Reduced Translocation of Glyphosate and Dicamba in Combination Contributes to Poor Control of Kochia scoparia: Evidence of Herbicide Antagonism

Kochia scoparia is a troublesome weed across the Great Plains of North America. Glyphosate and dicamba have been used for decades to control K. scoparia. Due to extensive selection, glyphosate- and dicamba-resistant (GDR) K. scoparia have evolved in the USA. Herbicide mixtures are routinely used to improve weed control. Herbicide interactions if result in an antagonistic effect can significantly affect the management of weeds, such as K. scoparia. To uncover the interaction of glyphosate and dicamba when applied in combination in K. scoparia management the efficacies of different doses of glyphosate plus dicamba were evaluated under greenhouse and field conditions using GDR and a known glyphosate- and dicamba-susceptible (GDS) K. scoparia. The results of greenhouse and field studies suggest that the combination of glyphosate and dicamba application controlled GDS, but glyphosate alone provided a better control of GDR K. scoparia compared to glyphosate plus dicamba combinations. Furthermore, investigation of the basis of this response suggested glyphosate and dicamba interact antagonistically and consequently, the translocation of both herbicides was significantly reduced resulting in poor control of K. scoparia. Therefore, a combination of glyphosate plus dicamba may not be a viable option to control GDR K. scoparia.

. Estimated values of ED 50 and GR 50 of glyphosate and dicamba in K. scoparia using the nonlinear regression analysis of four parameter log-logistic model*. * model: Y = C + (D−C)/ (1 + exp[b(log(x) − log(I 50 ))]); ED 50 (effective dose for 50% control of K. scoparia) and GR 50 (effective dose for 50% biomass reduction) values were estimated using the number of plants data and dry biomass data, respectively. # Values in parenthesis are standard error. + resistant level of GDR K. scoparia population comparing to GDS population using ED 50 or GR 50 values.
Scientific REPORTS | (2018) 8:5330 | DOI: 10.1038/s41598-018-23742-3 In general, glyphosate alone without mixing with dicamba showed the best control of both GDR and GDS K. scoparia, compared to the combinations containing the same dose of glyphosate. For example, 840 g ae·ha −1 of glyphosate (Trt 5) had 45 and 95% control of GDR and GDS K. scoparia, respectively. It rendered more control of GDR K. scoparia than glyphosate and dicamba combinations (Trt 6, 7, and 8); and had similar control of Trt 9, which was mixed with 560 g ae·ha −1 of dicamba. Also, the 840 g ae·ha −1 of glyphosate (Trt 5) controlled GDS K. scoparia more effectively than Trt 6 and 7 (Table 2).
When Trt 10 to 14 were compared, 1260 g ae·ha −1 of glyphosate alone (Trt 10) rendered higher or similar control of the combinations that contain 140 to 1400 g ae·ha −1 of dicamba with the same amount of glyphosate. In the case of combinations with 2100 g ae·ha −1 of glyphosate, the results suggest that 2100 g ae·ha −1 of glyphosate (Trt 15) alone controlled the 95% of GDR K. scoparia, which is higher than Trt 16,17, and 19 that were mixed with 140, 280 and 1400 g ae·ha −1 of dicamba, respectively. When 700 g ae·ha −1 of dicamba was mixed with 2100 g ae·ha −1 of glyphosate, the control of K. scoparia was similar to the application of 2100 g ae·ha −1 of glyphosate alone. However, all the combinations containing 2100 g ae·ha −1 of glyphosate rendered similar control of GDS K. scoparia except the Trt 19, which was mixed with 1400 g ae·ha −1 of dicamba and rendered only 91% control of GDS K. scoparia. GDR and GDS K. scoparia response to glyphosate plus dicamba combinations under field conditions. The results of K. scoparia control at 4 weeks after treatment (WAT) with combinations of glyphosate and dicamba are presented in Table 3. Similar to the results obtained under greenhouse conditions, the treatment 2.5 G (2.5 times of glyphosate at label recommended dose) controlled 98% of GDR K. scoparia, which is better than when treated with dicamba alone (e.g. 2.5D, 2.5 times of dicamba at label recommended dose). On the other hand, all the treatments with 2100 g ae·ha −1 of glyphosate, including the 2.5 G, 2.5 G + 1.25D (1.25 times of dicamba at label recommended dose), and 2.5 G + 2.5D, rendered 100% control of GDS K. scoparia. But, the treatment 2.5D that contained 1400 g ae·ha −1 of dicamba only provided 84% GDS K. scoparia control at 4 WAT, which is significantly less than the other treatments.

