Introduction

Glyphosate (N-phosphonomethyl glycine) is one of the world’s most extensively used broad-spectrum herbicides. It is effective against the majority of annual and perennial grasses and broad-leaved weeds. In plants, the 5-enolpyruvylshikimate-3-phosphate synthase (EPSP) enzyme normally catalyzes shikimate-3-phosphate (S3P) and phosphoenolpyruvate (PEP) into EPSP and inorganic phosphate (Haslam 1993) [Electronic Supplementary Material (ESM) Fig. 1). Glyphosate cannot bind free EPSPS enzyme but instead forms a EPSPS–S3P–glyphosate complex, which ultimately stops aromatic amino acid biosynthesis by the plant via the shikimate pathway (Hollander-Czytko and Amrhein 1980; Herrmann 1995). Over-expression of the EPSPS gene could reduce this competitive binding affinity and allow the plant to synthesize aromatic amino acids and function normally in the presence of glyphosate (Padgette et al. 1995a, b; Priestman et al. 2005a, b).

Glyphosate resistance was first reported in transgenic tobacco expressing the P101S substitution mutant of Salmonella typhimurium EPSPS (Comai et al. 1985; Stalker et al. 1985). Many EPSPS genes, including the wild-type class II EPSPS and the mutant class I EPSPS gene (aroA), have been reported to improve glyphosate tolerance when engineered in transgenic plants (Barry et al. 1992; Chen et al. 1999; Ye et al. 2001; He et al. 2003; Wang et al. 2003). Class I EPSPS enzymes include those present in Escherichia coli, Aeromonas salmonicida, and Arabidopsis thaliana, all of which are naturally sensitive to glyphosate (Lewendon and Coggins 1983). Resistance to glyphosate can be achieved through mutation of the target enzyme, such as by mutating the residues around the glyphosate-binding site (Stalker et al. 1985; Padgette et al. 1991; Shuttleworth et al. 1999; Eschenburg et al. 2002; He et al. 2003; Priestman et al. 2005a, b; Haghani et al. 2008). Class II EPSPS enzymes are naturally tolerant to glyphosate, and some such enzymes have been identified in Agrobacterium tumefaciens CP4, Bacillus subtilis, and Pseudomonas sp. strain PG2982 (Fitzgibbon and Braymer 1990; Barry et al. 1992). Unfortunately, transgenic plants that over-express most of the EPSPS genes described above have failed to show sufficient glyphosate resistance for commercial utilization (Bradshaw et al. 1997). The transgene in most commercial glyphosate-resistant crops is the EPSPS gene from Agrobacterium spp. CP4 (Pilacinski 2002; Funke et al. 2006).

Structural studies have confirmed that PEP and glyphosate have the same binding site and that glyphosate inhibition is competitive with PEP (Boocock and Coggins 1983; McDowell et al. 1996; SchÖnbrunn et al. 2001; Funke et al. 2006, 2009; Pollegioni et al. 2011). Class I EPSPS enzyme mutants always exhibit decreased affinity for substrate PEP with an increased tolerance to glyphosate (Padgette et al. 1995a, b). Only those glyphosate-resistant EPSPS enzymes with a high PEP binding affinity are suitable for commercial plants (Funke et al. 2009). It has been reported that many organisms isolated from glyphosate-contaminated soil show the spontaneous occurrence of glyphosate resistance. For example, AroA P. Putida from glyphosate-contaminated soil was found to exhibit high tolerance to glyphosate without any decrease in its affinity for PEP (Sun et al. 2005). The T97I/P101S-mutated E. coli EPSP synthase, which was found to confer insensitivity to glyphosate, has been used to generate herbicide-resistant varieties of corn GA21 (Funke et al. 2009). In summary, novel EPSPS genes from glyphosate-contaminated soil and EPSPS genes mutated at multiple residues are potential candidates for use in the engineering of crops with acceptable levels of herbicide tolerance for commercialization (Sun et al. 2005; Yi et al. 2007; Li et al. 2009; Tian et al. 2011; Sun et al. 2012; Zhou et al. 2012).

