Developing dual herbicide tolerant transgenic rice plants for sustainable weed management

Herbicides are important constituents of modern integrated weed management system. However, the continuous use of a single herbicide leads to the frequent evolution of resistant weeds which further challenges their management. To overcome this situation, alternating use of multiple herbicides along with conventional weed-management practices is suitable and recommended. The development of multiple herbicide-tolerant crops is still in its infancy, and only a few crops with herbicide tolerance traits have been reported and commercialized. In this study, we developed transgenic rice plants that were tolerant to both bensulfuron methyl (BM) and glufosinate herbicides. The herbicide tolerant mutant variant of rice AHAS (Acetohydroxyacid synthase) was overexpressed along with codon optimized bacterial bar gene. The developed transgenic lines showed significant tolerance to both herbicides at various stages of plant development. The selected transgenic lines displayed an increased tolerance against 100 μM BM and 30 mg/L phosphinothricin during seed germination stage. Foliar applications further confirmed the dual tolerance to 300 μM BM and 2% basta herbicides without any significant growth and yield penalties. The development of dual-herbicide-tolerant transgenic plants adds further information to the knowledge of crop herbicide tolerance for sustainable weed management in modern agricultural system.

The presence of weeds is very serious problem in any agricultural system because they aggressively compete with crops for nutrients, light and other important resources. Their infestation imposes a severe challenge to crop yield, productivity and survival of plants. Among various conventional and modern strategies, chemical herbicides are the least expensive, easiest and most effective way to combat weed invasion [1][2][3] . Advancements in genetically modified herbicide resistance technology, which began with the introduction of the first glyphosate-resistant soybean variety in 1996, opened a new way to manage weed populations in crop fields. Since then, many important genetically modified crops that are tolerant to various herbicides have been developed and commercialized 4,5 . Most of the transgenic crops are transformed with single herbicide-tolerance genes against glyphosate, basta, HPPD, PPO or AHAS-inhibiting herbicides. With the rise in global demand for food and other agricultural products, herbicide technology has contributed greatly during the past few decades. However, the continuous overuse of a single herbicide multiple times in a growing season increases the potential risk of evolution of resistant weeds which has become a major concern in agriculture worldwide. Further, higher doses of herbicide applications cause major threats to agricultural land and the environment 6,7 . Many weeds which are resistant to various herbicides have been identified in different parts of the world 1,[7][8][9] . According to the 'International Survey of Herbicide Resistant Weeds' , 486 unique cases of herbicide-resistant weed biotypes have been reported globally, of which 147 are in dicot and 106 are in monocot groups. A total of 92 crops have been reported to be infested by herbicide-resistant weeds. Glyphosate, basta, AHAS inhibitor, HPPD inhibitor and PPO inhibitor herbicides are widely used to kill or suppress weeds in crop fields. Among these herbicides, the highest 159 identified unique cases of weed resistance were of AHAS-inhibiting herbicides; in addition, there have been 39 cases of resistance identified against glyphosate, 3 against basta, 2 against HPPD-inhibitors and 13 against PPO-inhibiting herbicides, and the number is increasing daily 10 .

Results
In silico analysis and identification of P197S tolerant mutation in rice AHAS. A multiple sequence alignment involving the AHAS enzyme of different plants resulted in a consecutive 35 amino acid long conserved stretch from 190-224 amino acid positions in all AHAS enzymes (Fig. 1A). The amino acid proline was highly conserved in all the tested plant species at 197 with reference to Arabidopsis; this residue was located at position 171 in the rice AHAS protein. There have been many AHAS-inhibiting herbicide-resistant weeds with the P197S substitution mutation reported to exist in nature 10 and the multiple sequence alignment revealed the conserved position of this amino acid in the protein sequence, confirming the crucial role of proline in AHAS enzyme function. Thus, a naturally selected tolerance mutation in the AHAS gene could be a valuable resource for the development of herbicide-tolerant transgenic plants.

Development and molecular confirmation of transgenic plants.