GDR GDS
------g ae·ha −1 ------------%------   Fig. 1c and d). The difference in translocation of glyphosate was observed at both 72 and 168 HAT in GDR K. scoparia (Fig. 1c), and it was also found at 168 HAT in GDS K. scoparia (Fig. 1d). This indicates less [ 14 C] glyphosate was translocated away from TL when glyphosate was mixed with dicamba in both GDR and GDS K. scoparia ( Fig. 1c and d). Also, less [ 14 C] glyphosate was translocated to plant parts above-treated leaf (ATL) in both GDR (  (Fig. 2b). Translocation data indicate more [ 14 C] dicamba was retained in TL at 168 HAT in GDR K. scoparia ( Fig. 2c) for hot-DG than hot-D, and similar difference was observed in GDS K. scoparia at 24, 72, and 168 HAT (Fig. 2d). The translocation of dicamba to ATL or BTL also confirmed these differences. For instance, when dicamba was mixed with glyphosate (hot-DG), less [ 14 C] dicamba was translocated to ATL at 168 HAT in GDR K. scoparia (Fig. 2e), and at 24 and 168 HAT in GDS K. scoparia (Fig. 2f). Also, less [ 14 C] dicamba translocation to BTL at 168 HAT in both GDR ( Fig. 2g) and GDS K. scoparia ( Fig. 2h) was observed when dicamba and glyphosate were mixed. Furthermore, phosphor image analysis also supported that less [ 14 C] dicamba was translocated to shoots when dicamba was mixed with glyphosate than was applied alone in both GDR (Fig. 3d vs. c) and GDS ( Fig. 3h vs. g) K. scoparia.

Discussion
The dose-response results confirmed that GDR K. scoparia is resistant to both glyphosate and dicamba, whereas GDS K. scoparia is susceptible to both glyphosate and dicamba. Furthermore, the GDR K. scoparia exhibited low level of resistance to glyphosate, whereas, resistance to dicamba was high relative to the GDS K. scoparia. Because of low level of resistance to glyphosate in GDR K. scoparia, increased glyphosate dose provided better control of GDR K. scoparia under both greenhouse and field conditions. In contrast increase in dicamba dose did not provide satisfactory control of GDR K. scoparia in any conditions tested. Growers tend to increase the herbicide dose to achieve maximum weed control. However, our results suggest that increase in herbicide dose may not always provide good weed control, rather increase the selection pressure, which facilitates evolution of resistance. These practices are not sustainable and should not be recommended since they may drive weed populations to evolve a higher level of resistance 22,23 .
Mixing herbicides with different sites/modes of action has been used widely to broaden the spectrum of weed control and delay the development of herbicide resistance 24,25 . In this research both under greenhouse and field conditions, we found that combinations of glyphosate plus dicamba had antagonistic effect on GDR and GDS K. scoparia control. When glyphosate was mixed with dicamba, the GDR K. scoparia control was significantly decreased compared to the same dose of glyphosate applied by itself (Tables 2 and 3). The GDS K. scoparia was controlled using most of the herbicide combinations tested, primarily because high doses of glyphosate and/or dicamba can mask the antagonistic effect of reduced translocation of these herbicides in GDS K. scoparia.  Table 3. The treatments and efficacies (4 WAT) of glyphosate plus dicamba combinations applied on GDR and GDS K. scoparia in field conditions. * 2.5 G, 2.5D, and 1.25D represents 2.5 times of glyphosate at label recommended dose (840 g ae·ha −1 ), 2.5 times of dicamba at label recommended dose (560 g ae·ha −1 ), and 1.25 times of dicamba at label recommended dose for K. scoparia control, respectively. # Means of visual injury (n = 8), and the values in parentheses are standard error (n = 8). The values followed by different letters are significantly (P-value < 0.05) different among the four treatments within each population according to the Bonferroni's multiple comparisons test.
