Increased temperatures and elevated CO2 levels reduce the sensitivity of Conyza canadensis and Chenopodium album to glyphosate

Herbicides are the most commonly used means of controlling weeds. Recently, there has been growing concern over the potential impacts of global climate change, specifically, increasing temperatures and elevated carbon dioxide (CO2) concentrations, on the sensitivity of weeds to herbicides. Here, glyphosate response of both Conyza canadensis and Chenopodium album was evaluated under different environmental conditions. Reduced glyphosate sensitivity was observed in both species in response to increased temperature, elevated CO2 level, and the combination of both factors. Increased temperature had greater effect on plant survival than elevated CO2 level. In combination, high temperature and elevated CO2 level resulted in loss of apical dominance and rapid necrosis in glyphosate-treated plants. To investigate the mechanistic basis of reduced glyphosate sensitivity, translocation was examined using 14C-glyphosate. In plants that were subjected to high temperatures and elevated CO2 level, glyphosate was more rapidly translocated out of the treated leaf to shoot meristems and roots than in plants grown under control conditions. These results suggest that altered glyphosate translocation and tissue-specific sequestration may be the basis of reduced plant sensitivity. Therefore, overreliance on glyphosate for weed control under changing climatic conditions may result in more weed control failures.

changes in the translocation and distribution of the herbicide. Vacuolar sequestration, limited cellular uptake and rapid necrosis were all found to play a role in reduced plant sensitivity to glyphosate 16 .
Even though photosynthesis is not the primary inhibitory target of glyphosate, it has been reported to be affected by this herbicide. Glyphosate was suggested to cause inhibition of photosynthetic CO 2 assimilation 17 as well as a decrease in intermediates of the photosynthetic carbon reduction cycle 18 . Shikimic acid, one of the main products in the EPSPS pathway, is a precursor of pigments, defense compounds, lignin and other important molecules in plants 19 . Interestingly, glyphosate injury was also found to be correlated with chlorophyll content 20,21 .
This research was conducted to examine the joint effects of increased temperature and elevated CO 2 level on the sensitivity of weeds to glyphosate. To accomplish this objective, we chose two weed species, C. canadensis and C. album, that differ in leaf surface characteristics, flowering phenology and plant architecture. The specific research objectives were (1) to examine the influence of increased temperatures, elevated CO 2 levels, and the combination of both factors on the sensitivity of C. canadensis and C. album to glyphosate and (2) to investigate the mechanistic basis of plant response to glyphosate treatment under these environmental conditions. Results plant response to glyphosate. Plant sensitivity to glyphosate was reduced under high temperatures and elevated CO 2 levels (Table 1). For both species and all populations, plant survival was highest under the combined high temperature and elevated CO 2 (HT/ECO 2 ) treatment. Two out of four C. canadensis populations (CC4 and CC8) had a considerably higher percentage of plants surviving treatment with glyphosate under the combination of low temperature and ambient CO 2 level (LT/ACO 2 ) than all others (Table 1). Thus, differences in plant survival between the LT/ACO 2 and the HT/ECO 2 were not statistically significant for these two populations. However, for the remaining six populations of both C. album and C. canadensis, the survival percentage differed significantly between the LT/ACO 2 and HT/ECO 2 treatments (Table 1). Large differences in plant survival between current and projected environmental conditions were recorded for populations CA1, CA3 (C. album) and CCS (C. canadensis) in which no plants survived glyphosate treatment under LT/ACO 2 but 61.1%, 69.0% and 64.0% of the plants tested, respectively, survived under HT/ECO 2 conditions (Table 1). In addition, a higher percentage of glyphosate-treated plants survived under high temperature (HT/ACO 2 ) than under elevated CO 2 level (LT/ECO 2 ).