In this study, we isolated a new EPSPS gene from glyphosate-contaminated soil. The full gene [Gr5 aroA ; GenBank Acc. no. (Bankit tool) 1013095] was 1,819 bp and contained a 1,341-bp open reading frame (ORF) encoding a protein of 447 amino acids. Homology analysis and molecular modeling revealed that Gr5 aroA has a low homology with other EPSPS genes, such as the CP4 EPSPS gene from Agrobacterium spp. and the aroA genes from E. coli and Typhimurium. We also identified the function of this EPSPS gene by generating transgenic Gr5 aroA tobacco plants. The level of glyphosate tolerance and other agricultural characteristics of transgenic Gr5 aroA tobacco plants were evaluated.

Materials and methods

Construction of the metagenomic library from soil heavily contaminated with glyphosate

Glyphosate-contaminated soil was collected from the ground near the glyphosate storage and transportation room of the Hebei Qixing Glyphosate Production Co. (Hebei Province, China). Metagenomic DNA was extracted from the soil according to the method described by Griffiths (Griffiths et al. 2000; Kauffmann et al. 2004). Crude DNA was purified by agarose gel electrophoresis using a QIAEXII Gel Extraction kit (Qiagen, Venlo, The Netherlands). Purified DNA was then partially digested with Sau3AI (Promega, Madison, WI) and concentrated with electrophoresis on a 1 % agarose gel. DNA fragments of 3–5 kb were then ligated into Sau3AI-digested pACYC184 plasmid DNA using T4 DNA Ligase (Promega). The recombinant plasmids were electroporated into E. coli ER2799 (aroA-deleted mutant, New England Biolabs, Lpswich, MA) to generate a metagenomic library (Yanisch-Perron et al. 1985; Dower et al. 1988; Sambrook et al. 1989).

Cloning and sequence analysis of aroA gene

The DNA metagenomic library was screened on MOPS minimal medium supplemented with chloramphenicol, kanamycin, and 10 mM glyphosate. Positive clones were transferred onto MOPS minimal medium supplemented with chloramphenicol, kanamycin, and 150 mM glyphosate for the second screening. One colony harboring a plasmid with a 2-kb insert fragment was selected. The insertion fragment was then sequenced and named Gr5 aroA .

Structural analyses and phylogenic analysis of aroA genes

Gr5aroA protein was predicted based on sequence similarity by searching against the cDNA sequences in the GenBank database. For functional domain identification, we used Scan PROSITE on the ExPASy Web server for PROSITE motif database screening (http://ca.expasy.org/tools/scanprosite/) and NCBI RPS-BLAST for Pfam and SMART to screen conserved domain databases (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi).

EPSPS and AroA proteins from different organisms were used in our evolutionary analyses: Escherichia coli (P07638), Aeromonas salmonicida (Q03321), Arabidopsis thaliana (P05466), Nicotiana tabacum (P23981), Petunia hybrida (P11043), Zea mays (CAA44974), Bordetella pertussis (P12421), Agrobacterium tumefaciens CP4 (Q9R4E4), Bacillus subtilis (P20691), Staphylococcus aureus (Q05615), Dichelobacter nodosus (Q46550), Streptococcus pneumoniae (Q9S400), Pseudomonas putida (4G-1) (AJ812018), Pseudomonas sp. strain PG2982 (P56952), P. fluorescens Pf-5 (AAY93565), P. entomophila L48 (YP_607169), P. aeruginosa UCBPP-PA14 (ABJ12389), P. mendocina ymp (ZP_01526458), P. syringae pv. phaseolicola 1448A (YP_273200), and P. putida KT2440 (AAN67390). Deduced amino acid sequences of AroA were aligned by Clustal_X with manual adjustments (Jeanmougin et al. 1998). Phylogenetic trees were reconstructed using MEGA 4 (Kumar et al. 2004). These trees were based on the neighbor-joining method with the p-distance model. Gap sites in the alignment were not used in the phylogenetic reconstruction (the complete-deletion option). The reliability of the estimated trees was evaluated by the bootstrap method, with 1,000 pseudo-replications (Felsenstein 1985).