When applied in combination, the absorption of glyphosate and dicamba was enhanced at early hours than treated separately (Figs 1 and 2). Especially glyphosate absorption was increased in both GDR and GDS K. scoparia at 24 HAT ( Fig. 1a and b); after that the difference in absorption was minimal, suggesting that mixing these two herbicides can accelerate absorption of both herbicides immediately after application. Accelerated absorption of dicamba possibly occurred because of the inclusion of adjuvant ammonium sulfate 26 . The rapid absorption of ammonium ions can reduce the apoplastic pH 27 , which can enhance dissociation of the dicamba diglycolamine salt (in Clarity ® formulation of dicamba) to form non-ionized dicamba acid and become more lipophilic. Once dicamba becomes more lipophilic it can be absorbed more quickly via waxy leaf cuticles, which are highly lipophilic 28,29 . However, this process could also increase the volatility of dicamba and upsurge the potential of dicamba drift due to presence of acid form of dicamba 30 , yet not completely absorbed by the plant. On the other hand, glyphosate absorption could have been enhanced by the adjuvants included in Clarity ® formulation, but additional study is needed to test this hypothesis.
Translocation of glyphosate was affected by dicamba regardless of time after application. When glyphosate was mixed with dicamba, less [ 14 C] glyphosate was translocated and more was retained in treated leaves. This could occur as a result of rapid plant response to dicamba. As an auxinic herbicide, dicamba can cause rapid metabolic and physiological reactions within hours after application, which soon can lead to growth inhibition and reduction of transpiration and carbon assimilation 31 . Glyphosate is mainly transported via phloem 32 , which is highly dependent on the source-sink strength 33 . Therefore, due to weakened source upon dicamba application, the translocation of glyphosate may have been restricted compared to when glyphosate was applied alone. On the other side, reduced dicamba translocation was observed only at later time points when applied in combination. Glyphosate inhibits EPSPS enzyme and shuts down the shikimate pathway, which causes aromatic amino acid synthesis failure and stunts the growth of plants, and ultimately lead to plant death 34,35 . Within days, the glyphosate was translocated throughout the plant and shut down the shikimate pathway completely, soon after the carbon assimilation and phloem transport can cease. Therefore, the translocation of dicamba, which is also mainly facilitated by phloem 36,37 , would be significantly affected as a result of glyphosate-induced physiological alterations in plants.
In conclusion, though glyphosate plus dicamba combination is used to control a wide spectrum of monocot and dicot weeds in crops, this combination not necessarily is a good option to manage the stubborn weeds, such as K. scoparia, in North America Great Plains. Our results clearly suggest that glyphosate plus dicamba combination has significant antagonistic effect on both GDR and GDS K. scoparia, as a result of decreased translocation of these two herbicides resulting in reduced efficacy of both the herbicides. Therefore, if K. scoparia is the major issue in the field, glyphosate plus dicamba combination should not be recommended, especially when glyphosate and/or dicamba-resistant K. scoparia is present. Diversification of weed management tactics, such as inclusion of a third mode of action herbicide in the herbicide combination, or other non-chemical management practices such as tillage or cover crops are highly warranted to minimize the further development and spread of herbicide-resistant K. scoparia.

Materials and Methods
In 2012, K. scoparia seed were collected from a field in Haskell County, Kansas (37°29′48.5″N, 100°46′53.0″W). K. scoparia plants generated from these seeds were self-pollinated by keeping the plants in isolation from other K. scoparia plants and upon maturity seed were harvested separately from ten plants. One hundred seedlings were generated separately from seed harvested from above 10 plants. When plants reached 10-12 cm height, 50 plants each were treated with a label recommended field rate of glyphosate (840 g ae·ha −1 ) or dicamba (560 g ae·ha −1 ). In response to glyphosate or dicmaba treatment, all the progeny of a single plant tested that were completely killed, these were selected as glyphosate-and dicamba-susceptible (GDS) K. scoparia. The remaining seed harvested from the same GDS mother plant was used in all experiments in this research. Likewise, all the progeny of single plant tested that survived glyphosate or dicamba treatment, were selected as glyphosate-and dicamba-resistant (GDR) K. scoparia. Also, the rest of the seed harvested from the same GDR mother plant was used in this research.