Loss of apical dominance and outgrowth of multiple lateral shoots were observed in glyphosate-treated plants grown under high temperature (HT/ACO 2 ) alone and the combination of both high temperature and elevated CO 2 level (HT/ECO 2 ). This phenotype was consistently observed for C. album (Fig. 1a), but only approximately 10% of the C. canadensis plants exhibited a loss of apical dominance under HT/ACO 2 and HT/ECO 2 . Despite using the same photoperiod (11-h) for all treatments, variation in flowering phenology among C. album plants under different temperatures was detected. At the end of the experiment, 21 days after glyphosate treatment, both treated and untreated CA1 plants grown under HT/ACO 2 or HT/ECO 2 conditions had flower buds or flowers while plants grown under LT/ACO 2 and LT/ECO 2 did not have visible reproductive structures (Fig. 1a). spAD measurements. Over the four days following herbicide treatment, leaves of glyphosate-treated plants grown under HT/ECO 2 exhibited more rapid reduction in chlorophyll content (estimated in SPAD units) than leaves of plants grown under LT/ACO 2 (Fig. 2). The differences in SPAD measurements between the two environmental treatments were statistically significant for both C. album and C. canadensis (Table 2). Interestingly, differences between species were also observed as glyphosate-treated C. album plants grown under HT/ECO 2 exhibited faster reduction in chlorophyll (Fig. 2a)  grown under different environmental conditions (Fig. 3). For both species, differences in glyphosate translocation were mainly observed at 12, 24 and 48 hours after treatment (HAT). Higher 14 C-glyphosate signal intensity was detected in the shoot and roots of C. album plants grown under HT/ECO 2 than plants grown under LT/ACO 2 conditions at both 12 and 24 HAT (Fig. 3a). A similar pattern of glyphosate distribution was observed in C. canadensis although the differences in glyphosate distribution among plants grown under the different environmental conditions were not as visually distinguishable as in C. album plants (Fig. 3b). For both species, apparent differences in glyphosate translocation were also observed at 48 HAT. Based on the phosphor imaging results described above, which indicate that the largest differences in 14 C-glyphosate translocation between plants grown under different environmental conditions, occur at 12, 24 and 48 HAT, we investigated the absorption and quantified the distribution of 14 C-glyphosate in different plant parts of C. album and C. canadensis under different environmental conditions (LT/ACO 2 and HT/ECO 2 ) at these time points.
Glyphosate absorption differed markedly between the two species (Fig. 4). C. album plants grown under HT/ECO 2 conditions absorbed 14 C-glyphosate in a significantly greater amount than plants grown under LT/ ACO 2 within 12 and 24 HAT (Fig. 4a). However, at 48 HAT, no statistically significant difference in glyphosate  were not statistically significant (Fig. 4b). Overall, C. album absorbed substantially more 14 C-glyphosate than C. canadensis.
Quantification of 14 C-glyphosate translocation into different plant parts revealed that significantly more glyphosate was retained in the treated leaf of C. album plants grown under LT/ACO 2 than HT/ECO 2 conditions at both 24 and 48 HAT (Fig. 5a). In foliage leaves (i.e., all leaves except the treated leaf), low amounts of 14 C-glyphosate were found in plants grown under both LT/ACO 2 and HT/ECO 2 with no statistically significant differences between the environmental conditions ( Fig. 5b). Higher amounts of 14 C-glyphosate were found in plant stems under HT/ECO 2 than LT/ACO 2 although a statistically significant difference between treatments was only observed at 24 HAT (Fig. 5c). More glyphosate was found in both the shoot apical meristems (Fig. 5d) and the roots (Fig. 5e) of plants grown under HT/ECO 2 compared with plants grown under LT/ACO 2 . For both shoot apical meristems and roots, differences between environmental conditions were statistically significant at 24 and 48 HAT (Fig. 5d,e).