The in silico analysis revealed that the amino acid at position 171 in the rice AHAS protein is the conserved mutation point for herbicide tolerance; this amino acid was subsequently replaced by serine to develop herbicide-tolerant plants (Fig. 1A). While creating the P171S mutation in rice AHAS, the XbaI restriction site was introduced to help to distinguish transgenic plants from non-transgenic plant (Fig. 2E). We used the Agrobacterium-mediated transformation method for the generation of transgenic plants. The putative transgenic plants were initially screened for confirmation of the presence of transgene construct with the help of bar primers by PCR ( Fig. SI1A and B). Since AHAS is a native gene, it was not possible to screen putative transgenic plants with AHAS primers because there would be no difference between native and transgenic PCR bands.
The PCR positive T1 transgenic plants were analyzed to confirm the presence of the transgene integration by southern blotting. The hybridization of nylon membrane with AHAS DIG labelled probe resulted in a common band for all the transgenic lines (including wild type (wt) control plant) ( Fig. 2A); this band represents the native AHAS gene. In addition to native copy, there was an additional band found for all 5 transgenic lines (line 1, 2, 4, 5, 6), which showed the single copy AHAS transgene integration. Subsequently, the same nylon membrane was re-hybridized with bar DIG probe. The bar probe resulted in the same pattern as observed in AHAS hybridization except the common AHAS band due to the heterologous origin of bar gene (Fig. 2B) plants (lines 1, 2, 4, 5 and 6), lines 4 and 6 showed similar band patterns for both the AHAS and bar genes; thus considered as the same transgenic events. Further, transgenic line 3 and the wt control did not show signals for any of the genes; this line was considered non-transgenic. Thus, the southern blot analysis confirmed a total of 4 different positive AHAS and bar single copy integrated independent transgenic events.
The expression levels of the AHAS and bar transgenes in all 4 southern-positive transgenic lines were further validated by northern blotting and semi-quantitative RT-PCR. Northern hybridization was performed consecutively with AHAS and bar DIG probes. Initially, the membrane was hybridized with AHAS DIG probe which resulted in an ~2 kb band for all the plant samples; however, the hybridization signal was found to be higher in all the transgenic lines compared to the wt controls (Fig. 2C). Furthermore, the membrane was reprobed with bar DIG probe which resulted in an ~500 bp hybridization signal for all transgenic lines; this band was absent in the wt control. The experiment confirmed the expression of both transgenes. Further, the expression of the ActI constitutive gene was used as an internal standard which showed equal expression in all plant samples (Fig. 2C). Similarly, the semi-quantitative RT-PCR experiment showed a higher expression (almost 2-fold) of the AHAS gene in the transgenic lines compared to the wt control plants due to the active transcription of the AHAS transgene copy (Fig. 2D). All four transgenic lines showed expression of the bar gene in semi quantitative RT-PCR which was absent in wt control plants. The expression of the ActI gene was used as the internal standard to confirm equal quantities of cDNA used in all the experiments which showed its uniform expression in all plant samples (Fig. 2D).
The restriction digestion analysis of semi quantitative RT-PCR product of the AHAS gene from all the transgenic and wt control plants further helped to distinguish between the expression of OsmAHAS and OsAHAS. The internal AHAS gene primers were designed to amplify a region flanking the P171S (XbaI) mutation site. A 401 bp PCR band was amplified from all the transgenic and wt plants using cDNA and was digested with XbaI restriction enzyme. The PCR product of wt plant could not digested with XbaI restriction enzyme while all the other transgenic lines showed three distinct bands with 107, 294 and 401 bp DNA products. The PCR-based strategy was used as a molecular marker to estimate the relative expression of the native AHAS versus trans-AHAS genes in all transgenic lines (Fig. 2E).