Greenhouse experiments were conducted in weed science greenhouse attached to the Department of Agronomy at Kansas State University, Manhattan, Kansas, United States. The following greenhouse conditions were maintained: 25/20 °C (day/night, d/n) temperatures, 60 ± 10% relative humidity, and 15/9 h d/n photoperiod supplemented with 120 μmol·m −2 ·s −1 illumination provided with sodium vapor lamps. The physiological studies were conducted in growth chambers maintained at following conditions: 25/15 °C d/n temperature, 60 ± 10% relative humidity, and 15/9 h d/n photoperiod, light was provided by incandescent and fluorescent bulbs delivering 750 µmol·m −2 ·s −1 photon flux at plant canopy level.
Glyphosate-and dicamba-dose response of GDR and GDS K. scoparia. GDR and GDS K. scoparia seeds were germinated in trays (25 × 15 × 2.5 cm) filled with commercial potting mixture (Pro-Mix Potting-Mix, Premier Tech Horticulture, Ontario, CA). Individual seedlings at 6-leaf stage were transplanted into plastic pots (6.5 × 6.5 × 9 cm) containing the same type of soil and kept in the same greenhouse as above. When the K. scoparia seedlings were 10-12 cm height, they were treated with glyphosate (Roundup WeatherMax ® , Monsanto The above treatments were applied as follows. Herbicides were mixed according to the labels and applied using a bench-type sprayer (Research Track Sprayer, De Vries Manufacturing, Hollandale, MN, USA) equipped with a single moving even flat-fan nozzle tip (8002E TeeJet tip, Spraying Systems Co., Wheaton, IL, USA) delivering 187 L·ha −1 at 207 kPa in a single pass at 4.85 km·h −1 . At four weeks after herbicide treatment (WAT), glyphosateand dicamba-induced visual injury was rated based on composite visual estimation of growth inhibition, epinasty (downward curling of plant parts), necrosis, and plant vigor on a scale of 0 (no effect) to 100 (plant death). Plant were clipped off at soil level at 4 WAT and individual plants were placed in separate paper sacks. Dry biomass data was obtained by weighing after oven dried at 60 °C for 72 h.
GDR and GDS K. scoparia response to glyphosate plus dicamba combinations under greenhouse conditions. GDR and GDS K. scoparia seedlings were produced as described above. When plants reached 10-12 cm height in the greenhouse, 19 combinations of low to high doses of glyphosate plus dicamba ( Table 2) were applied (as described above) on both GDR and GDS K. scoparia to test their efficacy. At four WAT, the number of dead plants was recorded. In 2015, GDS and the GDR K. scoparia seeds were germinated in Planters Pride TM plastic greenhouse kit (72 cells, The HC Companies, Middlefield, OH, USA) in the greenhouse. When the seedlings reached 3-4 cm, twenty plants of either GDS or GDR K. scoparia seedlings were transplanted by hand into each field plot of 3 × 3 m. The field was sprinkler irrigated daily. After the seedlings were recovered from transplantation and reached to 10-12 cm height, five treatments including 2100 g ae·ha −1 of glyphosate, 1400 g ae·ha −1 of dicamba, 2100 g ae·ha −1 of glyphosate mixed with 700 g ae·ha −1 of dicamba, 2100 g ae·ha −1 of glyphosate mixed with 1400 g ae·ha −1 of dicamba, and a non-treated control were used and designated as 2.5 G, 2.5D, 2.5 G + 1.25D, 2.5 G + 2.5D, and non-treated, respectively (Table 3), were applied using a CO 2 -pressured backpack sprayer with a 2.74 m boom that was equipped with six TTI110015 tip at 275 kPa with a spray volume of 140 L·ha −1 by walking at 4.8 km·h −1 approximately. Visual injury data (as described above) were collected at 1, 2, 3, and 4 WAT.
In 2016, the experiment was repeated using the same method as described above for in the year 2015, except GDS and GDR K. scoparia seeds were directly planted into the 3 m × 3 m plots, and hand weeding was implemented to remove other weeds.  38 . In the same research, we also found that less than 5% of dicamba and glyphosate translocated to below ground tissue of K. scoparia 38 . Therefore, in this study, the K. scoparia plants were not treated with herbicide formulations prior to application of [ 14 C] labeled dicamba or glyphosate, and also the radioactivity in below ground parts of K. scoparia was not tested.