In C. canadensis, significantly more 14 C-glyphosate was translocated out of the treated leaf of plants grown under HT/ECO 2 at all harvest time points (Fig. 6a). No significant differences were observed in the amount of 14 C-glyphosate found in the rosette leaves (i.e., all leaves except the treated leaf) of plants grown under different grown under different environmental conditions and harvested at 6, 12, 24, 48 and 72 hours after treatment (HAT) with glyphosate. LT/ACO 2 -low temperature (18/12 °C) combined with ambient CO 2 (400 ppm), HT/ ECO 2 -high temperature (32/26 °C) combined with elevated CO 2 (720 ppm). Plants were divided into three parts: treated leaf (indicated horizontally by the red arrow), shoot (above treated leaf), and roots (below treated leaf) prior to imaging.   (Fig. 6c,d). Significant differences in the quantity of 14 C-glyphosate between environmental conditions were observed at both 12 and 24 HAT for shoot meristems (Fig. 6c), whereas in the roots, significant differences were observed for all harvest time points (Fig. 6d).

Discussion
Taken together, the results of our study clearly indicate that the control of two major weeds in California agriculture by glyphosate could be reduced under the projected changes in climatic conditions. Compared to current conditions, both C. canadensis and C. album plants were less sensitive to glyphosate under the higher temperatures, elevated CO 2 levels and the combination of both environmental conditions, which are predicted for the future. To the best of our knowledge, this research provides the first experimental evidence of the joint effects of both high temperatures and elevated CO 2 levels on weed sensitivity to glyphosate. Reduced glyphosate sensitivity under high temperature and CO 2 conditions was observed for all four populations of each species. Although the populations used for this study were primarily chosen from herbicide-free habitats, two C. canadensis populations (CC4 and CC8) exhibited a higher percentage of plants surviving glyphosate treatment at low temperature combined with ambient CO 2 level (LT/ACO 2 ) than all others. The wind-mediated seed dispersal, combined with the evolution and spread of glyphosate resistant C. canadensis populations across the Central valley of California 22 , may account for the higher percentage of plants surviving glyphosate under current (LT/ACO 2 ) conditions.
The rapid reduction in chlorophyll content (estimated in SPAD units), loss of apical dominance, and early initiation of reproductive structures observed in glyphosate-treated plants grown under high temperature combined with elevated CO 2 level (HT/ECO 2 ) provide insights into the mechanistic basis of the reduced plant sensitivity to glyphosate under climate change scenarios. It is generally claimed that glyphosate controls weedy plants by binding to, and inhibiting the EPSPS enzyme, which is essential for the biosynthesis of branched-chain amino acids 11 . Interestingly, several recent studies, in addition to this study, have revealed changes in phenological and physiological plant traits caused by glyphosate. Outgrowth of lateral shoots 23 , delayed flower development 24 and reduced stomatal conductance 21 , have been observed in response to glyphosate treatment. Additionally, as a phloem-mobile herbicide, glyphosate exhibits a classic source-to-sink translocation pattern 25 . The influence of glyphosate on photosynthesis-related processes, such as carbon fixation, starch accumulation and general carbohydrate formation, can eventually lead to self-induced limitation of glyphosate translocation 26 . Our findings suggest that most of the glyphosate that was not retained in the treated leaves was translocated into shoot apical meristems and young leaves (strong sinks), which caused rapid leaf decay and thus reduced glyphosate translocation to other plant organs.
Glyphosate absorption differed between the two species. In C. album, significantly higher 14 C-glyphosate absorption was observed in plants grown under HT/ECO 2 compared to LT/ACO 2 conditions. In C. canadensis, 14 C-glyphosate absorption was marginally higher, but not significantly, under LT/ACO 2 conditions. However, despite the differences in glyphosate absorption between species, the translocation and distribution pattern of 14 C-glyphosate within plants, once absorbed, was similar. In glyphosate-treated plants grown under HT/ECO 2 , glyphosate was translocated more quickly out of the treated leaf to other plant tissues than in plants grown under LT/ACO 2 conditions. Moreover, in plants grown under HT/ECO 2 , glyphosate translocation from the treated leaf into strong sinks (e.g. shoot meristems and roots) was rapid for both C. album and C. canadensis (Figs 5d,e and 6c,d, respectively). The rapid movement of glyphosate into shoot apical meristems and roots may reduce the mobility of the herbicide to other parts of the plant thereby reducing the overall sensitivity of plants to glyphosate under higher temperature and CO 2 (HT/ECO 2 ) conditions. It has been hypothesized for many glyphosate-resistant weeds that less glyphosate is translocated from the treated leaf to other plant parts compared to glyphosate-sensitive plants 16 . Interestingly, our results suggest a mechanistic basis for reduced plant sensitivity to glyphosate that differs from the altered glyphosate translocation mechanism hypothesized for many glyphosate-resistant weeds. For both C. album and C. canadensis, reduced translocation of glyphosate from the treated leaf was proposed as the mechanism for glyphosate tolerance 27,28 . Our results suggest that the mechanism leading to reduced glyphosate sensitivity under high temperatures and elevated CO 2 levels may differ from that conferring evolved glyphosate resistance in weeds. The pattern of glyphosate translocation observed in C. canadensis and C. album in this study can also explain the loss of apical dominance and the initiation of lateral shoots in glyphosate-treated plants grown under HT/ECO 2  . 1a). It is well-known that auxin moves basipetally from the apical shoot in order to suppress lateral bud growth 29 . Glyphosate translocation into the shoot apical meristem may cause severe damage to this tissue and, as a result, constrain auxin production. Low quantities of auxin and glyphosate at the whole plant level may enable the outgrowth of lateral shoots which, in turn, could lead to increased plant survival after glyphosate treatment and the phenotype observed in this study.
In conclusion, we have shown that glyphosate-treated plants grown under increased temperature and elevated CO 2 level exhibit reduced glyphosate sensitivity. Thus, the continued overreliance on glyphosate for weed control under changing climatic conditions may result in more weed control failures. In addition, from a practical point of view, the loss of apical dominance and early initiation of reproductive structures, as observed in glyphosate-treated plants grown under high temperature in this study, could further exacerbate weed problems by resulting in an unexpected increase in seed production per plant and rapid replenishment of the soil seed bank. Our translocation studies have revealed variation in glyphosate distribution pattern between plants grown under different environmental treatments. Tissue-specific glyphosate sequestration may be the leading cause for sub-lethal glyphosate quantities at the whole plant level reducing the overall efficacy of the herbicide. Further research is required to determine the exact mechanism leading to the reduced plant sensitivity to glyphosate under altered environmental conditions.  (Table 1). Seeds were collected from 30 randomly selected plants in each population and pooled. To increase the probability of collecting seeds from glyphosate-sensitive individuals, populations were primarily chosen from crops grown organically and from areas where herbicides are less likely to be used. In addition to the seeds sampled from those fields, seeds of a previously characterized glyphosate-susceptible C. canadensis population 30   Based on Intergovernmental Panel on Climate Change 31 predictions 32 , future projected extreme temperatures are estimated to be 3-5 °C higher than current maximum temperatures. Thus, the two temperature treatments chosen for this study were 18/12 °C (day/night) as the current average and 32/26 °C (day/night) as the projected maximum. The projected maximum temperature was calculated by adding 5 °C to the current maximum (27 °C). A difference of 6 °C between day and night temperatures was chosen in accordance with the current day/night temperature difference and with previous studies of C. album (Ziska et al. 13 ) and C. canadensis (Kleinman et al. 12 ). CO 2 treatment levels were ambient (400 ppm) and elevated (720 ppm), which corresponds to future climate projections and within the range of CO 2 levels projected by the year 2100 31,33,34 . Environmentally controlled growth chambers (Conviron-PGR15), equipped with non-dispersive infrared CO 2 analyzers (Horiba model APBA-250E) and valves injecting pure CO 2 to the incoming air stream, were set at either the near normal ambient CO 2 level (400 ppm) or at the elevated CO 2 level (720 ppm). Chamber CO 2 concentrations were logged at 30 second intervals and averaged for each 24 h period, showing that CO 2 levels averaged 490 ± 40 ppm for the ambient treatment and 720 ± 5 ppm for the elevated CO 2 treatment. plant response to glyphosate. Seeds from each C. album and C. canadensis population sampled were germinated in flats filled with commercial potting media (Professional growing mix, SunGro ® Horticulture Canada, Ltd., Vancouver, British Columbia, Canada). Seedlings of C. album at the two-to four-leaf stage and C. canadensis at the three-to four-leaf stage were transplanted into 5 by 5 cm plastic pots (one plant per pot) filled with the same potting media and grown in a growth chamber set at 25/15 °C (day/night) temperatures and 11-h photoperiod, representative of the day length for February/March in California, and a light intensity of 600 µmol m −2 s −1 provided by fluorescent and incandescent bulbs. Seedlings were watered daily.