Transgenic plants showed an increased tolerance to BM. The tolerance mutations of AHAS exhibit different degree of resistance to various classes of AHAS-inhibiting herbicides. Among all the tolerance mutations, the amino acid substitution at position P171 in the AHAS protein is frequently observed in many tolerant weed biotypes imparting a higher degree of tolerance to sulfonylureas (SU) and relatively lower level of tolerance to pyrimidinylcarboxylates (PC) herbicides 22 . To validate the level of tolerance offered by the introduced P171S mutation in rice AHAS, the transgenic rice seeds were screened against two popular AHAS-inhibiting herbicides BM (belongs to SU) and bispyribac sodium (BS; belongs to PC) at the germination stage. Initially we checked the natural tolerance of control rice seeds against various concentrations of BM and BS. The results showed a complete inhibition of root growth at 25 μM concentration for BM and 20 μM for BS. There was no significant difference in shoot growth observed (data not shown). The surface sterilized T3 homozygous seeds were inoculated on half-strength MS media containing 100 μM concentration of BS or BM herbicides (Figs 3 and 4). The wt seeds treated with and without herbicides were used as positive (+) and negative (−) controls respectively, and handled alike. In BM supplemented growth media, all the transgenic lines showed vigorous root growth which was comparable to that of the wt (−) control, as recorded after 7, 14 and 21 days of seed germination (Fig. 3). Further, the wt (+) control showed poor root growth even after 14 days; however, no significant difference in shoot lengths was observed which was similar to the wt (−) control. The root growth of the wt (+) control plants was completely inhibited after the 7 th day, and no subsequent growth was observed (Fig. 3A, 3B).
To compare the degree of tolerance offered by the transgenic lines to sulfonylureas (SU) and pyrimidinylcarboxylates (PC) class of herbicides, the T3 transgenic lines (1 and 2) were grown in half strength MS media supplemented with 100 μM BM and BS separately. The root growth of these transgenic lines in the BM supplemented media showed normal morphology, which was comparable to that of wt (−) controls (Fig. 4). However, the root growth was significantly arrested in both the transgenic lines in the BS media which was comparable to that of wt (+) plants (supplemented with 100 μM BS). Although, the transgenic plants showed a moderate level of tolerance to 50 μM BS at germination stage (data not shown). Furthermore, we did not observe any significant shoot growth differences among the transgenic lines, including the wt (−) and wt (+) controls, even at 21 days after seed germination ( Fig. 4A-C).   The tolerant transgenic lines (1 and 2), which were confirmed by seed germination tests, were further examined to determine the effect of BM herbicide on the overexpressed mutant OsmAHAS gene in transgenic plants via color tests. Transgenic lines 1 and 2 developed a light red color due to the accumulation of acetoin resulting from the enzymatic action of mutant OsmAHAS proteins in the presence of 0.1 µM BM. The color of the wt (−) control reaction was dark red because of the uninhibited accumulation of high acetoin due to the lack of BM. However, the color of the wt (+) control reaction remained yellowish due to the presence of the inhibitor BM and the absence of mutant OsmAHAS. The native OsAHAS protein is unable to produce acetolactate due to the inhibitory action of BM, thus leading to no acetoin formation and the color of the reaction remained yellow (Fig.  SI2A). The results showed active expression of the mutant OsmAHAS gene.
Confirmation of transgenic lines tolerance to phosphinothricin herbicide. Chlorophenol red is a color indicator dye that appears red at pH 6 and turns yellow with a decrease in pH. To validate the activity of the bar gene in transgenic plants, seeds from the wt control and transgenic groups were germinated on half strength MS media supplemented with 30 mg/L phosphinothricin herbicide. We observed a gradual change in media color from red to yellow after 10 days of seed inoculation in all transgenic lines, indicating active expression and function of the bar gene. The bar protein detoxified phosphinothricin rapidly which resulted in healthy growth of all transgenic plants (Fig. 5A-D); this result was comparable to that of the wt (−) controls. On the other hand, the wt (+) control plants were unable to detoxify phosphinothricin due to the absence of the bar, and consequently the media color remained unchanged (Fig. 5A,B). Unlike BM and BS herbicide, phosphinothricin has equal effects on both root and shoot growth (Fig. 5E). The root and shoot development was severely arrested in wt (+) plants while the transgenic plants showed normal growth.
We tested the effective tolerance of transgenic plants against commercially available phosphinothricin (marketed as 'Basta') during the vegetative stage. Basta (in which phosphinothricin is the active ingredient) is a contact herbicide that has a localized toxic effect on plant tissue. Phosphinothricin irreversibly binds to glutamine synthetase (GS), an important enzyme in the ammonia detoxification mechanism. The transgenic lines as well as control plants were painted (on leaf tip region) with 3% (v/v) basta herbicide at the 8-12 leaf stage. The localized toxic effect of basta killed the control plant leaves within 10 days of application due to the accumulation of ammonia, however, the transgenic lines (except line 4) did not show any leaf necrosis due to the detoxification of phosphinothricin by bar (Fig. SI2B). Phosphinothricin irreversibly binds to glutamine synthetase (GS), an important enzyme in the ammonia detoxification mechanism, and results in plant death.

