segregated with Mendelian ratios (Table 2). Five pBDL23-transformed rice lines showed Mendelian segregation ratios for GFP but not for mCherry, and one pBDL23 rice line showed Mendelian ratios for mCherry but not for GFP. Of 35 pBDL22-transformed rice lines, 32 (91.4%) showed GFP segregation with Mendelian ratios in the T1 generation (Table 3).

For each pBDL23-transformed rice line segregating in the T1 generation, double-fluorescent (GFP+ and mCherry+) and single-fluorescent (GFP+ or mCherry+) plants were discarded. Non-fluorescent T1 plants were assayed by PCR for Ds and the GOI (i.e., the Pi21-RNAi expression cassette). For Ds- and GOI-positive plants, PCR assays of the T-DNA-based GFP, mCherry, HPT, and AcTPase sequences were performed to confirm their absence. To streamline the screening of T1 rice populations, we used DNA pools of five non-fluorescent Ds/GOI plants from the same T1 line for extracts of rice genomic DNA (see Materials and Methods for further details). Each DNA pool was subjected to PCR assays for Ds and the GOI (Supplementary Image 1A), and further PCR assays confirmed the absence of GFP, mCherry, HPT, and AcTPase (Supplementary Image 1B). After identifying Ds-positive pools, individual plants from each DNA pool were analyzed by PCR to screen marker-free transgenic plants—i.e., Ds/GOI-positive and T-DNA-absent T1 plants (Supplementary Image 1C for pBDL23). For parallel experiments using the single fluorescent protein-expressing Ac/Ds vector pBDL22, single-fluorescent (GFP+) plants were discarded for each T1 rice line, and non-fluorescent plants were assayed by PCR to screen for marker-free Ds/GOI plants (Supplementary Image 1D) using the method described in Gao et al. (2015).

A total of 42 pBDL23-transformed T1 rice lines showed detectable mCherry and GFP fluorescence; after PCR screening of Ds/GOI, GFP, mCherry, HPT, and AcTPase sequences, marker-free Ds/GOI plants were obtained from 13 rice lines (Table 2). Non-fluorescent progeny from the abovementioned 13 T1 lines showed negative PCR results for GFP, mCherry, HPT, and AcTPase, suggesting that T1 progeny carrying the HPT selection marker, AcTPase, and other T-DNA sequences can be eliminated efficiently solely by GFP and mCherry fluorescence assays.

In a parallel experiment, we assayed 35 pBDL22-transformed T1 rice lines by detecting GFP fluorescence and PCR amplification of Ds/GOI, GFP, HPT, and AcTPase sequences. Marker-free Ds/GOI plants were obtained from six of these lines (Table 3). However, two of the pBDL22-transformed T1 lines segregated non-fluorescent plants that showed PCR amplification of T-DNA-based GFP, HPT, and AcTPase, suggesting that a single fluorescent protein marker (i.e., GFP in pBDL22) was less reliable than two fluorescent protein markers (i.e., GFP and mCherry in pBDL23) for counterselection against T-DNA-carrying progeny. Given that non-fluorescent plants derived from the single fluorescent protein-expressing Ac/Ds transposon vector pBDL22 might also carry T-DNA sequences, the screening of marker-free transgenic plants depends on both GFP fluorescence assays and additional PCR assays of T-DNA sequences.

The efficiencies of screening marker-free T1 plants from pBDL23 and pBDL22 rice lines were 31.0% (13/42) and 17.1% (6/35), respectively. In terms of the screening step of fluorescence assays, the occurrence of false marker-free transgenic plants was 0 for pBDL23 T1 lines and 25% (2/8) for pBDL22 lines.

To verify that the marker-free transgenic rice plants transformed with the double fluorescent protein-expressing Ac/Ds transposon vector pBDL23 contained Ds and the GOI (i.e., the RNAi Pi21 cassette) but not T-DNA sequences, rice genomic DNAs of marker-free T1 plants from three pBDL23-transformed rice lines were subjected to Southern blot hybridization using Ds, GFP, and HPT probes (Figure 4). Marker-free plants from the three rice lines showed distinctive patterns indicative of Ds hybridization (Figure 4A), suggesting that the Ds/GOI transgene was integrated into different sites in the rice genome. There were no hybridization signals for tested transgenic rice plants using GFP and HPT probes (Figures 4B,C). Therefore, these marker-free transgenic rice plants carried single Ds elements but -no T-DNA sequences, confirming that the transformation selection markers were eliminated.

To confirm whether the GOI functions properly in marker-free rice plants, T2 homozygous plants of the three rice lines were analyzed by qRT-PCR to determine the expression levels of the Pi21 gene (Figure 5). Pi21 expression levels were significantly suppressed in the marker-free rice plants compared with untransformed Nipponbare (NPB) rice.

Rice sequences flanking Ds insertions in marker-free transgenic rice lines 14YD270-1, 14YD273-2, and 14YD271-3 were isolated by thermal asymmetric interlaced (TAIL)-PCR (Supplementary Image 2A). The TAIL-PCR bands were sequenced, and the Ds flanking sequences were confirmed by PCR using a Ds-specific primer and a primer specific to the rice sequence derived from the TAIL-PCR bands (Supplementary Image 2B). BLAST search in the rice genome was performed using Ds-flanking rice sequences of the rice lines. As shown in Figure 6, the reintegrated Ds elements in 14YD270-1, 14YD273-2, and 14YD271-3 were mapped into the genomic positions Chr02:6618501, Chr03:369962, and Chr04:6664522, respectively. The Ds insertion sites in 14YD273-2 and 14YD271-3 were located in intergenic regions, while the Ds in 14YD270-1 was inserted in the coding region (CDS) of rice gene LOC_Os02g12670. Thus, the marker-free transgenic rice lines 14YD273-2 and 14YD271-3 are more suitable for utilization in rice disease resistance breeding.

To confirm whether the marker-free transgenic plants express rice blast resistance, three pBDL23-transformed marker-free T2 lines (14YD270-1, 14YD273-2, and 14YD271-3) were tested in rice blast inoculation. The rice blast lesions and disease scores of the marker-free rice lines (Supplementary Table 2) were significantly smaller than those of NPB rice (Figure 7).

We investigated the agronomic traits of marker-free transgenic rice. For each plant of the 14YD270-1 and 14YD273-2 T2 lines segregated for Ds, PCR for Ds was performed, and Ds-positive and -negative plants of each line were grouped separately. The agronomic traits spike number, flag leaf length, plant height, and spike length were evaluated for each plant. Student’s t-tests were conducted for Ds-positive and -negative plant groups in the lines 14YD270-1 and 14YD273-2. No significant difference was found between the Ds-positive and -negative groups in the same line for the above-mentioned agronomic traits (Table 4 and Supplementary Image 3). In addition, the heading date of each row of Ds-positive and -negative plants was investigated. The average heading dates of the Ds-positive and -negative groups of the 14YD270-1 line were 147.75 days, and the average heading dates for the Ds-positive and -negative groups in the line 14YD273-2 differed by only 1 day. Therefore, Ds insertion did not affect spike number, flag leaf length, plant height, spike length, and heading date in the 14YD270-1 and 14YD273-2 marker-free transgenic rice lines.

The “CaMV35S:mCherry:Tnos” expression cassette was amplified from pBDL09 using the following PCR primers: mCherryinF (5′GGGGGCCCGGTACCGAGCTCCCCC TCAGAAGACCAG3′) and mCherryinR (5′ATGATTACGAA TTCGAGCTAGCTAGTAACATAGATGACACCGC3′). The large fragment of SacI-digested pLJ26 (Gao et al., 2015) was ligated with the PCR-amplified “CaMV35S:mCherry:Tnos” cassette using an In-Fusion HD Cloning kit (Clontech, United States) to construct pBDL10A. One of the PCR primers was designed in such a way that the resulting pBDL10A construct retained a single SacI restriction site between the mCherry and GFP cassettes (likewise, in each of the following In-Fusion Cloning steps using AscI, one primer was designed in a similar way to retain a single AscI site in each new construct). The Ds fragment isolated from SacI-digested pLJ26 was inserted into the SacI site in pBDL10A to obtain pBDL11. The rice Actin 1 promoter was amplified from pZJ63 using PCR primers BDL14-F4 (5′CTAATTGAATGGaGCGCCGTCGAGGTCATTCATATGC T3′) and BDL14-R (5′GCAGGAATTCGGCGCCCCTACAAAAA AGCTCCGCAC3′). The Actin 1 promoter was inserted into the AscI site inside the Ds of pBDL11 via In-Fusion Cloning to construct pBDL14. The pBDL14 vector retained a single AscI site downstream of the Actin 1 promoter. T3A (the Pisum sativum RbcS-3A polyA signal sequence) was amplified from pZJ70 using primers PBDL15-F (5′TTTTTGTAGGGGCGCGCCAGCTTTCGTCCGTATCATC G3′) and PBDL15-R1 (5′GCAGGAATTCGGAGCACCTCGACA AAAAGCCTATAC3′) and then inserted into the AscI site, downstream of the Actin 1 promoter, in pBDL14 by In-Fusion cloning to construct pBDL15. The Pi21 -RNAi cassette for the expression of rice blast resistance (Fukuoka et al., 2009; Zhang et al., 2016) was digested from pBDL18 using AscI and then cloned to a position between the Actin 1 promoter and T3A in pBDL15, thereby constructing the double fluorescent protein-expressing Ac/Ds transposon vector pBDL23.

The Pi21-RNAi cassette was cloned into the Ds element in pLJ26 (Gao et al., 2015) to construct a single fluorescent protein-expressing Ac/Ds transposon vector pBDL22. pBDL22 was used as a control for comparison with the double fluorescent protein-expressing vector pBDL23, which was constructed as follows. First, the Actin 1 promoter was amplified from pBDL07 and inserted into the AscI site within the Ds of pLJ26 by In-Fusion cloning to construct pBDL20. T3A was then amplified from pBDL07 and inserted into the AscI site downstream of the Actin 1 promoter in pBDL20 by in-fusion cloning to construct pBDL21. The Pi21-RNA cassette was released from pBDL18 by AscI digestion and inserted between the Actin 1 promoter and T3A in pBDL21 to obtain pBDL22.

The full-length sequences of pBDL11, pBDL23, and pBDL22 are deposited in the GenBank sequence database (their GenBank accession numbers were MW653811, MW653812, and MW653813, respectively).

pBDL23 and pBDL22 were transformed into Agrobacterium strain EHA105 via electroporation. Rice transformation was performed as previously described (Gao et al., 2015; Chen et al., 1998). The seeds of rice cultivar NPB were dehusked, surface-sterilized, and cultured on NB medium containing 2,4-D to induce embryogenic calli. Rice calli were co-cultivated with Agrobacterium strains harboring pBDL23 and pBDL22, respectively, for 3 days. The calli were then washed with sterilized water to remove Agrobacterium cells and cultured on NB medium containing 50 mg/L hygromycin. Hygromycin-resistant rice calli derived from one or two rounds of selection culture were transferred onto a differentiation medium to regenerate rice plants. The hygromycin-resistant rice plants were examined under a lamp-type fluorescence detector (Fluorescent Protein Macro Detector Set, Biological Laboratory Equipment Maintenance and Service Ltd., Budapest, Hungary) to detect GFP and mCherry fluorescence signals, which were used as criteria to select T0 plants. For BDL23, double-fluorescent (GFP+ and mCherry+) rice plants were selected while single-fluorescent (GFP+ or mCherry+) and non-fluorescent (GFP- and mCherry-) plants were discarded. For pBDL22, GFP-fluorescent T0 plants were selected. Rice transformants were transplanted to soil, and T1 rice seeds were harvested at the maturity stage.

Independently transformed rice lines with at least 100 T1 seeds per line were tested for transgene segregation analysis and marker-free selection. For NPB and each of the pBDL23 and pBDL22 T1 lines, 100 seeds were soaked in water at 37°C for 2 days. The seeds were then washed with water, wrapped in a wet towel, and covered with plastic wrap to retain moisture; incubation continued at 37°C. After 5 days of germination, rice seedlings were observed under the BLS lamp-type fluorescence detector with blue and green excitation light sources to detect GFP and mCherry fluorescence, respectively, to distinguish between mCherry- and GFP-positive plants. Fluorescent and non-fluorescent plants from each line were counted, and chi-square statistical tests were performed. Fluorescent and non-fluorescent plants from each T1 line were grown together in the same pot and kept in a Conviron PGR15 plant growth chamber (Canada) with a photoperiod of 14 h (light)/10 h (dark), light intensity 4 (370 μmol/m²/s), 26°C (light)/24°C (dark), and 85% humidity.

To extract rice genomic DNA, we grew germinated T1 rice seeds in soil for 8 days, and leaf tissue was then sampled from rice seedlings. To screen marker-free transgenic plants, we sampled an equal amount of leaf tissue from each of five non-fluorescent plants from the same T1 line (i.e., to create a DNA pool), and these samples were mixed for rice DNA extraction. The DNA pools were PCR assayed to detect Ds and the GOI (see below). When DNA pools showed positive PCR assays for Ds and the GOI, rice DNA was extracted from each of the single plants from each pool and assayed individually by PCR. The steps of rice DNA extraction were as follows: 2 cm of leaf tissue was cut from each seedling and placed into a 1.5 ml centrifuge tube, which was added with steel beads and 500 μl of TPS extraction solution (0.1 M Tris-HCl, 0.01 M EDTA, 1 M KCl, pH 8.0). Rice leaf tissue was homogenized using a Tissuelyser-FE III tissue grinder (Jingxin Industrial Development, Shanghai, China) at 70 Hz for 90 s. Homogenized leaf tissue was incubated in TPS extraction solution at 75°C for 30 min (in a water bath) and centrifuged at 12,000 rpm (Centrifuge5430, Eppendorf, Germany) for 10 min. About 200 μl of supernatant was transferred to a new tube and mixed with an equal volume of isopropanol. This mixture was placed at −20°C for 30 min and centrifuged at 12,000 rpm (Centrifuge5430, Eppendorf, Germany) for 5 min. The DNA pellet was briefly washed with 75% of ethanol, air-dried, and dissolved in 50 μl of double-distilled H₂O containing 0.2 mg/ml RNase A.

The Ds transposon empty donor sites (EDS) in pBDL23-transformed rice plants were amplified using the PCR primers GFP-EF (5′ACAACCACTACCTGAGCACC3′) and mCherry-ER (5′TTGGAGCCGTACATGAACTG3′), and the primers GFP-EF and HPT-ER (5′ATCGAAGCTGAAAGCACGAG3′) were for the EDS-PCR of pBDL22 transformants. The cycling parameters for EDS-PCR were as follows: 95°C for 5 min; 33 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 2 min; and 72°C for 7 min. The primer pair of mCherry-F1 (5′CAAGGCCTACGTGAAGCACC3′) and TNOS-R (5′CCATCTCATAAATAACGTCATG3′) and the pair of GFP-EF and TNOS-R were used to amplify mCherry and GFP, respectively. The cycling parameters for mCherry and GFP were as follows: 95°C for 5 min; 33 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 1 min; and 72°C for 7 min.

PCR primers specific to pBDL22 and pBDL23 T-DNA sequences are listed in Supplementary Table 1. Rice DNA pools derived from the non-fluorescent T1 segregants were PCR assayed using four pairs of primers specific to Ds and the GOI. DNA pools showing positive PCR assays for Ds and the GOI were further analyzed using primers specific to T-DNA sequences other than Ds and the GOI. For DNA pools showing negative results for PCR assays of T-DNA sequences, rice DNA was extracted from individual plants corresponding to each pool and assayed by PCR for both Ds and the GOI to screen for marker-free Ds/GOI plants. PCR reactions were performed on an ABI Applied Biosystems Veriti 96-Well Thermal cycler (Gene Company, United States) with Taq DNA polymerase (TsingKe Biological Technology, China). The cycling parameters for both GOI and Ds assays were as follows: 95°C for 5 min; 36 cycles of 95°C for 30 s, 56°C for 30 s, and 72°C for 1 min; and 72°C for 7 min. The cycling parameters for T-DNA assays were as follows: 95°C for 5 min; 26 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 1 min; and 72°C for 7 min.

Southern hybridization was carried out using a Southern blot hybridization kit (Towin Biotechnology, China). About 10 μg of rice DNA was first digested with EcoRI (for the Ds probe), HindIII (for the HPT probe), or BamHI (for the GFP probe). The digestion was then loaded onto an agarose gel and electrophoresed at 30V for 16 h. Enzymatic products fractionated in the electrophoresis gels were transferred onto nylon membranes and fixed by UV light. About 30 min of pre-hybridization at 37°C was followed by overnight hybridization with a digoxigenin-labeled probe. The hybridization signal was revealed by reaction with the chemiluminescent substrate CSPD and detected visually by X-ray. We used 822 bp, 721 bp, and 840 bp PCR fragment sequences as hybridization probes for the detection of Ds, GFP, and HPT, respectively, as described previously (Gao et al., 2015).

In this study, 2 cm of rice leaf tissue was sampled from single rice seedlings and placed into a 2 ml centrifuge tube. Steel beads were added, and the tissue samples were homogenized at 50 Hz for 90 s using a tissue grinder after liquid nitrogen precooling of centrifuge tube racks. Subsequently, 1 ml of Trizol was added to the leaf powder, and the mixture was incubated at room temperature for 5 min. Rice RNA was extracted by adding 200 μl of chloroform/isoproterenol (24:1) and centrifuging the resulting solution at 12,000 rpm (Centrifuge5415R, Eppendorf, Germany) for 10 min at 4°C. The supernatant was transferred to a new tube to which 500 μl of isopropanol was added. The solution was gently mixed and stored at −20°C for 10 min and then centrifuged at 12,000 rpm (Centrifuge5415R, Eppendorf, Germany) and 4°C for 10 min. The nucleic acid precipitate was washed by adding 1 ml of 75% ethanol, followed by centrifugation at 9,900 rpm (Centrifuge5415R, Eppendorf, Germany) and 4°C for 5 min and drying in air. The resulting RNA pellet was then dissolved in RNase-free water and stored at −80°C.

Reverse transcription of rice RNA was performed using the TransScript All-in-one First Strand cDNA Synthesis SuperMix for qPCR (TransGen Biotech, China) Reverse Transcription Kit as per the manufacturer’s instructions. qPCR was carried out with the TranStart Top/Tip Green qPCR SuperMix (TransGen Biotech, China) kit using the Bio-Rad CFX96 Real-time System. The cycling parameters were as follows: 94°C for 30 s; 39 cycles of 94°C for 5 s, 58°C for 15 s, and 72°C for 10 s. OsEF-Tu was used as an internal reference gene, and relative expression levels of the target gene were calculated by the 2^{–Δ Δ} CT method using three technical replicates for each sample. The OsEF-Tu primers were EF-Tu-F (5′ATGATCACGGGTACCTCCCA3′) and EF-Tu- R (5′TTGTCAGGGTTGTAGCCGAC3′); the Pi21 primers were Pi21-RT-F (5′CGGCAAATTTGACAGATGGGTAT3′) and Pi21-RT-R (5′CTTCTCCGGGTCGAACTTC3′).

Ds flanking sequences in marker-free transgenic rice lines were amplified by TAIL-PCR (Liu et al., 1995) and hiTAIL-PCR (Liu and Chen, 2007). Two to three successive rounds of nested PCR reactions were performed using the primers specific to the 5′ Ds- and 3′ Ds terminal sequences (Kolesnik et al., 2004) and arbitrary degenerate (AD) primers. For each TAIL-PCR or hiTAIL-PCR reaction, a 5′ Ds- or 3′ Ds-specific primer was paired with an AD primer (AD2; Liu et al., 1995) or a long arbitrary degenerate (LAD) primer (LAD-1, LAD-2, LAD-3 or AC1; Liu and Chen, 2007). PCR products were fractionated on 1 to 1.5% (w/v) agarose gels. PCR bands were cut from the gels and purified for sequencing. The Ds flanking sequences were confirmed by PCR using a Ds-specific primer and a primer specific to the rice sequence derived from the sequencing results of TAIL-PCR bands.

Rice blast disease inoculation was performed as previously described (Yang et al., 2017) with slight modifications. Rice blast spores that had been stored on filter paper at −20°C were cultured on CM medium at 25°C with a photoperiod of 12 h light/12 h dark. After culturing for 12 days, rice blast spores were washed from CM medium using sterile water containing 0.02% Tween 20. The collected spores were counted under a microscope using a hemocyte counting plate. Spore concentration was adjusted to 5 × 10⁵ spores/ml with 0.02% Tween 20 (Li et al., 2016). Germinated rice seeds were sown in soil pots and grown for 10 days in a Conviron plant growth chamber (Canada) with a photoperiod of 14 h (light)/10 h (dark), light intensity 4 (370 μmol/m²/s), and 26°C (light)/24°C (dark) at 85% humidity. Each pot contained 18–20 rice seedlings, and 3.5 ml of the spore suspension was used for rice blast inoculation. About 2 ml of spore suspension was initially sprayed onto the rice seedlings using a spray gun to coat all leaves with water droplets. Rice seedlings were then protected using a plastic bottle whose cap and bottom had been removed, and 1.5 ml of spore suspension was sprayed into the mouth of the bottle to ensure that the bottle remained full of fog. The bottle mouth was then sealed with plastic wrap. The rice seedlings were kept in a plant growth chamber at 25°C in the dark for 34 h, after which the photoperiod of 26°C (light)/24°C (dark) was resumed. Disease symptoms were examined 5 days post inoculation.

Rice seeds of marker-free T2 transgenic rice lines that segregated for Ds (GOI) were planted. Total genomic DNA was extracted from 50 mg of leaf tissue sampled from each rice seedling for PCR assay of Ds. Ds-positive and -negative T2 plants were grouped separately. For the Ds-positive and -negative plant groups of each transgenic rice line, each group was planted into two rows. Spike number, flag leaf length, plant height, and spike length were investigated for each plant. Spike number was counted at rice maturity. Five tillers were obtained randomly for each plant to measure the flag leaf length and spike length separately, and average flag leaf length and average spike length of individual plants were calculated. Date of rice heading was observed for each row (six plants in total) of Ds-positive or negative group, and the date of 50% plants heading was taken as the date of rice heading. The number of days from the date of root and bud emergence during seed germination to the date of rice heading was calculated as the heading date. The average heading date of two rows of Ds-positive or negative plants was calculated for each T2 transgenic rice line.