Materials and Methods
Three days after transplanting, 20-40 healthy seedlings from each population were moved to two growth chambers that differed in the following temperature and CO 2 conditions: [1] LT/ACO 2 -low temperature (18/12 °C) combined with ambient CO 2 (400 ppm), and [2] HT/ECO 2 -high temperature (32/26 °C) combined with elevated CO 2 (720 ppm) but with the same photoperiod and light intensity as described above. Seedlings of C. album were grown to a height of 6-8 cm, whereas seedlings of C. canadensis were grown to the 8-10 rosette leaf stage (5-6 cm in diameter), then treated with glyphosate (Roundup PowerMax ® , Monsanto, St. Louis, MO, USA) at the labeled field rate of 867 g ae ha −1 using an automated spray chamber equipped with a flat-fan 8001E nozzle (TeeJet ® , Spraying Systems Co., Wheaton, IL, USA). The sprayer was calibrated to deliver 187 L ha −1 of glyphosate solution at a pressure of 296 kPa. For each treatment, five unsprayed individual plants were designated as untreated controls. One hour after glyphosate treatment, plants were returned to their respective growth chambers. Plant survival was assessed 21 days after treatment (DAT). The experiment was repeated 2-3 times. Treatment combinations and experimental runs were rotated between the two chambers.
In addition, seedlings of two populations (CA1 for C. album and population CCS for C. canadensis) were assessed for plant response to glyphosate under two additional temperature and CO 2 combinations: [3] LT/ECO 2 -low temperature (18/12 °C) combined with elevated CO 2 (720 ppm), and [4] HT/ACO 2 -high temperature (32/26 °C) combined with ambient CO 2 (400 ppm). Photoperiod, light intensity, glyphosate application and data collection were the same as described above. Due to a shortage of available growth chambers in which CO 2 levels could be regulated, only one population of each species could be tested at these environmental conditions. Konica Minolta Sensing, Inc., Osaka, Japan), following the method of Yannicccari et al. 35 . Three independent measurements were taken at the middle section of the youngest fully expanded leaf four days following treatment with glyphosate. Chlorophyll content (estimated in SPAD units) was calculated as the average of the three measurements and expressed for glyphosate-treated plants as a percentage of the respective values obtained for untreated control plants.
Absorption and translocation of 14 C-glyphosate. Glyphosate absorption and translocation under different temperature and CO 2 conditions was assessed using a completely randomized experimental design with four replicates. Seeds from C. album population CA1 and C. canadensis population CCS were germinated and seedlings grown as described above. Seedlings of C. album at the two-to four-leaf stage and C. canadensis at the three-to four-leaf stage were transplanted into 40 ml vials and grown hydroponically with a dilute nutrient solution, as described in Moretti and Hanson 28 , in the growth chambers maintained at LT/ACO 2 and HT/ECO 2 conditions. A solution containing glyphosate at a final concentration that approximated an 867 g ae ha −1 spray solution at 187 L ha −1 carrier volume was prepared by mixing 14  Phosphor image analysis was used to visualize herbicide translocation. 14 C-glyphosate treated and dissected plant parts (treated leaves, shoots, and roots) were pressed between two layers of paper and dried at 60 °C for 72 h. After cooling to room temperature, each sample was placed in a 20 × 40 cm exposure cassette (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) and brought into contact with a standard storage phosphor screen (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) for 24 h. Glyphosate translocation was visualized using the Storm 860 PhosphorImager system (Molecular Dynamics, Sunnyvale, CA, USA). Image analysis was conducted using the ImageQuant 5.0 software (Amersham Biotech-Molecular Dynamics, Sunnyvale, CA, USA).
Following phosphor image analysis, 14 C-glyphosate translocation was quantified at three harvest time points, 12, 24 and 48 HAT, for both species. To measure the amount of non-absorbed glyphosate, rinsate (i.e., the treated leaf wash) was evaporated to dryness and resuspended in 10 ml of scintillation cocktail (Ultima Gold ™ , Perkin Elmer, Walthan, MA). Rinsate radioactivity was quantified using a liquid scintillation spectrometer (LSS) device (LS 6500, Beckman Coulter, Fullerton, CA). The oven-dried plant samples used for phosphor image analysis were also used to assess the distribution of 14 C-glyphosate. Treated leaves and roots were combusted with no further dissection whereas shoots were divided into several subsections as illustrated in Supplementary Fig. S1. For C. album, each shoot was divided into three parts: 1) shoot apical meristems including young undeveloped leaves, 2) leaves + petioles below the treated leaf, and 3) stem. For C. canadensis, each shoot was divided into two parts: 1) shoot meristems including young undeveloped leaves and 2) the remaining rosette leaves. Different plant parts were placed separately into a combustion cone and dried at 60 °C for 96 h. Each cone was combusted in a biological oxidizer (Sample Oxidizer Model 307, PerkinElmer, Waltham, MA, USA). The evolved 14 CO 2 was trapped in 10 ml of a carbon dioxide adsorbent solvent (Carbo-Sorb ® E, PerkinElmer, Waltham, MA, USA) and mixed with 10 ml of scintillation cocktail (Permaflour ® E + , PerkinElmer, Waltham, MA, USA). Radioactivity was quantified using the LSS device described above. statistical analysis. Data on the survival of glyphosate-treated plants grown under LT/ACO 2 and HT/ECO 2 were analyzed using a generalized linear model (GLM) with PROC GENMOD of SAS (ver 9.4., SAS Institute Inc., Cary, NC, USA). The loglikelihood ratio test was used to assess the significance of the interaction between experimental runs and treatments as well as the main effects of experimental run. Probabilities of plant survival and the 95% confidence intervals for all possible combinations of populations by treatment were estimated using the LSMEANS statement of SAS. For populations CA1 and CCS, an additional analysis of plant survival data was conducted across all four temperature and CO 2 combinations (LT/ACO 2 , LT/ECO 2 , HT/ACO 2 and HT/ECO 2 ). Data were analyzed using ANOVA in JMP (ver. 13) statistical package (SAS Institute Inc., Cary, NC, USA). Means were compared using Tukey-Kramer honestly significant difference (HSD) test (α = 0.05).
SPAD measurements were pooled for each species and means were compared in agreement with a Levene's ANOVA test for homoscedasticity of variance (P ≥ 0.05). No outliers were identified with the studentized residuals technique based on a t-distribution with α = 0.05. Normality of residues (Shapiro-Wilk's test) and homoscedasticity of variance (Fligner-Killeen's test) were tested with α = 0.05. Multiple linear regressions of leaf greenness as a function of days after treatment with glyphosate were performed separately for each treatment (LT/ACO 2 and HT/ECO 2 ) and weed species, and regression slopes were obtained with their 95% confidence intervals.
For the absorption and translocation studies, total glyphosate quantity was converted into percentages according to equations 1 and 2. R signifies the recovered radioactivity. Translocation of 14 C-glyphosate to different plant sections was calculated using equation 3 where R signifies the recovered radioactivity; ME, the shoot meristems including young undeveloped leaves; LS, the remaining rosette leaves, SM, the stem; RS, the roots, and TL, the treated leaf. Data from absorption and translocation studies were analyzed using ANOVA in JMP (ver. 13) statistical package (SAS Institute Inc., Cary, NC, USA). Means were compared using Student's t-test (α = 0.05). Data were visualized separately for each treatment (LT/ACO 2 and HT/ECO 2 ) and weed species using SigmaPlot (ver. 12) software (Systat Software Inc., San Jose, CA, USA).