Gr5 aroA protein expression

The DNA fragment containing the coding region of Gr5 aroA was obtained by PCR using the pACYCG2 plasmid as the template with the following two primers: primer 1 (5′-CGGGA TCCAT GGCGT GTTTG CCTGA TGA-3′) and primer 2 (5′-CCAAG CTTTC AGGCA AACAC CTCGAG-3′). The fragment was digested with BamHI and HindIII and then cloned into the corresponding restriction sites of the pET-28a vector (Novagen, Madison, WI). The expression vector pET-28a harboring Gr5 aroA was confirmed by DNA sequencing. The plasmid was then transformed into E. coli BL21 (DE3) competent cells, and the positive E. coli BL21 clone was grown at 37 °C in 100 ml LB medium containing 50 μg/ml kanamycin. Cells were collected and diluted to a density 0.60 at OD600. Isopropyl β-d-1-thiogalactopyranoside was added to a final concentration of 1 mM for the induction of Gr5aroA protein expression. After 4 h, the cells were collected by centrifugation and re-suspended in 10 ml of 50 mM Tris buffer (pH 7.0) with 0.1 mM dithiothreitol. The cell suspension was frozen at −70 °C and then thawed to room temperature. Cells were lysed by sonication. The crude homogenate was clarified by centrifugation at 12,000 rpm for 30 min at 4 °C. The resulting supernatants were assayed for protein expression.

Enzyme activities assay of Gr5 aroA

Gr5aroA protein activity was measured by the production of inorganic phosphate using the malachite green dye assay method (Lanzetta et al. 1979). The standard reaction was carried out at 28 °C in a final reaction volume of 100 μl containing 50 mM HEPES, pH 7.0, 1 mM S3P, 1 mM PEP, 0.1 mM (NH4)6Mo7O24·2H2O, and purified enzyme. After incubation for 3 min, 1 ml of malachite green–ammonium molybdate colorimetric solution was added and the solution mixed thoroughly, following which 0.1 ml of 34 % sodium citrate solution was added. After a 30-min incubation at room temperature, samples were measured at 660 nm. The same reaction solution without S3P was used as the control.

Transformation and regeneration of transgenic Gr5 aroA tobacco plants

The Gr5 aroA coding sequence was inserted into the region between the CaMV35S promoter and the nopaline synthase terminator (Nos-ter) on Agrobacterium binary vector pBI121 (Rogers et al. 1986). The resulting vector, pBI-Gr5 aroA , contained the selectable marker neomycin phosphotransferase gene (nptII) and the Gr5 aroA gene driven by the CaMV35S promoter (ESM Fig. 2). The expression cassette was then introduced into A. tumefaciens strain EHA 105 by tri-parental mating using a previously described protocol (Hood et al. 1993; Horsch et al. 1988). and subsequently used to generate transgenic tobacco plants.

Tobacco (Nicotiana tobacum) cv. Petit Havana SR1 was used for transformation following the protocol described by Horsch et al. (1988) (see Murashige and Skoog 1962). The kanamycin-resistant plants were subjected to PCR, Southern blot, and quantitative real time (qRT)-PCR analyses to establish the presence and expression of the Gr5 aroA gene.

Verification of transgenic plants

Genomic DNA was extracted from young kanamycin-resistant leaves according to the method described by Doyle and Doyle (1990), and then PCR analysis was carried out for the detection of the Gr5 aroA gene in transgenic plants using the forward primer Gr5 aroA F1 (5′-ATGGC GTGTT TGCCT GATGAT-3′) and the reverse primer Gr5 aroA R1 (5′-GTGGA TTTGC TGACT GTGTGT-3′). The expected product size was 1.34 kb. The PCR reactions were carried out in a total volume of 25 μl containing 50 ng tobacco genomic DNA, 50 mM KCl, 10 mM Tris–HCl (pH 8.3), 1.5 mM MgCl2, 200 μM dNTPs, 1.25 U of Taq DNA polymerase, and 25 pmol of each primer. For PCR amplification, DNA was denatured at 94 °C for 3 min followed by 30 amplification cycles of 94 °C for 50 s, 53 °C for 60 s, and 72 °C for 90 s, with a final extension at 72 °C for 10 min.