Absorption and translocation of [
One Radioactive herbicides were applied on GDR and GDS K. scoparia as follows. K. scoparia seedlings were grown in a growth chamber and when plants were 10-12 cm height, two newly expanded leaves were marked.
Ten µL of hot-G or hot-GD solution (5 µL per leaf) was applied using Wiretrol ® (10 µL, Drummond Scientific Co., Broomall, PA, USA). Thirty minutes after herbicide application, plants were returned to the same growth chamber. Plant tissue was harvested at 24, 72 and 168 hours after treatment (HAT) and dissected into treated-leaves (TL), tissue above the treated leaves (ATL), and tissue below the treated leaves (BTL). TL were gently washed twice with 5 mL of 10% (v/v) aqueous ethanol solution with 0.5% of Tween-20 for one minute. Radioactivity in the rinsate was quantified using liquid scintillation spectrometry ( , hot-GD' , and hot-DG' , respectively), were prepared using the same method as described above.
GDR and GDS K. scoparia seeds were germinated in trays filled with the commercial potting mixture as described above. Individual seedlings 2 to 3 cm height were transplanted into plastic pots (6.5 × 6.5 × 9 cm) that filled with silica sand (Granusil ® Handy Sand, Fairmount Santrol, Sugar Land, TX, USA) and rinsed in 1% (w/v) of Miricle-Gro water soluble All Purpose Plant Food (N:P:K = 24:8:16, Scotts Miracle-Gro Products Inc. Marysville, OH, USA) and kept in growth chamber. When the K. scoparia seedlings were 6-8 cm height (10-12 cm height plants were not selected, because the plants were taller to manuplate for phosphor image analysis), they were treated with 1 µL droplet of hot-G' , hot-D' , hot-GD' , and hot-DG' on one newly expanded leaf. At 24, 72, and 168 HAT, K. scoparia plants were gently uprooted, and the roots were washed with water carefully. Then, the whole plant was washed twice with 10 mL of 10% (v/v) ethanol aqueous solution with 0.5% of Tween-20 for 1 minute, and then pressed using a handmade plant press 39 and dried at 60 °C for 72 h. The pressed K. scoparia plants were exposed to BAS-IP MS 2040 E Multipurpose Standard Storage Phosphor Screen (GE Healthcare Life Sciences, Pittsburgh, PA, USA) for 44 h (the hot-G' and hot-GD' treated plants) or 24 h (the hot-D' and hot-DG' treated plants), and the screen was read using Bio-Rad molecular imager FX (Bio-Rad Laboratories, Inc. Hercules, CA, USA).
Experimental design and data analysis. Split plot design was used in the experiment of glyphosate and dicamba dose response on GDR and GDS K. scoparia. K. scoparia population and herbicide dose were mainand subplot, respectively. Treatments were arranged in a factorial combination with GDR and GDS K. scoparia and different herbicide doses. No interaction between experimental runs was observed; hence, data from the repeated experiments were pooled prior to analysis. Then, visual injury and dry biomass data were subjected to non-linear regression analysis using four parameter log-logistic model 40  In Eq. 1, Y refers to the percentage of untreated, C and D are the lower limit and upper limit of the data, respectively, b is the slope, and I 50 is the dose required for 50% response of visual injury or biomass reduction, which was used to estimate ED 50 (effective dose for 50% control of K. scoparia) and GR 50 (effective dose for 50% biomass reduction) values from the visual injury and dry biomass data, respectively.
Split plot experimental design was also used in greenhouse screening experiments and efficacy study of different glyphosate plus dicamba combinations in field conditions. K. scoparia population and rate of herbicide combination were the main-and subplot, respectively. Data from the repeated experiments were pooled prior to analysis due to no interaction between experimental runs was found. Two-way analysis of variance was performed in GraphPad Prism 7 (GraphPad Software, Inc., La Jolla, CA, USA) using Bonferroni's multiple comparisons test (p-value < 0.05).