Metabolic Profiling and Enzyme Analyses Indicate a Potential Role of

Subscriber access provided by FLORIDA ATLANTIC UNIV

Article

Metabolic profiling and enzyme analyses indicate a potential role of antioxidant systems in complementing glyphosate resistance in an Amaranthus palmeri biotype Amith Sadananda Maroli, Vijay Nandula, Franck E. Dayan, Stephen Duke, Patrick Gerard, and Nishanth Tharayil J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.5b04223 • Publication Date (Web): 02 Sep 2015 Downloaded from http://pubs.acs.org on September 8, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Agricultural and Food Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 44

Journal of Agricultural and Food Chemistry

1 1

Metabolic profiling and enzyme analyses indicate a potential role of antioxidant systems in

2

complementing glyphosate resistance in an Amaranthus palmeri biotype

3

Amith S. Maroli1, Vijay K. Nandula2, Franck E. Dayan3, Stephen O. Duke3, Patrick Gerard4,

4

Nishanth Tharayil1*

5

1

Department of Agricultural & Environmental Sciences, Clemson University, Clemson, South Carolina 29634 USA,

6

2

Crop Production Systems Research Unit, United States Department of Agriculture, Stoneville, Mississippi 38776,

7

USA, Natural Products Utilization Research Unit, Agricultural Research Service, United States Department of

8

Agriculture, University, Mississippi 38677, USA, Department of Mathematical Sciences, Clemson University,

9

Clemson University, South Carolina 29634, USA,

10

3

4

*Corresponding Author (Phone: 864-656-4453, Email: [email protected])

11

12

13

14

15

16

17

18

19

20

ACS Paragon Plus Environment


2 21

Abstract

22

Metabolomics and biochemical assays were employed to identify physiological

23

perturbations induced by glyphosate in a susceptible (S) and resistant (R) biotype of

24

Amaranthus palmeri. At 8 h after treatment (HAT), shikimic acid accumulated in both R-

25

and S-biotypes in response to glyphosate application, which was accompanied by an

26

increase in organic acids and aromatic amino acids in the R-biotype and an increase in

27

sugars and branched-chain amino acids in the S-biotype. However, by 80 HAT the

28

metabolite pool of glyphosate-treated R-biotype was similar to that of the water-treated

29

control S and R-biotype, indicating physiological recovery. Furthermore, glyphosate-

30

treated R-biotype had lower reactive oxygen species (ROS) damage, higher ROS

31

scavenging activity, and higher levels of secondary compounds of the shikimate

32

pathway. Thus metabolomics, in conjunction with biochemical assays, indicate that

33

glyphosate induced metabolic perturbations are not limited to shikimate pathway, and

34

the oxidant quenching efficiency could potentially complement the glyphosate

35

resistance in this R-biotype.

36

Keywords: Non-targeted metabolomics, anti-oxidant enzymes, Amaranthus palmeri,

37

GC-MS metabolite profiling, LC-MS/MS profiling, Free radicals, Glyphosate resistance,

38

Reactive Oxygen Species (ROS)


Page 2 of 44

Page 3 of 44


3 39

40

Introduction Globally,

many

crop

production

systems

employ

glyphosate

(N-

41

(phosphonomethyl)glycine) to control annual and perennial grasses and broad-leaf

42

weeds.1,2

43

enolpyruvylshikimate-3-phosphate synthase (EPSPS) that catalyzes the penultimate

44

step of shikimate pathway, the conversion of shikimic acid to chorismate, the precursor

45

for aromatic amino acids (tyrosine, phenylalanine and tryptophan) and other secondary

46

plant metabolites.3 Glyphosate competes with phosphoenolpyruvate (PEP), a substrate

47

for EPSPS enzyme, to form a very stable enzyme-herbicide complex that inhibits the

48

product-formation reaction.4 The broad spectrum of herbicidal activity of glyphosate is

49

optimally employed in modern agriculture by utilizing crops engineered for glyphosate

50

resistance. However, due to the over-reliance of glyphosate to combat weeds, at least

51

32 weed species have evolved resistance to glyphosate.5 Of these, glyphosate-resistant

52

Amaranthus palmeri is the most economically damaging weed, mainly in US cotton,

53

corn and soybean production systems, causing yield losses ranging from 50 to 95%

54

depending on several factors, including the competitive ability of crops.6,7

Glyphosate

works

by

specifically

inhibiting

the

enzyme

5-

55

Application of genomics and transcriptomics technologies have helped to identity

56

the common causes/mechanisms of evolved glyphosate resistance in some of the

57

weeds,8,9 including a higher abundance of the EPSPS enzyme resulting from an

58

increased EPSPS gene copy number,10 and reduced translocation of glyphosate to the

59

target site tissues (especially meristems) in A. palmeri.11 Though application of these

60

approaches provide a robust understanding of the genetic regulation on the potential

61

functioning of the organism, a comprehensive picture of the phenotypic manifestations



4 62

of evolved glyphosate resistance is still lacking. This is partly because the molecular

63

markers that are tracked using genomics, transcriptomics and proteomics (DNA, RNA

64

and proteins, respectively) are susceptible to functional alteration either by epigenetic

65

modifications or post-transcriptional and post-translational modifications, resulting in

66

altered phenotypes.12 Metabolomics, which focuses on comprehensive identification

67

and quantitation of low-molecular weight metabolites (metabolome) broadens the

68

understanding of the functioning of a physiological system. Since metabolomics

69

measures the final products of gene expression, protein expression and enzymatic

70

activity as affected by the environment- the metabolites, it offers a powerful approach

71

for molecular phenotyping.13 Thus, while genomics and proteomics provide information

72

about the processes that are genetically programmed to happen, metabolomics gives a

73

more definite and real-time representation of gene-environment interactions.

74

In plants, under stressful conditions, accumulation and depletion of several

75

primary metabolites are commonly observed.14,15 Therefore, monitoring the changes in

76

metabolic levels can provide clues to their roles in stress responses and regulation. A

77

key metabolic pathway in plants is the shikimate pathway that links carbon and nitrogen

78

metabolism.16 Apart from its role in aromatic amino acid synthesis, it serves as a major

79

sink for intermediate metabolites from the central carbon metabolism pathways

80

(glycolytic and pentose phosphate pathways). Any perturbations in the shikimate

81

pathway will lead to the disruption in aromatic amino acids synthesis, as well as alter

82

the metabolic stoichiometry of other carbon intermediates resulting in system-wide

83

perturbations. Metabolomics would therefore be an ideal approach not only to map the


Page 4 of 44

Page 5 of 44


5 84

glyphosate-induced disruption of plant metabolism, but also to illuminate the physiology

85

that confers resistance to some biotypes.