Southern blot analysis was further used to confirm the transgenic status of plants. Nine transgenic tobacco plants (No. 1, 3, 7, 12, 13, 15, 18, and 29) were analyzed by Southern blot hybridization for the presence of the Gr5 aroA gene in the tobacco plant. Aliquots of genomic DNA (8 μg) were digested overnight at 37 °C with XbaI (which cuts only once in pBI Gr5 aroA ), fractionated by 0.8 % agarose gel electrophoresis, and transferred to a Hybond-N+ membrane (Amersham Bioscience, Amersham, UK). The Gr5 aroA probe (ORF) was labeled with [32P]dATP using a random-priming kit (Ready-to-Go; Pharmacia, Freiburg, Germany). Hybridization was carried out according to the method described by Thomas (Thomas 1980). Signals were visualized by exposure to Fuji X-ray film (Fuji, Tokyo, Japan) at −70 °C for 2 days.

The qRT-PCR analysis was used to investigate Gr5 aroA gene expression according to the manual of SYBR premix Ex-Taq (Takara, Otsu, Japan). The specific primers (5′-CAC CACCT GGCGC GTCGC-3′, 5′-GGTCG GGATC GTATC CTG-3′) were used to amplify the Gr5 aroA gene. The ubiquitin gene (Accession number: U66264.1) was used as the control. Each sample was tested three times. The results were analyzed by the comparative ΔCT method and Option 3 software.

Glyphosate tolerance assay of T 0 transgenic plants and T 1 transgenic lines

Nine independent transgenic tobacco T 0 lines (No. 1, 3, 7, 12, 13, 15, 18, and 29) were used in the glyphosate tolerance assay. Plants of the independent transgenic lines and wild-type (WT) (untransformed) plants were micro-propagated, and six plants of each were grown under a 14/10-h light/dark photoperiod with an irradiance of 50 μmol m−2 s−1 provided by white fluorescent tubes. Both the transgenic and WT tobacco plants were then transferred into soil in the greenhouse to determine their glyphosate tolerance. In this assay, all plants were sprayed at the five- to six-leaf stage (about 2 weeks after transplanting) with a 41 % glyphosate salt solution (the active ingredient) to generate six levels of injury: 2.35 l/ha (1.0 kg a.i./ha; 0.5× recommended concentration), 4.7 l/ha (2.0 kg a.i./ha; 1×), 9.3 l/ha (3.9 kg a.i./ha; 2×), 14.0 l/ha (6.0 kg a.i./ha; 3×), 18.6 l/ha (7.8 kg a.i./ha; 4×), and 23.0 l/ha (9.8 kg a.i./ha; 5×). Vegetative injury was determined by visual observation 2 weeks after treatment. Agronomic characteristics were investigated at plant maturity.

T 1 seeds were harvested from different glyphosate-tolerant T 0 plants. Both WT and T 1 seeds were germinated in MS medium without kanamycin. Positive and negative plants were confirmed using PCR. At least ten plants from each independent transgenic line and WT and non-transgenic tobacco plants were transferred into soil in the greenhouse for a second glyphosate tolerance assay using the same procedure as in the assay of T 0 plants.

Analysis of herbicide resistance was performed according to the method described by Ye et al. (2001). Individual plants were scored for tolerance to glyphosate using a scale ranging from 0 (normal plant, fertile, no delay in maturity, seeds setting normally), through to 1 (small chlorosis on the leaves, fertile, no delay in maturity, seeds setting normally), 2 (less than one leaf wilted 2 weeks after treatment, fertile, no delay in maturity, seeds setting normally), 3 (two or more leaves wilted after treatment, stunt, delay in maturity, seeds setting un-normally), and 4 (wilt, dead plant). The injury data for all treated plants were averaged. Average values of <2.0 were considered to indicate tolerance relative to the treatment, while average values >2.0 were considered to indicate intolerance relative to the treatment.

Genetic analysis of segregation of the Gr5 aroA gene in T 1 progenies

Seven independent T 0 transgenic plants (No. 1, 3, 7, 12, 13, 15, 18, and 29) were grown to maturity. T 1 seeds were harvested and sown in soil in the greenhouse. The germinated T 1 plants (2 weeks old) were analyzed for the presence of the Gr5 aroA gene by the PCR method previously described for the segregation pattern analysis. Northern bolt analysis was performed to determine Gr5 aroA gene expression in T 1 progenies of the same seven lines.

Results

Isolation of genes involved in glyphosate tolerance from the metagenomic library

In order to clone novel glyphosate tolerance genes, we constructed a soil metagenomic library with 5.06 × 105 clones. This library was then screened on glyphosate-resistant MOPS plates using mutant ER2799 as the recipient. After the first round of screening, 28 clones were found to tolerate 10 mM glyphosate. These positive clones were then transferred onto MOPS-minimal-medium plates containing chloramphenicol, kanamycin, and 150 mM glyphosate for the second screening. One colony harboring a plasmid with an insert of about 2 kb was selected based on the results of the second screening; the insert was named Gr5 aroA , and was subsequently sequenced and analyzed.

Bioinformatic analysis of Gr5 aroA gene

Sequence analysis showed that the Gr5 aroA gene was 1,819 bp and contained a 1,341-bp ORF encoding a protein of 447 amino acids (ESM Fig. 3). The deduced Gr5 aroA aroA protein was found to have 30 % identity with E. coli aroA and 70 % identity with Pseudomonas syringae pv. B728a at the amino acid level.

To determine whether the Gr5 aroA gene could be translated, we inserted the coding region of Gr5 aroA into pET-28a (E. coli BL21 as the host cell) to detect protein expression. The soluble proteins in cell-free extracts were separated by 10 % sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) to determine the expression level of Gr5 aroA protein. One major band appeared at around 45 kDa, and this band corresponded with the putative molecular weight of the protein detected in all induced samples (ESM Fig. 4).

Phylogenetic analysis revealed that the EPSPS and AroAs from different organisms could be grouped into two different classes. Class I AroA proteins consisted of typical AroA proteins from E. coli, Arabidopsis thaliana, Nicotiana tabacum, and Zea mays. EPSP synthase proteins from A. tumefaciens CP4, Bacillus subtilis, and Pseudomonas fluorescens Pf-5 were grouped as class II AroA proteins. Based on the phylogenetic tree, we assigned Gr5aroA to class I (Fig. 1; ESM Fig. 5). Gr5aroA contains three conserved domains (49GSKS52, 196SSQYV200, and 319SQMQ322), which is characteristic of class I EPSPS enzymes.

Fig. 1
figure 1

A neighbor-joining (NJ) tree was constructed based on AroA amino acids sequences using the p-distance model. The complete deletion option was selected for the gaps or missing data. This tree is unrooted and bootstrapped 1,000 times

Full size image

In order to identify the residues implicated in the insensitivity of the Gr5aroA protein, we aligned Gr5aroA protein with other class I and II proteins from 14 different accessions. Among these reported residues (T42M, G96A, P101S, G101A/G137D, G101A/P158S, T97I/P101S) whose mutations could increase glyphosate tolerance, we identified the 101st amino acid of Gr5aroA protein to be phe (Padgette et al. 1991; Shuttleworth and Evans 1994; Eschenburg et al. 2002; He et al. 2003; Priestman et al. 2005a, b; Funke et al. 2009;). Gr5aroA protein was also found to have the residues Glu-351 and Arg-354, which are involved in PEP and S3P binding for CP4 EPSP synthase (SchÖnbrunn et al. 2001). Gr5aroA protein contains the strictly conserved EPSP synthase residues Glu-351 and Arg-354 in Gr5 aroA , which are involved in PEP glyphosate binding. Taken together, these results indicated that the Gr5 aroA gene is an EPSPS gene and the result of extreme glyphosate selection.