Characterization of transgenic plants under the foliar application of BM and basta herbicides.
The T3 homozygous transgenic lines as well as two wt control plants (one month old) were analyzed for their field level tolerance against the herbicides BM and basta. The first set of transgenic lines and wt controls were foliar sprayed with 300 μM BM (Fig. 6A, 6B). Necrotic symptoms started on wt (+) control plants after 4-5 days of BM application, and the leaves turned yellow and completely died after 15-20 days. However, all four transgenic lines showed healthy growth without any necrotic symptoms in response to the same BM application; their leaves stayed green which was comparable to the wt (−) control plants (without BM application) (Fig. 6A, 6B). There was no significant difference between the transgenic and wt (−) plants in terms of growth and morphology (Fig. 6C). The measurements of the total fresh weight of the herbicide treated transgenic lines did not show any growth penalty which suggests that the introduced P171S mutation added an advantage to the transgenic plants in providing tolerance to BM herbicide.
Similarly, a second set of plants (transgenic lines and wt control) were sprayed with 2% (v/v) commercially available basta solution. The transgenic lines showed normal morphology with healthy leaves without any leaf necrosis similar to wt (−) plants. Furthermore, the total fresh biomass was analyzed which did not show any noticeable differences between the transgenic and wt (−) control plants, indicating no morphological penalty on the transgenic plants due to the active expression of bar (Fig. 7C). However, the wt (+) plants started showing necrotic symptoms after 6-7 days and completely died within 10-15 days (Fig. 7A, 7B).
We also examined the effects of dual applications of the herbicides basta and BM on the transgenic plants. A set of transgenic lines and wt control plants were initially sprayed with a 2% (v/v) basta solution. At 7 days of basta application, the same set of plants was similarly sprayed with 300 μM BM. There was no phenological injury observed in the transgenic lines which remained green and healthy even at 20 days post-application (Fig. 8A). Further, we allowed the transgenic plants to grown until the maturity stage to determine the effects of dual applications of herbicides on yield. The transgenic plants did not show any yield penalty having normal seed setting and seed number per panicle which was similar to the wt (−) control plants (Fig. 8B, 8C).

Discussion
To overcome the problem of frequent evolution of herbicide-resistant weeds, the alternating or combined use of multiple herbicides in conjunction with traditional practices is always suggested. The alternating or combining of herbicides provides the least chance for a weed species to acquire resistance against respective herbicides in such a short span of time 23 . Thus, developing transgenic plants that have more than one herbicide tolerance could be a great solution to overcome this problem. However, only a few genetically modified crop plants that are tolerant to multiple herbicides have been reported in the public and private sectors. There is only a single report from the public sector, to the best of our knowledge, mentioning the development of PGMS (photoperiod-sensitive genic male-sterile) transgenic rice which was engineered with dual herbicide tolerance traits against the glyphosate and glufosinate herbicides 24 . A double herbicide tolerant soybean event 356043 that contained a mutated soybean gm-hra (Glycine max herbicide-resistant acetolactate synthase) and glyphosate resistance gene was developed by Pioneer Hi-Bred, Johnston, IA which was deregulated in 2008 by the United States Department of Agriculture  IN) has also planned to commercialize transgenic corn, cotton and soybean that exhibit tolerance to glyphosate along with 2, 4-dichlorophenoxyaceticacid (2, 4-D) (Enlist TM ), with both 2,4-D and glufosinate (Enlist E3 TM ) and together with aryloxyphenoxypropionates herbicides. In addition, Syngenta and Bayer Crop Science have in their pipeline to develop dual herbicide tolerant soybean against HPPDi and glufosinate 4,11,25,26 .