86

The glyphosate-mediated disruption of the shikimate pathway could have potential

87

adverse effects on other cellular processes resulting in secondary effects such as

88

creation of reactive radicals. Earlier studies reported that in addition to site-specific

89

action, glyphosate also disrupts the photosynthetic machinery by reducing ribulose-1,5-

90

bisphosphate carboxylase/oxygenase (RuBisCO) activity and 3-phosphoglyceric acid

91

(3-PGA) levels17, which can result in the generation of reactive oxygen species (ROS)

92

causing several fold increase in lipid peroxidation.18,19 This ROS-mediated glyphosate

93

damage is further supported by the proteomic analysis that shows glyphosate-treated

94

rice plants have a compositionally similar, but quantitatively lower, proteomic profile as

95

that of rice plants treated with paraquat,20 a free radical-producing herbicide. Thus,

96

disruption of EPSPS could result in secondary toxic effects mediated through the

97

production of ROS. Such an effect might be enhanced in glyphosate-resistant A.palmeri

98

if some tissues have inadequate protection by EPSPS gene amplification as there is

99

evidence that EPSPS gene expression is not equal in all tissue types in these plants.21

100

Previous application of metabolomics in understanding the effect of glyphosate

101

on the physiology of plants was limited to the model plant species Arabidopsis thaliana

102

and transgenic crop plants.22-24 Also metabolomics approach has been recently adopted

103

to understand effect of chemical stresses on Lolium perenne upon exposure to a non-

104

lethal dose of glyphosate.25 However, to date no studies have employed metabolomics

105

to characterize the physiology of herbicide resistance in weeds. The objective of the

106

present study is to use non-targeted metabolite profiling to elucidate the differential



6 107

metabolite-level responses to glyphosate in a resistant and a susceptible biotype of A.

108

palmeri, and to assess the potential role of anti-oxidant machinery in complementing

109

glyphosate resistance.

110

Materials and Methods

111

Plants Biotypes

112

Glyphosate-susceptible (S-biotype) and glyphosate-resistant (T4B1, R-biotype)

113

Amaranthus palmeri seeds were obtained from Stoneville, Mississippi (Crop Production

114

Systems Research Unit, USDA-ARS, Stoneville, MS). These biotypes have been well

115

characterized and GR50 values for the R-biotype and S-biotype were previously

116

calculated as 1.3 kg ha-1 and 0.09 kg ha-1, respectively (~15-fold resistance).11

117

Glyphosate resistance in T4B1 were verified by growing field collected seeds from

118

suspected resistant biotype under greenhouse conditions and screened with glyphosate

119

at 0.42 kg ae ha-1. Only those plants that survived this treatment were classified to be

120

glyphosate resistant (first-generation resistant, F-1) and represented that population.

121

Second generation (F-2) resistant biotypes were obtained by physically spreading the

122

pollen (generated from the greenhouse grown glyphosate resistant F-1 generations)

123

from male plants on the female plants (greenhouse grown glyphosate resistant F-1

124

generations) every morning over a period of 2 weeks. Each female plant that produced

125

viable seeds was considered a unique F-2 biotype within a population, with all biotypes

126

within the population having the same male parent. Biotypes raised from the F-2 seeds

127

were confirmed to be resistant to glyphosate following screening with glyphosate at a

128

rate of 0.42 kg ae ha-1. F-2 generated seeds were planted in germination trays


Page 6 of 44

Page 7 of 44


7 129

containing a commercial germination mixture and then were transplanted to individual

130

pots (10 cm diameter x 9 cm deep) containing the germination mixture upon

131

germination. The plants were grown in a greenhouse at 30°C/20°C day/night temp with

132

a 14 h photoperiod. The pots were fertilized five days after transplanting with 50 ml of 4

133

g

134

Marysville, OH, USA) and sub-irrigated every alternate day until harvest.

135

Experimental Design and Glyphosate Application

l-1

fertilizer

(MiracleGrow®,

24%-8%-16%,

Scotts Miracle-Gro Products,

Inc.,

136

Eight days after transplanting, the S- and R-biotypes were randomly assigned to

137

two treatment groups:- control and glyphosate (Roundup WeatherMAX®, Monsanto Co,

138

St. Louis, MO) sprayed at a rate of 0.4 kg ae ha-1 which corresponds to approximately

139

half the recommended field application rate of glyphosate Roundup WeatherMax has

140

about 49% of glyphosate-phosphate salt. We were not able to obtain the inert

141

proprietary ingredients that could be used to treat the control plants, hence the control

142

plants were treated with water. Two weeks after transplanting, the respective treatments

143

(water or glyphosate) were applied to 12 plants per biotype using an enclosed spray

144

chamber (DeVries Manufacturing, Hollandale, MN) calibrated to deliver 374 l ha-1

145

through an 8001E ﬂat fan nozzle (Tee Jet Spraying Systems Co., Wheaton, IL). Both R-

146

and S- biotypes were morphologically similar (plant height ~20cm;) at the time of

147

treatment application. Three young apical leaves along with the meristem were

148

harvested from each treatment replicate (six plants per glyphosate and control

149

treatments) at 8 h after treatment (HAT) and 80 HAT and were immediately frozen in

150

dry-ice and stored at -80°C. Six independent treatment replicates were maintained for

151

all subsequent analyses. At 8 HAT, no visible injuries were observed on either



8 152

glyphosate treated S- or R-biotypes. At 80 HAT, the glyphosate treated S-biotype

153

displayed visible leaf necrosis while the glyphosate treated R-biotype and the water

154

treated S- and R-biotypes showed no visible injuries. To minimize the variations due to

155

circadian rhythm, the plants were harvested at the same time of the day (~8 h after

156

sunrise) for both the sampling time. The harvested leaves were finely ground with dry

157

ice using a mortar and pestle and stored at -80°C and the finely powdered leaves were

158

used for all subsequent analyses.

159

Metabolite Profiling by GC/MS

160

Low-molecular weight polar metabolites in the tissues were identified and

161

quantified using gas chromatography/mass-spectrometry. Polar metabolites were

162

extracted from the ground tissues as described by Lisec et al26 and Suseela et al,15 with

163

slight modifications. Briefly, about 100 mg of finely powdered leaf tissues was weighed

164

into 1 ml methanol and homogenized by sonicating in an ice-bath for 20 sec at 50%

165

amplitude, centrifuged at 12,000 × g and rapidly cooled on ice. The supernatant was

166

transferred to pre-cooled glass tubes, and equal volume of cold chloroform was added

167

and cooled at 4°C. Metabolites were fractioned into polar and non-polar phases with

168

addition of 1 ml of cold water and then centrifuged at 671 × g for 1 min. About 1.5 ml of

169

the aqueous-methanol phase was transferred to microfuge tubes. A subsample (150 µl)

170

of this extract was transferred into vials with glass inserts and 5 µl of 5 mg ml-1 ribitol

171

(internal standard) in hexane and 5 µl of 5 mg ml-1 of d27-myristic acid (retention time