Kinetic properties of Gr5 aroA protein

Protein extracts prepared from E. coli BL21 cells containing the plasmids harboring Gr5 aroA were used to analyze Gr5 aroA enzyme activity. The E. coli EPSP synthase from the mutant AB2829 host cells was used as the control. The soluble proteins in cell-free extracts were separated by 10 % SDS-PAGE to determine the expression level of the Gr5 aroA protein. One major band of protein was detected at about 45 kDa in all induced samples that corresponded with the molecular weight of EPSPS (ESM Fig. 4) (Lewendon and Coggins 1983). Kinetic analysis revealed the half maximal inhibitory concentration (IC50) of Gr5 aroA protein to be 24.06 mM, and its Km (PEP) and Ki (glyphosate) values were 220 and 217 μM, respectively. Relative to E. coli AroA, Gr5 aroA protein showed a 360-fold higher Ki (glyphosate) value. Its catalytic efficiency relative to PEP utilization was increased by about threefold (Table 1). These results showed that Gr5aroA had a high tolerance to glyphosate and PEP catalytic efficiency.

Table 1 Kinetic constants of Gr5aroA and Escherichia coli EPSPS
Full size table

Generating transgenic Gr5 aroA tobacco plants

To determine whether the Gr5 aroA gene induced glyphosate tolerance in plants, we constructed an expressing vector, pBI121-CaMV35S::Gr5 aroA ::Nos, and used A. tumefaciens-mediated transformation to generate transgenic Gr5 aroA tobacco plants. Tobacco leaf discs were transformed with A. tumefaciens strain EHA 105 containing pBI-Gr5 aroA (ESM Fig. 2) using the method described by Horsch et al. (1988). Following two rounds of kanamycin (50 mg/l) selection, 152 kanamycin-resistant plants were regenerated. Forty-eight plants showed amplification of the predicted fragment of Gr5 aroA , while no amplification was observed in the control plants (data not shown).

The results of Southern hybridization showed that each plant contained one or more Gr5 aroA -specific hybridizing bands, and the unique hybridization patterns observed indicated that each plant was derived from an independent transformation event (representative sample shown in Fig. 2a). The copy number of Gr5 aroA in the independent transgenic plants ranged from one to more than five.

Fig. 2
figure 2

Verification of transgenic Gr5aroA gene tobacco plants. a Representative Southern blot analysis for the presence of the 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) gene (Gr5aroA) gene in transgenic tobacco plants. Plant genomic DNA (8 μg) and pBI Gr5aroA were digested with XbaI, separated by 1.0 % agarose gel electrophoresis, and hybridized to the 32P-labeled Gr5aroA probe. Lanes 1 Gr5aroA plasmid (positive control), 2 untransformed plant (negative control), 3–11 independent transgenic plants corresponding to lines No. 1, 2, 3, 7, 12, 13, 15, 18, and 29, respectively. b Representative real time-PCR analysis of the expression of Gr5aroA in transgenic tobacco plants

Full size image

The expression of Gr5 aroA in independent transgenic lines was also detected by qRT-PCR analysis. The expression level varied across different lines, with the maximum expression level being about threefold higher than the lowest expression level (Fig. 2b).

Glyphosate tolerance in transgenic Gr5 aroA plants

No significant differences were observed in the growth processes of WT and transgenic plants; therefore, the seedlings of both transgenic and WT plants were treated with six different concentrations of glyphosate salt solution. When treated with 0.5× the recommended dose of herbicide, WT plants maintained the same growth rate as transgenic plants even though their leaves showed some chlorosis. With treatment of 1× the recommended herbicide treatment, the leaves of the WT plants wilted after 7 days and no WT plant survived >14 days after treatment. In order to determine the highest concentration of glyphosate that transgenic plants could tolerate, we then sprayed 2×, 3×, 4×, and 5× the recommend dose of herbicide onto the plants. WT tobacco plants showed severe vegetative damage and died within 1 week. In contrast, transgenic tobacco plants showed no significant differences in newly emerged leaves, meristem tissues, or seed set characteristics following the 2× or 3× treatment, respectively. With the 4× the recommended dose of herbicide treatment, however, all transgenic plants showed a slight chlorosis on newly emerged leaves and their flowering times were delayed by about 2–3 days without stunting. Finally, when transgenic plants were sprayed with 5× the recommended dose of herbicide, growth was inhibited and severely stunted and no seeds were produced (Fig. 3; Table 2; ESM Table 1). In summary, the maximum glyphosate concentration endured by transgenic Gr5 aroA plants was fourfold the recommended field dose. At this level, transgenic plants could still grow, flower, and produce seeds.

Fig. 3
figure 3

Glyphosate tolerance assay of tobacco control and transgenic lines and their T1 progenies. a Left wild-type (WT) control plant following treatment with 0.5× recommended herbicide dose, right transgenic plant No. 18 following treatment with 0.5× recommended herbicide dose. b Left WT control plant following treatment with 1× recommended herbicide dose, right representative transgenic T1 plant (No. 18) following treatment with 1× recommended herbicide dose. c Left WT control plant following treatment with 2× recommended herbicide dose, right representative transgenic T1 plant (No. 18) following treatment with 2× herbicide dose. d Left WT control plant following treatment with 3× recommended herbicide dose, right representative transgenic T1 plant (No. 18) following treatment with 3× recommended herbicide dose. e Left WT control plant following treatment with 4× recommended herbicide dose, right representative transgenic T1 plant (No. 18) following treatment with 4× recommended herbicide treatment. f Left Representative transgenic T1 plant (No. 18) following treatment with 5× recommended herbicide treatment, right WT control plant following treatment with 5× recommended herbicide dose

Full size image
Table 2 Tolerance of transgenic Gr5 aroA tobacco T 0 lines and wild-type line to different levels of glyphosate
Full size table

Gr5 aroA glyphosate tolerance could be inherited in T 1 progenies

To prove that Gr5 aroA glyphosate tolerance could be inherited in T 1 progeny, Gr5 aroA gene inheritance was confirmed by PCR, gene expression, and glyphosate tolerance testing. The PCR analysis showed that the transgenes in five of the lines tested (No. 1, 12, 15, 18, and 29) were inherited at a segregation ratio of 3:1 (ESM Table 2). Inheritance of the transgenes was more complicated in the other two lines, indicating that the integration of the transgenes into tobacco genomes of the five lines was Mendelian.

Gene expression analysis was carried out to determine Gr5 aroA expression in the seven transgenic T 1 progenies (No. 1, 3, 12, 13, 15, 18, and 29). The results showed that the transgenic T 1 lines containing Gr5 aroA could express Gr5 aroA at levels similar to their T 0 parental lines (data not shown).

Because T 1 seeds harvested from different glyphosate-tolerant T 0 plants were segregated, positive and negative seedlings were first confirmed by PCR amplification and then transferred into soil in the greenhouse for the glyphosate tolerance assay. The results of this assay showed that transgenic T 1 progenies and T 0 plants had the same level of glyphosate tolerance.

Under normal conditions (no glyphosate treatment), there was no significant difference in growth between WT and T 1 plants. Both WT and transgenic T 1 plants survived the application of 0.5× the recommended herbicide treatment. With the application of higher levels of herbicide (≥1×), the leaves of both WT and T 1 non-transgenic plants turned yellow and wilted quickly, and all plants in both of these groups died within 5–10 days. In contrast, T 1 transgenic plants grew normally after treatment with less than 4× the recommended herbicide dose, and no significant differences in growth were observed. The agronomic characteristics of transgenic plants treated with 4× the recommended herbicide dose changed, with a delay in the flowering time and the development of a slight chlorosis of newly emerged leaves. No transgenic plant grew normally following treatment with 5× the recommended herbicide dose: The plants were severely stunted and no seeds were produced. These results supported the notion that Gr5 aroA expression in tobacco plants enhances their glyphosate tolerance and that herbicide resistance is strictly conserved in their offspring.

Discussion

Expression of the EPSPS gene has been found to cause glyphosate tolerance in crop plants, enabling more effective weed control via the herbicide (Franz et al. 1997). Unfortunately, the transgene in most commercial glyphosate-resistant crops is the EPSPS gene derived from Agrobacterium spp. CP4 (Padgette et al. 1995a, b). This single source of the EPSPS gene is probably what has caused the decrease in herbicide tolerance that has become a concern in field management programs. Improving EPSPS gene diversity and generating multi-herbicide-resistant crops is an effective approach to this problem.

Recent studies have shown that continued glyphosate selection pressure on organisms will favor mutations that reduce glyphosate sensitivity while still maintaining catalytic efficiency in nature. In 2005, Sun et al. (2005) isolated an Aroa p.putida gene from Pseudomonas putida 4G-1 in glyphosate-contaminated soil. The protein for which it coded showed a higher glyphosate tolerance and the same PEP binding activity as that of E. coli. Barry et al. (1997) isolated the widely used EPSPS gene from Agrobacterium sp. strain CP4 from a waste-fed column at a glyphosate production company. Taken together, these results show that there are EPSPS gene mutations which can occur spontaneously and that a large portion of EPSPS gene diversity exists in organisms exposed to highly glyphosate-contaminated environments.

Class I EPSPS mutant enzymes have been found to exhibit a decreased affinity for substrate PEP and an increased tolerance to glyphosate (Padgette et al. 1995a, b). Because glyphosate and PEP bind to the same site, the PEP binding ability of EPSPS enzyme is one of crucial factors that can determine its suitability for use in commercial glyphosate-resistant crops. Gr5aroA protein contains the strictly conserved EPSP synthase residue Glu-351 (corresponding to the E. coli Glu-341 residue), which is involved in stabilizing PEP binding (SchÖnbrunn et al. 2001). The Gr5aroA protein has also been shown to have a P101F mutation. Several studies have shown that P101 mutations can directly improve PEP binding ability and glyphosate tolerance (Comai et al. 1985; Stalker et al. 1985; Pilacinski 2002; Healy-Fried et al. 2007; Funke et al. 2009). Enzyme kinetic analysis has revealed that the glyphosate tolerance and PEP binding activity of Gr5 aroA protein are about 360- and threefold higher than that of E. coli AroA. The glyphosate tolerance of Gr5aroA protein is fourfold higher than that of the AroA from P. fluorescens (Zhou et al. 2012). It is therefore reasonable to deduce that both of these two amino acid residues contribute to the glyphosate tolerance and PEP binding activity of the Gr5 aroA protein.

Our analysis of transgenic tobacco plants showed that Gr5aroA protein can affect the plant genome. The results of our PCR and Southern blot analyses of transgenic Gr5 aroA plants confirmed that the Gr5aroA gene was successfully inserted into the tobacco genome and highly expressed (Linn et al. 1990; Tang et al. 1999; Wang et al. 2004). Segregation analysis of T 1 progeny demonstrated that the Gr5 aroA gene was inherited by Mendelian rules in most T 1 individuals. The expression level of Gr5 aroA in the T 1 progeny was unchanged relative to the T 0 parental lines (Bano-Maqbool and Christou 1999; Tang et al. 1999; Wang et al. 2004).

Gr5 aroA expression can significantly improve glyphosate tolerance in plants. In our study, following application of the recommended dose of herbicide, the leaves of both the WT plants and T 1 non-transgenic tobacco plants died within 5–10 days. In contrast, after being sprayed with the same dose, transgenic plants showed no significant effects of the herbicide. Transgenic plants can grow normally without agronomic characteristics under even higher concentrations of herbicide. In our study, the highest concentration that the transgenic plants could tolerate was 4× the recommended dose. The glyphosate treatment assay was repeated in T 1 plants with similar results. These results indicate that glyphosate tolerance in our transgenic Gr5 aroA plants was much higher than the dose recommended (1×) for commercial crops. As such, they clearly demonstrate that there is a sufficient safety margin with regard to glyphosate spraying for Gr5 aroA engineered crops to be used commercially in fields that rely on glyphosate as an herbicide.

In conclusion, we report the identification of the Gr5 aroA gene from glyphosate-contaminated soil and confirm it as a promising candidate for the development of transgenic crops in the future.