Here, we reported the development of dual herbicide tolerant transgenic indica rice (cultivar Swarna) against BM and basta herbicides for sustainable weed management. There have been 22 amino acid substitutions at seven different sites with in the AHAS reported in many tolerant weeds. The common mutation point P197 (equivalent to P171 in rice) is quite vulnerable to 11 different amino acid substitution mutations that provide resistance against several AHAS-inhibiting herbicide compounds and occurs very frequently in AHAS-inhibiting herbicide resistant biotypes 1 . Previously, transgenic plants generated by overexpressing AHAS with P197S mutation exhibited a high degree of tolerance against herbicide bensulfuron methyl (BM) and moderate resistance against bispyribac sodium (BS) 27 . Similar to the above finding, in present study, the transgenic lines showed tolerance to 100 µM BM and exhibited normal phenotype; while a complete root growth inhibition was observed in these plants in 100 µM BS supplemented media (Figs 3 and 4). Odell and colleagues reported that the mutated form of AtAHAS overexpressed in transgenic plants showed a 300-fold increased tolerance against chlorsulfuron while the wild type AtAHAS gene conferred only 3-fold tolerance to those plants 28 . Similarly, Kawai and coworkers compared nine different kinds of P171 rice AHAS mutations for their sensitivities against sulfonylurea herbicides. The authors confirmed the high resistance of the P171S mutation against BM herbicide with an RS ratio greater than 8000 29 . Yu and colleagues conducted a similar study in which they compared and functionally validated the pleotropic effects of four different P197 substitution mutations of AHAS in Lolium rigidum. Among the mutations, P197S showed no effects on AHAS kinetics and subsequently had no penalty on overall plant growth. The study further highlighted the preference for a serine substitution for proline in AHAS over any other mutation in nature 22 . Mutant variants of AHAS against BM and related herbicide derivatives have been well documented in many crops such as rice 27,[30][31][32][33] , soybean 34 , sugarcane 35 and wheat 36   Basta is a contact herbicide that irreversibly inhibits the glutamine synthetase (GS) enzyme resulting in the accumulation of ammonia in the cell and subsequent plant death. The bacterial bar gene is responsible for the detoxification of phosphinothricin herbicide by acetylating its amino group and helps in the survival of the plants [38][39][40][41] . This gene has been commonly used in many herbicide-tolerant transgenic crops, including rice 24 , tobacco 20 , soybean 42 , maize 43 and cotton 44 .
The AHAS and bar enzymes have different mechanisms of action against their respective herbicides; these mechanisms are very effective. We developed four independent dual herbicide-tolerant transgenic indica rice (Swarna) lines against BM and phosphinothricin (basta) herbicides. The molecular analysis confirmed the expression of both transgenes in all transgenic plant lines. The seed germination analysis of wt control rice plants revealed its natural tolerance to BM up to 25 µM and tolerance to BS up to 20 µM concentrations which is relatively high compared to other indica and japonica rice varieties. Previously, two indica, five indica-derived and one japonica rice variety were checked for their tolerance limit against BS herbicide 30 . With tolerance limit up to 2.5 µM BS concentration (moderately inhibited) at the seedling stage, the wt indica rice cultivar Kasalath was the most tolerant among the selected cultivars, although its roots were completely inhibited at 0.5 µM. However, all the other cultivars showed high susceptibility to 2.5 µM concentration. There are no previous reports regarding the effects of BM herbicide on the seed germination of rice to the best of our knowledge. AHAS color test on crude enzyme extracts showed functional activity of OsmAHAS trans-protein in the transgenic lines in the presence of 0.1 µM BM. Further, the transgenic lines were able to tolerate 300 µM BM spraying at the 8-12 leaf stage. Shimizu and colleagues transformed tobacco plants with a mutated Arabidopsis P197S AHAS gene and generated transplastomic plants. These plants were able to show functional activity of trans-AHAS proteins in the presence of 0.1 µM BM herbicide in color tests, and the excised tobacco leaves were also able to survive on MS media that were supplemented with the same concentration of BM herbicide 45 . A mutant P169S Monochoria vaginalis AHAS gene was used by Song and colleagues to transform Arabidopsis. The transgenic plants were able to tolerate a high concentration of BM herbicide, as confirmed by hypocotyl length estimations and color test 46 .
Our transgenic lines were also able to detoxify 30 mg/L of phosphinothricin herbicide as recorded at the seed germination stage. In addition, the generated transgenic plants showed high tolerance to 2% and 3% basta herbicide as confirmed by foliar spraying and leaf paint assay respectively. The application of basta to whole plants at the vegetative stage showed no visible injury and fitness costs to the transgenic plants. Similar results have been reported in previous studies involving the bar in different crop plants 20,42,47,48 . This investigation highlights the use of effective simultaneous dual herbicide tolerance of transgenic rice plants without causing any yield penalty. However, additional experiments are needed to analyze the tolerance against both herbicides under natural field conditions.

Conclusion
We have successfully developed dual herbicide-tolerant transgenic rice plants which were confirmed at various plant developmental stages i.e. at the seed germination, mature leaves and seedling stages; the transgenic plants showed a higher level of tolerance against BM and basta herbicides. Further, the transgenic plants performed well without any yield penalty. This study highlights the importance of the P171S mutation for imparting herbicide tolerance and the development of dual herbicide-tolerant transgenic plants which will add additional knowledge to weed management practices of rice cultivation.

Methods
In silico analysis of AHAS-inhibiting herbicide tolerant P to S mutation in rice. A  Preparation of gene construct and generation of transgenic plants. The final gene construct with mutated AHAS and bar genes was cloned into the pMDC99 Gateway compatible cloning vector and transformed into Agrobacterium strain EHA 105, which was further used for transformation in rice calli (refer to supplementary information for details) (Fig. 1B, 1C).

Molecular and expression analyses of putative transgenic plants.
The tissue culture generated putative transgenic plants were confirmed by PCR and southern blotting, and the expression analysis of the confirmed transgenic plants was performed by northern blotting and semi-quantitative RT-PCR. The presence and expression of the mutaned AHAS gene was also confirmed by restriction digestion analysis (refer to supplementary information for details).
Analysis of herbicide tolerance during the seed germination stage. To verify the level of tolerance offered by the overexpression of OsmAHAS, the natural tolerance of control Swarna plants against BS and BM herbicides were analyzed by growing the sterilized seeds on half-strength MS media that contained the various concentrations of herbicides (BS and BM). The concentrations of herbicides used were 0, 10, 20, 30, 40, 50 and 100 μM. The seeds were surface sterilized and grown for 15-20 days to regularly monitor the effects of the herbicides on plant growth. The tolerance of the transgenic plants at the seed germination stage against these herbicides was analyzed by growing the T3 seeds in half-strength MS media supplemented with 100 μM BS or BM herbicides separately. In both experiments, the positive and negative wt controls were also maintained along with the transgenic lines.
To assess the activity of the bar at the seed germination stage, the sterilized seeds of all four transgenic lines, in addition to the positive and negative wt control lines, were inoculated on half-strength MS media (pH 6.0) with 50 mg/L of chlorophenol red (CR) and 3 mg/L phosphinothricin herbicide. The wt seeds inoculated on the media with phosphinothricin served as positive controls, while the wt seeds germinated on media without phosphinothricin were used as negative controls.
Characterization of transgenic plants during the vegetative stage. The toxic level of BM herbicide was initially assessed by foliar application of various concentrations of BM including 50, 100, 150, 200, 250 and 300 μM on one month-old wt control plants, and the lethal dose was chosen based on the appearance of necrotic symptoms on the leaves. Two separate sets of all four T3 homozygous transgenic lines along with wt control plants were grown to analyze the field level tolerance against both herbicides i.e. BM and basta. The plant groups were foliar sprayed with respective herbicides and the results were monitored based on the appearance of injury symptoms.
The first set of transgenic plants were treated with 300 μM BM (tissue culture grade) solution using 0.1% Tween-20 as a surfactant and the second set with 2% (v/v) basta herbicide (Bayer Crop Science, Germany). The third set of transgenic plants were sprayed with both herbicides BM and basta at an interval of 7 days. The negative control wt (wt (−)) plants in all experiments were sprayed with water along with 0.1% Tween-20 while the positive wt (wt (+)) plants were sprayed with the same concentrations of the respective herbicides. The results were monitored regularly at 7-days intervals in each experiment.
For these experiments, a total of 20 plants were taken from a particular group and subjected to herbicide treatment. The herbicide-treated plants were analyzed and compared for any phenological injury or impact of transgene integration on the grain yield between transgenic plants and wt (−) plants.
Leaf paint assays and AHAS color tests. The leaf paint assay and AHAS color tests were performed on mature leaves of transgenic plants to confirm the activity of bar and AHAS proteins respectively (refer to supplementary information of full detail).