172

lock) in hexane were added to the vials and then completely dried under a nitrogen

173

stream. Dried samples were methoxylaminated at 60°C with 20 µl of methoxylamine

174

hydrochloride (20 mg ml-1) in pyridine for 90 min followed by silylation of the metabolites


Page 8 of 44

Page 9 of 44


9 175

with 90 µl of N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA) with 1%

176

trimethylchlorosilane (TMCS) for 30 min at 40°C. The derivatized metabolites were

177

separated by gas chromatography (Agilent 7980; Agilent Technologies, Santa Clara,

178

CA) on a J&W DB-5ms column (30 m x 0.25 mm x 0.25 µm, Agilent Technologies), and

179

analyzed using a transmission quadrupole mass detector (Agilent 5975 C series) with

180

an electron ionization interface. The initial oven temperature was maintained at 60°C for

181

1 min, followed by temperature ramp at 10°C per min to 300°C, with a 7 min hold at

182

300°C. Carrier gas (He) flow was maintained at a constant pressure of 76.53 kPa and

183

the injection port and the MS interphase were maintained at a constant temperature of

184

270°C; the MS quad temperature was maintained at 150°C; and the MS source

185

temperature was set at 260°C. The electron multiplier was operated at a constant gain

186

of 10 (EMV = 1478 V), and the scanning range was set at 50–600 amu, achieving 2.66

187

scans sec-1. Metabolite peaks were identified by comparing the absolute retention time,

188

retention-time index of the sample with that of the in-house metabolomics library

189

supplemented with Fiehn Library (Agilent Technologies, Wilmington, DE, USA,

190

G1676AA).27 Data deconvolution was performed using Automatic Mass Spectral

191

Deconvolution and Identification System (AMDIS v2.71, NIST) using the following

192

parameters: Peak identification was limited to a minimum match factor of 70 relative to

193

the Retention Index (RI) library and having a RI threshold window of 10 x 0.01 RI. A

194

maximum match factor penalty of 30 was deducted from the final calculated net value in

195

case of RI mismatch. The peak abundance (integrated signal) was normalized with that

196

of the ribitol (internal standard) abundance and expressed as percent of ribitol

197

((metabolite intensity/ribitol intensity)x100).



Page 10 of 44

10 198

Phenylalanine Ammonia Lyase (PAL) Assay

199

The PAL assay was carried according to Shang et al.28 with slight modifications.

200

Briefly, 0.15 g of the ground leaves were homogenized in 1 ml of ice-cold sodium borate

201

buffer (pH 8.3) by sonication and the homogenates were immediately centrifuged

202

(12,000 × g, 5 min). The reaction mixtures consisting of the buffer and homogenates

203

were pre-incubated at 40°C for 5 min. The reaction was started by the addition of 50

204

mM L-phenylalanine. Controls (without L-phenylalanine) were prepared to determine

205

plant endogenous trans-cinnamic acid (t-CA) content. The reaction was terminated after

206

1 h with 5 N HCl and analyzed on HPLC-UV (LC 20-AT UFLC and Prominence SPD

207

M20-A photodiode array, Shimadzu Scientific, Columbia, MD) by monitoring the

208

abundance of t-CA at 290 nm. HPLC separation of t-CA was performed on Kinetex® XB-

209

C18 column (100 mm x 4.60 mm x 2.6 µm, 100 Å, Phenomenex, Torrance, CA) at an

210

isocratic flow rate of 0.7 ml min-1 using a solvent composition of 50:50 0.05% formic acid

211

in water and methanol. The t-CA levels in the samples were calculated with respect to

212

the calibration curve of the t-CA standard solution prepared in the same sodium-borate

213

buffer as that of the samples.

214

Glutathione Reductase and Thiobarbituric Acid Reactive Substances (TBARS) Assays

215

Glutathione reductase (GR) activity was determined based on the oxidation of

216

NADPH at 340 nm in presence of oxidized glutathione disulfide (GSSG).29 GR enzyme

217

was extracted from the leaves by sonication with extraction buffer [0.1 M potassium

218

phosphate (pH 7.6) and 2 mM EDTA] and then transferred into the assay mixture

219

composed of 100 mM potassium phosphate buffer (pH 7.8), 0.5 M GSSG and 0.2 M

220

DTT. The assay was initiated with the addition of 0.2 M NADPH and incubated at 25°C


Page 11 of 44


11 221

for 30 min. Endogenous levels of NADPH was determined with a blank (No NADPH

222

addition). The GR activity was determined spectrophotometrically based on the

223

extinction coefficient of NADPH measured at 340 nm using a JASCO V-550 UV-Vis

224

spectrophotometer (Supporting Information Equation 1).

225

Malondialdehyde (MDA) was quantified by a modified thiobarbituric acid reactive

226

substances (TBARS) assay as reported by Hodges et al.30 Briefly, about 0.1 g of ground

227

tissue was weighed into 1 ml extraction buffer (80:20, methanol:water solution) and

228

sonicated thrice for 20 sec. A 250 µl volume of the extract was added to a final volume

229

of 1 ml of assay mixture (50 mM NaOH, 25% (vol/vol) HCl and 7.5 mM thiobarbituric

230

acid (TBA)) and the reaction was started by incubating the assay mixture at 95°C. After

231

30 min, the reaction was stopped by rapidly cooling in ice for about 10 min. To account

232

for interfering compounds, the absorbance was read at wavelengths of 440 nm, 532 nm

233

and 600 nm. False positive absorbance (absorbance by non MDA compounds) was

234

corrected by incubating the extracts in a TBA free assay solution. MDA equivalents

235

were calculated as described by Hodges et al30 (Supporting Information Equation 2).

236

Total Soluble Protein Quantification

237

Proteins were quantified with the bicinchoninic acid (BCA)-protein assay kit

238

according to the manufacture’s recommended protocol (BCA Protein Assay kit; Pierce

239

Chemical Co., Rockford, Ill.). Proteins were extracted from ground tissue into extraction

240

buffer (20 mM phosphate buffer, pH 7.6) by sonication. The extract was diluted 10X with

241

water and added to 400 µl of BCA working reagent (BCA-WR, 50:1, BCA Reagent A:

242

BCA Reagent B) and made up to 1 ml using deionized (DI) water. The assay mix was



Page 12 of 44

12 243

incubated for 30 min at 37°C and then rapidly cooled in ice for 5 min. The absorbance of

244

all the samples was measured at 562 nm using a UV-Vis spectrophotometer within 10

245

min of cooling. The concentrations of total soluble proteins (TSPs) were quantified using

246

a protein concentration curve with bovine serum albumin as the standard protein.

247

Statistical Analysis

248

The relative concentration of metabolites across the biotypes and treatments

249

were analyzed by multivariate and univariate statistical analyses. To evaluate the main

250

and interactive effects of biotypes and herbicide treatments we first analyzed the

251

metabolomics data with permutational multivariate analysis of variance (PERMANOVA;

252

Primer 7 PERMANOVA+, version 7.0.5, Primer-E Ltd, Plymouth, UK). Due to the non-

253

normal distribution nature of metabolic data, PERMANOVA analysis is a more robust

254

test than ANOVA/MANNOVA to identify statistical significance of the treatments.31,32

255

The metabolomics data were auto-scaled to satisfy the assumptions of normality and

256

equal variance before multivariate analyses. Supervised Partial Least Squares-

257

Discriminant Analysis (PLS-DA) models were constructed using Metaboanalyst33 and

258

the Microsoft® Excel® add-in Multibase package (Numerical Dynamics, Japan) to test

259

the significance of the effects of biotype and herbicide treatments on the metabolite

260

profiles, and the PLS-DA model was validated using permutation testing.34 Hierarchical

261

cluster analysis, using Euclidean distance as similarity measure, of the metabolite

262

responses with respect to the treatments were visualized using heatmaps. Univariate

263

statistical analysis (ANOVA) was also performed to determine the statistical significance

264

of metabolite among the different treatment groups (biotypes and herbicide) followed by

265

Tukey’s HSD post-hoc test. Statistical significance of the biochemical assay responses


Page 13 of 44


13 266

(concentrations of t-CA, NADPH and malondialdehyde (MDA)) were tested by two-way

267

analysis of variance. Differences among individual means were tested using Tukey’s

268

HSD multiple comparison tests with Pvalue < 0.05 considered statistically significant.

269

Statistical analyses were done using SAS (v 9.2, SAS Institute, Cary, NC).

270

Results/Discussion

271

GC-MS Metabolite Profiling

272

Based on the pre-defined deconvolution parameters (see methods section), GC-

273

MS analysis positively identified more than 60 metabolites across all samples based on

274

their mass-spectral fingerprints and retention-index matches (Supporting Information

275

Table 1). The pairwise correlations between these metabolites were quantified using the

276

Pearson correlation coefficient as similarity measure with a threshold window of 0.5,

277

and the resulting 51 metabolites that responded to treatments were used for further

278

statistical analyses. PERMANOVA analysis confirmed that there was a significant

279

interaction effect between biotype and herbicide treatment for both the time points of

280

harvest (Supporting Information Table 2). From the total identified metabolite pool at 8

281

and 80 HAT, 49 and 52 compounds were selected that had a false discovery rate (FDR;

282

percent of false positive that are predicted to be significant) of less than 0.05 (delta =

283

0.6; Supporting Information Table 3). The robustness of the class discrimination was

284

verified through permutation testing of separation distance based on the ratio of the

285

between group sum of the squares and the within group sum of squares. The

286

permutation cross validation of PLS-DA model for 8 and 80 HAT had P =0.0005 over

287

2000 iterations and the models goodness of fit was > 0.85.


PLS-DA revealed a


Page 14 of 44

14 288

clustering of the treatments for both the harvest times (Figure 1). At 8HAT, component 1

289

axis differentiated between the two biotypes (S- and R-biotype) while the PC-2 axis

290

delineated the two treatments (water and glyphosate) (Figure 1c). Interestingly, at

291

80HAT, along PC-1 axis, the glyphosate-treated R-biotype had a metabolic pool similar

292

to that of control R- and S- biotypes, indicating a potential recovery of R-biotype from

293

glyphosate toxicity (Figure 1d). The distinction of the glyphosate treated S-biotype at 80

294

HAT from the rest of the treatments, could be due to the continued glyphosate induced

295

disruption of the shikimate pathway resulting in the differential abundance of several key

296

metabolites such as shikimate, 3-dehydroshikimate, tyrosine, tryptophan, asparagine

297

etc. (Figure 1b)

298

Univariate analysis of the relative abundance of the metabolites in response to

299

the treatments indicated significant differences in key metabolites (Supporting

300

Information Table 3). At 8HAT, most of the organic acids (such as succinic, glyceric,

301

malonic, citric, etc.) and aromatic amino acids were higher in the glyphosate-treated R-

302

biotype, while the S-biotype had a higher abundance of sugar metabolites (glucose,

303

sucrose, fructose, talose) and branched chain amino acids (Figure 2a). However, at 80

304

HAT, the glyphosate-treated S-biotype had the highest concentration of metabolites

305

involved in primary carbon and nitrogen metabolism, possibly due to growth retardation

306

and subsequent non-utilization of these primary metabolites due to glyphosate toxicity

307

(Figure 2b). At 8 HAT, hierarchical clustering grouped the water-control and glyphosate

308

treatment into major clusters indicating that, irrespective of their resistance level, the

309

metabolism of both biotypes were similarly influenced by glyphosate (Supporting

310

Information Figure 1a). Within the herbicide treatment-clusters the metabolic profile of


Page 15 of 44


15 311

the S-biotype was different from that of the R-biotype. Also, the metabolite pool of non-

312

treated S- and R-biotypes were more similar to each other compared to the metabolic

313

responses between glyphosate-treated S- and R-biotypes (Figure 2a). At 80 HAT,

314

hierarchical clustering analysis grouped the glyphosate treated S-biotype into a distinct

315

cluster while the glyphosate treated R-biotype was more closely associated with the

316

water treated (control) S- and R-biotypes (Supporting Information Figure 1b). The

317

metabolite pool of glyphosate treated R-biotype resembled that of the water treated S-

318

biotype rather than the water treated R-biotype (Figure 2b).

319

Shikimic acid accumulated in both the glyphosate-treated S- and R-biotypes in 8

320

HAT, indicating an initial inhibition of EPSPS enzyme in both biotypes. Compared to the

321

8 HAT, at 80 HAT the shikimate accumulation did not differ significantly in the R-

322

biotype, while the shikimate concentration increased by 3-fold in the S-biotype (Figure

323

3). This continued accumulation of shikimic acid in the S-biotype could be due to the

324

continued inhibition of EPSPS by glyphosate, and/or due to a loss of feedback control of

325

the shikimate pathway. In plants, about 20% of the fixed carbon that flows through the

326

shikimate pathway is regulated by first enzyme 3-deoxy-D-arabinoheptulosonate 7-

327

phosphate (DAHP) synthase.35a Though no known inhibitor of plant DAHP synthase has

328

been identified, arogenate is considered as a potential candidate for the allosteric

329

inhibition of DAHP synthase.35b As arogenate is synthesized downstream of shikimic

330

acid, inhibition of EPSPS by glyphosate could result in decreased levels of arogenate

331

and therefore impairing the allosteric regulation of DAHP synthase in the S-biotype,

332

leading to continued accumulation of shikimic acid. Studies have reported an initial

333

increase, followed by subsequent decrease, in shikimate levels following exposure to



Page 16 of 44

16 334

sub-lethal doses of glyphosate in both glyphosate susceptible and resistant plants.36-38

335

The stable levels of shikimic acid at 80 HAT in the R-biotype could be possibly due to its

336

conversion into downstream metabolites such as simple phenolic acid derivatives

337

and/or loss of shikimic acid pathway inhibition due to the induction of expression of the

338

multiple EPSPS gene copies, thereby diluting glyphosate effects on EPSPS.39,40

339

The inhibition of the shikimate pathway by glyphosate results in the disruption of

340

aromatic amino acid biosynthesis. The aromatic amino acid concentrations in the

341

control S- and R- biotypes were similar at both 8 and 80 HAT indicating no inherent

342

differences between the biotypes in the absence of herbicide-induced stress. However,

343

after 80 HAT, there were significant changes in the concentrations of the aromatic

344

amino acids in both the biotypes. The levels of phenylalanine in the S-biotype did not

345

differ between 8 and 80 HAT but there was a 3-fold increase in the levels of tryptophan

346

and tyrosine. A similar observation of tyrosine accumulation after glyphosate application

347

was reported in glyphosate susceptible purple nutsedge.41 In the R-biotype the

348

concentration of phenylalanine increased by more than 10-fold while the levels of

349

tyrosine and tryptophan decreased at 80 HAT (Figure 4). However, phenylalanine

350

increased similarly in the untreated S biotype between 8 and 80 HAT. A few studies

351

have reported the effect of these aromatic amino acids in reversing glyphosate

352

toxicity,42-44 although other studies have reported little or no effects of exogenous

353

application of phenylalanine and tryptophan in reducing the toxicity of glyphosate.45a,45b

354

The several fold larger phenylalanine pool in the glyphosate treated R-biotype than in

355

the S-biotype at 80 HAT could serve as a significant pool for antioxidant

356

phenylpropanoid metabolite synthesis in the R-biotype, which could partially mitigate the


Page 17 of 44


17 357

toxicity of glyphosate. Although, the non-aromatic amino acids did not differ in the

358

control S- and R-biotypes, the glyphosate-treated S-biotype had a higher pool of non-

359

aromatic amino acids, possibly due to reduced protein synthesis or other amino acid

360

utilization because of metabolic disruption caused by inhibition of the shikimate

361

pathway. However, it should be noted that the interpretation of the physiology of plants

362

based on pool sizes of pathway intermediates is difficult because pool size does not

363

reflect pool flux.

364

Glyphosate has only one enzyme binding target in the plant, EPSPS. Blockage

365

of the shikimate pathway at this site leads to many adverse secondary and tertiary

366

effects on other pathways and processes. Most of the measured metabolite changes

367

occur in the linked pathways of glycolysis, oxidative pentose pathway and the

368

tricarboxylic acid (TCA) cycle (Figure 5a, 5b). In plants, the energy producing stage of

369

glycolysis is regulated by at least six known mechanisms.46 The rate limiting step of

370

glycolysis, conversion of fructose-6-phosphate to fructose-1,6-bis-phosphate by

371

pyrophosphate dependent phosphofructokinase (PPi-PFK), is activated by inorganic

372

phosphates (Pi) and inhibited by PEP.47 The ratio of Pi:PEP regulates the activity of

373

PPPi-PFK. As glyphosate competes with PEP for binding to EPSPS, the glyphosate

374

sensitive biotype will have higher levels of under-utilized PEP which could then inhibit

375

the conversion of fructose-6-phosphate to fructose-1, 6-bis-phosphate. This block in the

376

rate limiting step of glycolysis could then lead to an overall accumulation of sugars due

377

to impaired carbon metabolism.48,49 This influence of glyphosate on carbohydrate

378

metabolism is reflected in the accumulation of sugars in both the S- and R- biotypes

379

within 8 HAT (Figure 6a). Sugar moves from sources (photosynthetically active tissues)



Page 18 of 44

18 380

to metabolically active sinks (e.g., meristems and young leaves), so inhibition of growth

381

of the sinks results in sugar accumulation. Interestingly, at 80 HAT, the sugar levels in

382

the glyphosate-treated S-biotype continued to accumulate, while that of the glyphosate-

383

treated R-biotype was similar to that of the control biotypes (Figure 6b), indicating a

384

recovery of the R-biotype.

385

The blockage of the shikimate pathway by glyphosate reverberated across other

386

biochemical pathways and processes as evident by the fluxes of intermediates of the

387

TCA cycle wherein all the major metabolites of TCA cycle (pyruvic, t-aconitic, citric, α-

388

keto glutaric, succinic and malic acid) responded to glyphosate application (Supporting

389

Information Table 4). At 8 HAT, a differential abundance in the concentrations of organic

390

acids was observed in the glyphosate treated S- and R-biotypes compared to the

391

respective water control, with the glyphosate-treated R-biotype exhibiting a substantial

392

increase in concentrations (Figure 5a). However, at 80 HAT, in the glyphosate-treated

393

S-biotype, except for succinate and α-ketoglutarate, all other metabolites of the TCA

394

cycle responded to glyphosate application with a drastic reduction in citrate. In contrast,

395

no significant perturbations of the TCA cycle were observed in the glyphosate-treated

396

R-biotype except for the reduction in the levels of citric acid, similar to that of the

397

glyphosate-treated S-biotype (Figure 5b). Overall, our results not only highlight the

398

robustness of metabolomics approach in identifying differential physiological responses

399

of S- and R-biotypes to glyphosate, but also identify metabolic-level differences

400

between the R- and S-biotypes in the absence of herbicidal stress.

401

Biochemical Assays


Page 19 of 44


19 402

The observed disruption of the oxidative pentose pathway and the energy-

403

producing tricarboxylic acid cycle, following glyphosate application, could result in

404

generation of reactive free radicals, which could initiate lipid peroxidation resulting in the

405

formation of the reactive compound malondialdehyde (MDA). Compared to the

406

glyphosate-treated R-biotype, the glyphosate-treated S-biotype accumulated 49%

407

higher MDA after 80 HAT (Figure 7a), indicating that higher levels of free radicals are

408

generated in the glyphosate-treated S-biotype. A similar observation of deleterious

409

effect of glyphosate on maize plants due to increased lipid peroxidation via MDA

410

production was reported by Sergiev et al.19 Thus, the death of the glyphosate-treated S-

411

biotype could be partially accelerated by the complementary action of free radicals that

412

result in observed lipid peroxidation.

413

Reduced levels of MDA accumulation in the R-biotype following glyphosate

414

application (Figure 7a) could indicate an apparent increase in the activities of enzymes

415

and metabolites which can effectively quench the free radicals. Plants have a well-

416

defined and extensive enzymatic and non-enzymatic anti-oxidant machinery to protect

417

them from the free radicals generated during normal physiological metabolism.50 In

418

plants, the glutathione reductase (GR) enzyme system serves as a major anti-oxidant

419

defence. GR catalyzes the reduction of oxidized glutathione disulfide (GSSG) to

420

glutathione (GSH), its anti-oxidant compound which neutralizes highly reactive ROS

421

molecules. Increased GR activity results in increased production of GSH, enhancing the

422

antioxidant potential. The reduction of GSSG to GSH is energy dependent and utilizes

423

NADPH as the energy supplier. Therefore the GR activity is captured by monitoring the

424

levels of NADPH29 wherein a decrease/increase in NADPH levels indicate a



Page 20 of 44

20 425

proportional increase/decrease in GR activity. At 8 HAT, compared to the water-treated

426

S-biotype, the glyphosate-treated S-biotype had a 26% decrease in NADPH levels

427

thereby indicating higher GR activity than in the non-treated control. At 8 HAT, in the

428

glyphosate treated R-biotype, there was an 85% increase in NADPH levels indicating

429

lower GR activity in the glyphosate-treated R-biotype compared to its non-treated

430

control (Figure 7b). However, at 80 HAT, compared to the water control R-biotype, there

431

was a 51% reduction in the NADPH levels in the GR activity of glyphosate-treated R-

432

biotype which potentially indicates higher GR activity in the glyphosate-treated R-

433

biotype. Contrastingly, at 80 HAT, compared to the water treated S-biotype, the NADPH

434

levels increased 37% in the glyphosate-treated S-biotype, indicating a decrease in

435

NADPH consumption. Lower GR activity (as indicated by higher NADPH levels) in

436

glyphosate-treated S-biotype at 80 HAT could indicate a decrease in synthesis of

437

glutathione by GR, resulting in lower anti-oxidant potential despite higher levels of

438

oxidative damage. Although it is possible that the lower utilization of NADPH in the S-

439

biotype could also be due to the cessation of carbon fixation induced by the glyphosate

440

toxicity, the concomitant higher accumulation of MDA in the glyphosate-treated S-

441

biotype at 80 HAT (Figure 7a) indicates a lower GR activity (Figure 7b). Thus, the

442

increased GR activity in glyphosate-treated R-biotype at 80 HAT indicates a responsive

443

but delayed GR-dependent antioxidant activity. Similar observations of the induction of

444

plant stress response genes following glyphosate application was shown in Festuca

445

arundinacea by Unver et al.51 following miRNA and transcriptome analysis.

446

Furthermore, transcriptome analysis of glyphosate-treated susceptible and resistant

447

soybean plants (Glycine max (L.) Merr.) by Zhu et al.52 had similar conclusions. Thus,


Page 21 of 44


21 448

the observed higher GR activity combined with the lower accumulation of MDA in

449

glyphosate-treated R-biotypes indicates the existence of a robust free radical

450

scavenging system in this biotype that could potentially complement the glyphosate

451

resistance conferred by EPSPS gene copy amplification.

452

The phenylpropanoid pathway is a major secondary metabolic pathway that

453

synthesizes a wide array of phenolic compounds (such as flavonoids and betalains) with

454

radical-scavenging capacity to reduce oxidative stress. As the rate limiting step of this

455

pathway is the conversion of phenylalanine to t-CA catalyzed by the phenylalanine

456

ammonia-lyase (PAL) enzyme, the rate of formation of t-CA could partially reflect the

457

antioxidant capacity of the plant. At 8 HAT extracted PAL activity was the same in the

458

control groups for both biotypes and in the treated R biotype (Figure 8a). The activity

459

from the treated S biotype was much higher at this time. However, at 80 HAT the PAL

460

activity was same in all treatments except for reduced activity in the treated S biotype

461

(Figure 8b). This is in accordance to previous studies which observed rapid increases

462

in PAL activity in glyphosate-treated plants that were transient, often peaking at about

463

24 HAT.53,54 In this previous work, extracted PAL activity was temporarily enhanced by

464

glyphosate treatment, while phenylpropanoid levels were decreased, suggesting that

465

the PAL gene(s) is up-regulated by blockage of the shikimate pathway.

466

Based on the metabolite profiles of the S- and R-biotype under both water and

467

glyphosate treatment, two phenylpropanoid secondary metabolites, ferulic and caffeic

468

acid, were identified. While ferulic acid was detected in the R-biotype in both the water

469

control and glyphosate treatments at 8 HAT and only in the glyphosate treatment at 80

470

HAT, no ferulic acid was detected in the S biotype in either water- or glyphosate-treated



Page 22 of 44

22 471

plants at both 8 and 80 HAT (Supporting Information Figure 2a). Similarly, at 8 HAT,

472

caffeic acid was detected in both the control and glyphosate-treated S- and R-biotypes,

473

but at 80 HAT, it was not detected in either control or glyphosate-treated S-biotype, but

474

was still present in the R biotype in all treatments at all times. Interestingly, while the

475

caffeic acid levels were elevated in the glyphosate treatment over the control at 8 HAT

476

and not at 80 HAT, the level of caffeic acid in the control was much higher at 80 than 8

477

HAT (Supporting Information Figure 2b). This suggests that due to the extra copies of

478

EPSPS gene in the R biotype, there could be an increase in the levels of metabolites

479

downstream of the shikimate pathway, including metabolites of the phenylpropanoid

480

pathway. In addition to phenolic acids, LC-MS profiling of flavonoids (Supporting

481

Information

482

phenylpropanoids,55 particularly 4-O-feruloylquinic acid (4-FQA), a known anti-oxidant,

483

in the R-biotype (Supporting Information Figure 3). This further supports our hypothesis

484

that the R-biotype could possess an enhanced anti-oxidant capability which may play a

485

complementary role in conferring resistance to glyphosate. The enhanced levels of

486

certain phenylpropanoid compounds in the R biotype may play a role in the different

487

spectral reflectivities of the two A.palmeri biotypes.56

Method

1)

identified

significantly

higher

abundance

of

certain

488

At 80 HAT, the control R-biotype contained 60% higher total soluble proteins

489

(TSP) compared to the S-biotype, and the TSP in both biotypes decreased 80 HAT

490

following glyphosate application. The R-biotype had about 35% reduction in protein

491

levels as compared to the S- biotype (Figure 9). This decrease in protein levels in the R-

492

biotype could be attributed to the initial inhibition of shikimate pathway as evident in the

493

accumulation of shikimic acid at 8 HAT. The potential transient inhibition of the


Page 23 of 44


23 494

shikimate pathway could lead to protein catabolism to sustain growth, which might partly

495

explain this observation. The reduced protein levels in the S-biotype could be attributed

496

to the combined effect of catabolism and termination of protein synthesis.

497

Our results demonstrate the application of metabolomics as a complementary

498

tool for further elucidating glyphosate resistance at the phenotypic level in A. palmeri.

499

Metabolite profiling differentiated physiological differences between S- and R-biotypes,

500

both in the presence and absence of glyphosate treatment. Though both the S- and R-

501

biotypes were exposed to identical environmental conditions under the two treatment

502

conditions (water or glyphosate), the metabolite profile patterns of water-treated

503

(control), plants within each biotype were not similar between the two harvest times.

504

This could be attributed to the rapid fluxes of the measured metabolites that were

505

altered in the intervening period between the two sampling time-point. Importantly, the

506

metabolic changes observed after glyphosate application captures the differences in

507

response to glyphosate application by the two biotypes. The resistance to glyphosate,

508

though primarily conferred by the multiple copies of EPSPS gene in this resistant A.

509

palmeri biotype, may be complemented by the anti-oxidative protective mechanisms.

510

Considering the recent findings that EPSPS gene expression is not equal in all tissue

511

types within a single individual of glyphosate resistant A.palmeri biotypes21, the

512

enhanced antioxidant potential might help to mitigate the secondary toxicity of

513

glyphosate generated through generation of free radicals in tissues with lower EPSPS

514

copies.

515

Abbreviations Used



Page 24 of 44

24 516

5-enolpyruvylshikimate-3-phosphate synthase (EPSPS); h after treatment (HAT);

517

ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO); 3-phosphoglyceric acid

518

(3-PGA); malondialdehyde (MDA); glutathione reductase (GR); oxidized glutathione

519

(GSSG); glutathione (GS); phenylalanine ammonia-lyase (PAL); total soluble proteins

520

(TSP); tricarboxylic acid cycle (TCA); phosphoenolpyruvate (PEP); pyrophosphate-

521

dependent

522

phosphate synthase (DAHP);.

phosphofructokinase

(PPi-PFK);

3-deoxy-D-arabinoheptulosonate

7-

Acknowledgement We thank Vidya Suseela for helping with the metabolomics data interpretation, proofreading and giving valuable comments on presenting the manuscript in its current form. Authors also acknowledge William Molin for providing constructive comments on a previous version of this manuscript. Supporting Information Available. Information contains the list of metabolites identified in water- and glyphosate-treated S- and R-biotypes of Amaranthus palmeri, statistical values for the metabolic responses to the treatments and the equations used to interpret the biochemical assays. This material is available free of charge via the Internet at http://pubs.acs.org. Funding Sources: Statistical approaches for this research was partly supported by NIH grant

5P206M10344

received


by

PG.

Page 25 of 44


25

References 1. Duke, S.O.; Powles, S.B. Glyphosate: a once-in-a-century herbicide. Pest Manag. Sci. 2008, 64, 319-325. 2. Dill, G.M. Glyphosate-resistant crops: history, status and future. Pest Manag. Sci. 2005, 61, 219-224. 3. Duke, S.O.; Baerson, S.R.; Rimando, A.M. Herbicides: Glyphosate. In Encyclopedia of Agrochemicals; Plimmer, J.R., Gammon, D.W., Ragsdale, N.N., Eds.; John Wiley & Sons, New York, 2003. 4. Sammons, R.D.; Gruys, K.J.; Anderson, K.S.; Johnson, K.A.; Sikorski, J.A. Reevaluating glyphosate as a transition-state inhibitor of EPSP synthase: Identification of an EPSP synthaseEPSP glyphosate ternary complex. Biochemistry 1995, 34, 6433-6440. 5. Heap, I. The international survey of herbicide resistant weeds. Online. URL. (http:// www.weedscience.org) (Monday, May 11, 2015). 6. Moore, J.W.; Murray, D.S.; Westerman, R.B. Palmer amaranth (Amaranthus palmeri) effects on the harvest and yield of grain sorghum (Sorghum bicolor). Weed Technol. 2004, 18, 23-29. 7. Morgan, G.D.; Baumann, P.A.; Chandler, J.M. Competitive impact of Palmer amaranth (Amaranthus palmeri) on cotton (Gossypium hirsutum) development and yield. Weed Technol. 2001, 15, 408-412.



Page 26 of 44

26 8. Délye, C. Unravelling the genetic bases of non-target-site-based resistance (NTSR) to herbicides: a major challenge for weed science in the forthcoming decade. Pest Manag. Sci. 2013, 69, 176-187. 9. Délye, C.; Jasieniuk, M.; Le Corre, V. Deciphering the evolution of herbicide resistance in weeds. Trends Genet. 2013, 29, 649-658. 10. Gaines, T.A.; Zhang, W.; Wang, D.; Bukun, B.; Chisholm, S.T.; Shaner, D.L.; Nissen, S.J.; Patzoldt, W.L.; Tranel, P.J.; Culpepper, A.S.; Grey, T.L.; Webster, T.M.; Vencill, W.K.; Sammons, R.D.; Jiang, J.; Preston, C.; Leach, J.E.; Westra, P. Gene amplification confers glyphosate resistance in Amaranthus palmeri. Proc. Natl. Acad. Sci. U.S.A 2010, 107, 10291034. 11. Nandula, V.K.; Reddy, K.N.; Koger, C.H.; Poston, D.H.; Rimando, A.M.; Duke, S.O.; Bond, J.A.; Ribeiro, D.N. Multiple resistance to glyphosate and pyrithiobac in Palmer amaranth (Amaranthus palmeri) from Mississippi and response to flumiclorac. Weed Sci. 2012, 60, 179188. 12. Patti, G.J.; Yanes, O.; Siuzdak, G. Innovation: Metabolomics: the apogee of the omics trilogy. Nat. Rev. Mol. Cell Biol. 2012, 13, 263-269. 13. Anderson, J.V. Emerging technologies: An opportunity for weed biology research. Weed Sci. 2008, 56, 281-282. 14. Molinari, H.B.C.; Marur, C.J.; Daros, E.; de Campos, M., Freitas, K.; de Carvalho, J.F., Portela R.; Filho, J.C.B.; Pereira, L.F.P.; Vieira, L.G.E. Evaluation of the stress-inducible production of proline in transgenic sugarcane (Saccharum spp.): osmotic adjustment, chlorophyll fluorescence and oxidative stress. Physiol. Plant. 2007, 130, 218-229.


Page 27 of 44


27 15. Suseela, V.; Triebwasser-Freese, D.; Linscheid, N.; Morgan, J.A.; Tharayil, N. Litters of photosynthetically divergent grasses exhibit differential metabolic responses to warming and elevated CO2. Ecosphere 2014, 5, art106. 16. Tzin, V.; Galili, G. New insights into the shikimate and aromatic amino acids biosynthesis pathways in plants. Mol. Plant 2010, 3, 956-972. 17. Servaites, J.C.; Tucci, M.A.; Geiger, D.R. Glyphosate effects on carbon assimilation, ribulose bisphosphate carboxylase activity, and metabolite levels in sugar beet leaves. Plant Physiol. 1987, 85, 370-374. 18. de María, N.; de Felipe, M.R.; Fernández-Pascual, M. Alterations induced by glyphosate on lupin photosynthetic apparatus and nodule ultrastructure and some oxygen diffusion related proteins. Plant Physiol. Biochem. 2005, 43, 985-996. 19. Sergiev, I.G.; Alexieva, V.S.; Ivanov, S.V.; Moskova, I.I.; Karanov, E.N. The phenylurea cytokinin 4PU-30 protects maize plants against glyphosate action. Pestic. Biochem. Physiol. 2006, 85, 139-146. 20. Ahsan, N.; Lee, D.; Lee, K.; Alam, I.; Lee, S.; Bahk, J.D.; Lee, B. Glyphosate-induced oxidative stress in rice leaves revealed by proteomic approach. Plant Physiol. Biochem. 2008, 46, 1062-1070. 21. Godar, A.; Koo, D.; Peterson, D.; Gill, B.; Jugulam, M. Stability of EPSPS gene copies in glyphosate-resistant Palmer amaranth (Amaranthus palmeri). Proc. Weed. Sci. Soc. Am. Annu. Meet. 2015, 220. 22. Pedersen, H.A.; Kudsk, P.; Fomsgaard, I.S. Metabolic Profiling of Arabidopsis thaliana reveals herbicide and allelochemical dependent alterations before they become apparent in plant growth. J. Plant Growth Regul. 2015, 34, 96-107.



Page 28 of 44

28 23. Trenkamp, S.; Eckes, P.; Busch, M.; Fernie, A.R. Temporally resolved GC-MS-based metabolic profiling of herbicide treated plants treated reveals that changes in polar primary metabolites alone can distinguish herbicides of differing mode of action. Metabolomics 2009, 5, 277-291. 24. García-Villalba, R.; León, C.; Dinelli, G.; Segura-Carretero, A.; Fernández-Gutiérrez, A.; Garcia-Cañas, V.; Cifuentes, A. Comparative metabolomic study of transgenic versus conventional soybean using capillary electrophoresis-time-of-flight mass spectrometry. J. Chromat. A. 2008, 1195, 164-173. 25. Serra, A.A.; Couée, I.; Renault, D.; Gouesbet, G.; Sulmon, C. Metabolic profiling of Lolium perenne shows functional integration of metabolic responses to diverse subtoxic conditions of chemical stress. J. Exp. Bot. 2015, 66, 1801-1816. 26. Lisec, J.; Schauer, N.; Kopka, J.; Willmitzer, L.; Fernie, A.R. Gas chromatography mass spectrometry-based metabolite profiling in plants. Nat. Protoc. 2006, 1, 387-396. 27. Fiehn, O.; Wohlgemuth, G.; Scholz, M.; Kind, T.; Lee, D.Y.; Lu, Y.; Moon, S.; Nikolau, B. Quality control for plant metabolomics: reporting MSI-compliant studies. Plant J. 2008, 53, 691– 704. 28. Shang, Q.M.; Li, L.; Dong, C.J. Multiple tandem duplication of the phenylalanine ammonialyase genes in Cucumis sativus L. Planta 2012, 236, 1093-1105. 29. Yannarelli, G.G.; Fernández-Alvarez, A.J.; Santa-Cruz, D.M.; Tomaro, M.L. Glutathione reductase activity and isoforms in leaves and roots of wheat plants subjected to cadmium stress. Phytochemistry 2007, 68, 505-512.


Page 29 of 44


29 30. Hodges, D.M.; DeLong, J.M.; Forney, C.F.; Prange, R.K. Improving the thiobarbituric acidreactive-substances assay for estimating lipid peroxidation in plant tissues containing anthocyanin and other interfering compounds. Planta 1999, 207, 604-611. 31. Anderson, M.J. A new method for non‐parametric multivariate analysis of variance. Austral. ecology. 2001, 26, 32-46. 32. Anderson, M.J. "Permutational multivariate analysis of variance." Department of Statistics, University of Auckland, Auckland, 2005. 33. Xia, J.; Mandal, R.; Sinelnikov, I.V.; Broadhurst, D.; Wishart, D.S. MetaboAnalyst 2.0--a comprehensive server for metabolomic data analysis. Nucleic Acids Res. 2012, 40, 127-133. 34. Szymańska, E.; Saccenti, E.; Smilde, A.K.; Westerhuis, J A. Double-check: validation of diagnostic statistics for PLS-DA models in metabolomics studies. Metabolomics 2012, 8, 3-16. 35a. Herrmann, K.M. The shikimate pathway: early steps in the biosynthesis of aromatic compounds. Plant Cell. 1995, 7, 907. 35b. Herrmann, K.M. The shikimate pathway as an entry to aromatic secondary metabolism. Plant Physiol. 1995, 107, 7-12. 36. Steckel, L.E.; Main, C.L.; Ellis, A.T.; Mueller, T.C. Palmer amaranth (Amaranthus palmeri) in Tennessee has low level glyphosate resistance. Weed Technol. 2008, 22, 119-123. 37. Mueller, T.C.; Massey, J.H.; Hayes, R.M.; Main, C.L.; Stewart, C.N. Shikimate accumulates in both glyphosate-sensitive and glyphosate-resistant horseweed (Conyza canadensis L. Cronq.). J. Agric. Food Chem. 2003, 51, 680-684. 38. Becerril, J.M.; Duke, S.O.; Lydon, J. Glyphosate effects on shikimate pathway products in leaves and flowers of velvetleaf. Phytochemistry 1989, 28, 695-699.



Page 30 of 44

30 39. Weaver, L.M.; Herrmann, K.M. (1997). Dynamics of the shikimate pathway in plants. Trends Plant Sci. 1997, 2, 346-351. 40. Görlach, J.; Raesecke, H.R.; Rentsch, D.; Regenass, M.; Roy, P.; Zala, M.; Keel, C.; Boller, T.; Amrhein, N.; Schmid, J. Temporally distinct accumulation of transcripts encoding enzymes of the prechorismate pathway in elicitor-treated, cultured tomato cells. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 3166-3170. 41. Wang, C. Effect of glyphosate on aromatic amino acid metabolism in purple nutsedge (Cyperus rotundus). Weed Technol. 2001, 15, 628-635. 42. Jaworski, E.G. Mode of action of N-phosphonomethylglycine. Inhibition of aromatic amino acid biosynthesis. J. Agric. Food Chem. 1972, 20, 1195-1198. 43. Gresshoff, P.M. Growth inhibition by glyphosate and reversal of its action by phenylalanine and tyrosine. Funct. Plant Biol. 1979, 6, 177-185. 44. Imaoka, A.; Ueno, M.; Kihara, J.; Arase, S. Effect of glyphosate on tryptamine production and Sekiguchi lesion formation in rice infected with Magnaporthe grisea. J. Gener. Plant Pathol. 2008, 74, 109-116. 45a. Duke, S.O.; Hoagland, R.E. Effects of glyphosate on metabolism of phenolic compounds I. