The Chrysanthemum x morifolium ribulose-1,5-bisphosphate carboxylase small subunit chloroplast signal peptide gene CTS was amplified from Impactvector 1.4 by PCR with primers CTS-1/CTS-2, and the G2-EPSPS gene was cloned from Pseudomonas fluorescens by PCR with primers G2-1/G2-2. The CTS::G2-EPSPS fusion gene was amplified by overlap PCR, using primers containing a BamH I restriction site at the 5′-end of forward primer (BamH I-CTS-1: GGATCCggatcctatggcctcgatctcttcc) and a Sac I restriction site at the 5′-end of reverse primer (Sac I-G2-2: GAGCTCgagctctcagtcgtttaggtgaacg), respectively. The PCR conditions were five cycles of 95°C for 25 s, 65°C for 25 s, and then 72°C for 70 s with no primers. Subsequently, primers were added into the PCR tube, and the cycle was run at 95°C for 5 min, followed by 29 cycles at 95°C for 25 s, 57°C for 25 s, and 72°C for 70 min; and finally 72°C for 5 min. The UBI promoter from pHAC25 was sub-cloned into pBI121 with Hind III and BamH I restriction sites to generate the pUBI121 vector. The fusion gene CTS::G2-EPSPS was inserted into the pUBI121 vector by BamH I and Sac I to generate the pUBI121::CTS::G2-EPSPS vector. Subsequently, the new vector was digested by Hind III and EcoR I and cloned into the multiple cloning site of pCAMBIA1301 to generate the vector 13UG2 (Figure 1A). The vector was introduced into Agrobacterium tumefaciens EHA105. Transformation of the Japonica cultivar ZH11 was performed based on the method reported by Hiei et al. (1994).

Rice genomic DNA was isolated from hygromycin resistance rice leaves using the modified CTAB method (Porebski et al., 1997) and amplified with the primers G2-F and G2-R1 (Supplementary Table S1). PCR was carried out with 50 ng of rice genomic DNA, 10 μl of 2× PCR mix (2× EasyTaq PCR SuperMix, Transgene, Beijing, China), and 0.5 μl each of G2-F (10 μM) and G2-R1 (10 μM). The PCR conditions were 95°C for 8 min, followed by 32 cycles of 95°C for 25 s, 53°C for 40 s, and 72°C for 1 min; and finally, 72°C for 5 min.

Total protein was extracted from 0.1 g of transgenic rice leaves using a Plant Protein Extraction Kit (CWBIO, CW0885, Beijing, China), and the amount of protein was calculated using the bicinchoninic acid (BCA) method. About 10 μg of protein was denatured at 95°C for 8 min, and then resolved by SDS-PAGE on 12% gels; proteins were then wet-blotted onto a nitrocellulose filter membrane. The G2-EPSPS protein was detected using the primary monoclonal antibody, anti-G2-EPSPS (3 mg/ml, 1:2000 dilution), which was obtained from immune mice using G2-EPSPS protein purified from the prokaryotic expression system. Horseradish peroxidase- (HRP)-labeled goat anti-mouse antibodies were used as the secondary antibody diluted at a ratio of 1:5000, and enhanced chemiluminescence (ECL) was used to detect the blot.

Aliquots of 40 μg purified DNA were digested with Hind III and EcoR I, respectively. The digested DNA fragments were separated on a 0.8% agarose gel and then blotted onto Hybond^TM-N⁺ nylon membrane (GE Healthcare UK Limited). The probe was amplified using primers G2-F/G2-R2 (Supplementary Table S1), and labeled with DIG (PCR DIG Probe Synthesis Kit, Roche, Germany). The blot was washed and detected according to the product instructions (DIG High Prime DNA Labeling and Detection Starter KitII, Roche, Germany).

For glyphosate-tolerance studies, seeds from each of the non-transgenic control ZH11 and homozygous transgenic plants (T₃ generation) were germinated in ultrapure water containing different concentrations of glyphosate at 37°C for 1 day in the dark. To test seed germination in ZH11 and transgenic lines, 40 seeds per dish were grown in 0, 50, and 100 ppm concentrations of glyphosate for 6 days. The glyphosate used in the report was Roundup, which contains 41% isopropyl amine salt of glyphosate (Monsanto, Malaysia), and 1000 ppm glyphosate, equivalent to 0.25% (v/v) Roundup.

For experimental analyses, seeds from the non-transgenic control ZH11 and transgenic lines (T₃ generation) were germinated in ultrapure water at 37°C for 1 day in the dark, and then transferred into soil in the greenhouse under controlled conditions. About 30 seedlings of ZH11 per pot were grown for 4 weeks, and then sprayed with 0, 1000, 2000, or 3000 ppm of glyphosate to test for lethal concentrations. The two homozygous transgenic lines were each sprayed with 0, 3000, 5000, 8000, 10,000, 12,000, 15,000, or 20,000 ppm of glyphosate at the tillering stage. Twelve days later, all treated plants were photographed and shoot height was measured.

For the glyphosate-tolerance assay, seeds of ZH11 and transgenic lines (T₄ generation) were transferred to 1/2 Yoshida hydroponic culture with 1 mM NH₄NO₃ (pH 5.5) containing different concentrations of glyphosate for 10 days, and the nutrient solution was changed every day. ZH11 was grown in 0, 0.01, 0.1, 0.2, 0.5, 1, 4, and 16 ppm glyphosate; transgenic plants were grown in 0, 5, 10, 25, 50, 100, 200, 800 ppm glyphosate for 10 days. The seedlings were kept in a growth chamber [28°C; 14-h/10-h light/dark photoperiod; 90% humidity; photon flux density, 500 μmol/m²/s photosynthetically active radiation (PAR)]. After 10 days, all the plants were collected. Plant height data at each glyphosate concentration were expressed as relative height, by comparing the mean height of each transgenic line at each concentration with its mean height at 0 ppm glyphosate. The relative heights at different concentrations could be fitted to a sigmoidal logistic model to produce dose-response curves (Seefeldt et al., 1995; Cui et al., 2016).

Total RNA was extracted from control and transgenic plant shoot tissues using RNAiso Plus (Takara Biotechnology, Dalian, China). Approximately 1.5 μg of total RNA was used to synthesize first-strand cDNA using a PrimeScript^TM RT reagent Kit with gDNA Eraser (Takara). In the transgenic plant expression analyses, transcript levels were normalized to those of the OsUBQ5 gene, and G2-EPSPS gene expression level was detected by the primers G2-151F/G2-151R (Supplementary Table S1). Real-time PCR was carried out in 7500 Real-Time PCR System (Applied Biosystems, United States) in 20 μl reaction mixtures containing 2 μl cDNA sample (1:20 dilution), 0.4 μl forward primers (10 μm), 0.4 μl reverse primers (10 μm), 0.4 μl Dye II and 10 μl SYBR Green Real-time PCR Mix (Takara Biotechnology, Dalian, China), the PCR conduction was at 95°C for 15 s, followed by 40 cycles of 95°C for 5 s and 60°C for 34 s.

The GenomeWalker^TM Universal Kit (Clontech, Takara) was used to isolate the flanking sequence of T-DNA; 25 μg of transgenic rice genomic DNA was digested with restriction endonucleases DraI, EcoR V, PvuII, and StuI, respectively, and the adaptor supplied in the kit was added. The specific primer GPS1/2 was designed and combined with the left boundary of the T-DNA. PCR was conducted according to the kit protocol. The sequence was analyzed using BLAST software in the NCBI database to search for the integration feature of T-DNA on the rice genome using the Rice Genome Annotation Project database².

For integration event-specific PCR detection analysis, the multiplex PCR method was used. The event-specific primers G₂-OsF/G₂-GR were designed according to the 5′-flanking sequence of the T-DNA of G2-6 transgenic rice and the sequence near the left-border of 13UG2. Additionally, OsSPS (sucrose-phosphate synthase 1, LOC_Os01g69030) specific primers SPSF/SPSR were designed, as a positive control in the multiplex amplification system (Supplementary Table S1). Multiplex PCR was conducted using genomic DNA isolated from transgenic rice with different integration events as the template. The PCR conditions were 95°C for 8 min, followed by 32 cycles at 95°C for 25 s, 54°C for 25 s, and 72°C for 30 s; and then extension at 72°C for 5 min.

To determine the detection limit for transgenic plants, the concentration of template DNA isolated from G2-6 was measured and diluted to 5, 10, 20, 40, 80, 160, and 320 copies, respectively. Then, to identify the detection limit using the integration-event-specific PCR system, a negative control was performed without using the DNA template.

A total of 46 independent transgenic plants were obtained. Rice genomic DNA was isolated from T₀ generation transgenic plants, and PCR amplification was performed with G2-EPSPS specific primers G2-F/G2-R1. A 1201 bp fragment was produced in the transgenic rice, whereas PCR products could not be amplified in ZH11, showing that the target gene was successfully integrated (Figure 1B). To determine G2-EPSPS protein expression in transgenic rice, proteins were extracted from PCR-positive transgenic rice leaves, and all the selected transgenic lines had G2-EPSPS protein expression; here only eight results were listed, as shown in Figure 1C. Southern blotting was performed with the T₀-generation plant DNA. Results showed that G2-6 and G2-7 contained a single copy of the G2-EPSPS gene (Figure 1D). The G2-EPSPS gene had similar expression levels in leaves of both G2-6 and G2-7 (Supplementary Figure S1).

Glyphosate can affect seed germination, and seeds are usually inhibited and bleached under glyphosate stress (Dun et al., 2007). To investigate the agronomic performance of the two transgenic lines, seeds and soil culture seedlings were treated with different concentrations of glyphosate. Seeds from ZH11, G2-6, and G2-7 were germinated in ultrapure water with 0, 50, and 100 ppm glyphosate for 6 days (Figure 2). The results showed that G2-6 and G2-7 could germinate well at 50 ppm and 100 ppm concentration of glyphosate; the radicle also grew normally at 100 ppm glyphosate, whereas ZH11 germination was obviously inhibited at 50 ppm glyphosate, and completely inhibited at 100 ppm glyphosate. These findings indicated that transgenic lines were more tolerant to glyphosate than the non-transgenic variety ZH11.

Four-weeks-old seedlings of the control ZH11 were treated with 0, 1000, 2000, and 3000 ppm glyphosate. After 12 days, all treated ZH11 plants were dead, leaving only the 0 ppm glyphosate treatment group alive (Figure 3A). This result indicated that in the field, 1000 ppm glyphosate is a lethal concentration for rice.

To determine the tolerance of the transgenic cultivars G2-6 and G2-7 to much higher glyphosate concentrations, they were treated with 0, 3000, 5000, 8000, 10,000, 12,000, 15,000, or 20,000 ppm glyphosate in the field at the tillering stage. After 12 days, plants under all treatments were fully expanded and green, but the shoot heights were different and were therefore used for statistical analysis of the agronomic performance of transgenic lines (Figures 3B,C). There were significant differences in plant shoot height between homozygous lines and glyphosate dosages (Table 1).

When treated with 3000 ppm or higher glyphosate, shoot heights of both transgenic lines were significantly decreased compared with untreated group, whereas at 8000, 10,000, 12,000, 15,000, and 20,000 ppm of glyphosate, the shoot heights of G2-6 and G2-7 exhibited no significant differences, respectively. Furthermore, shoot height of G2-6 was decreased by 15.3% when compared with G2-7 in the 20,000 glyphosate treatment group. These results indicated that G2-7 may show better agronomic performance than G2-6, although they exhibited similar expression levels. With further cultivation, the differences of shoot height disappeared at the heading stage; all treatments then progressed to the same height as the untreated group, and produced seeds normally.

There were significant differences in plant shoot height and plant fresh weight between homozygous lines, and with glyphosate dosage. In the hydroponic system, increase in shoot height and plant fresh weight of ZH11 could be significantly inhibited at 0.2 ppm glyphosate. The ZH11 seedlings were decomposed at 4 ppm glyphosate concentration, and according to the dose response curve, the relative seedling height was not obviously reduced with increasing glyphosate concentration; the results indicated that ZH11 was completely inhibited by 4 ppm glyphosate (Figure 4A and Table 2).

For G2-6 and G2-7, at increased treatment concentrations of glyphosate, shoot heights were significantly decreased to differing degrees (Table 3 and Figure 4A). For G2-6, shoot height and plant fresh weight was significantly inhibited at 10 ppm glyphosate, whereas for G2-7 the corresponding inhibitory concentration was 50 ppm. For both G2-6 and G2-7, shoot height and plant fresh weight could be completely inhibited at 800 ppm glyphosate; at this application rate the seedlings performed similar to ZH11 when treated with 4 ppm glyphosate.

To investigate the glyphosate-tolerance of the transgenic lines, dose-response curves were created using the relative shoot heights of ZH11, G2-6, and G2-7 under different glyphosate concentrations (Figure 4B). The results revealed that the I₅₀ for each line was 0.93, 96.4, and 114.07 ppm, respectively. The glyphosate-tolerance of G2-6 and G2-7 was 104- and 123-times that of ZH11, respectively.

To better understand the glyphosate-tolerance phenotype of the transgenic lines, we further analyzed G2-EPSPS gene expression levels in the shoots of plants under the 0 and 100 ppm glyphosate treatments. For both G2-6 and G2-7, there was no significant difference in G2-EPSPS gene expression level between 0 and 100 ppm glyphosate treatment (Figure 4C).

To analyze the integrity of transgene insertion in the plant genome, we isolated the left-border flanking sequences of T-DNA from the G2-6 transgenic line. Within the obtained 863 bp sequence, there were 612 bp showing 100% similarity to the 13UG2 vector sequence. The 251 bp upstream sequences exhibited 99% similarity to the sequence of Oryza sativa chromosome 8 from 23685037 to 23685287. The results indicated that the 5’-terminus of T-DNA was inserted at position 23685287 of the Oryza sativa chromosome 8. Further analysis showed that, in G2-6, the distance of the T-DNA integration sites to the upstream protein coding sequence (CDS) was 14 kb and to the downstream protein CDS it was 26 kb.

To develop event-specific PCR detection of G2-6 and ensure credibility of the PCR system, an endogenous reference gene OsSPS was selected as the positive control. The specific primers G2-OsF/G2-GR were designed according to the sequence of the rice genome and the left-border region of 13UG2, respectively (Figure 5A). A 554 bp fragment could be amplified from G2-6, whereas no PCR product could be amplified from other transgenic lines, ZH11, or the negative control without DNA template. A 151 bp fragment could be amplified from all transgenic lines and the non-transgenic line ZH11, whereas no PCR product could be amplified from the negative control (Supplementary Figure S2). The results confirmed that conventional PCR could identify the specific integration events of G2-6 and could therefore be used successfully for detecting transgenic rice G2-6.

To test the limits of the event-specific PCR system, the DNA template of transgenic rice G2-6 was diluted to 5, 10, 20, 40, 80, 160, and 320 copies. A weak 554 bp fragment could be amplified from five copies, and with an increase in the template copy number, the brightness increased (Figure 5B). The detection limits were acceptable under the regulatory requirements of China.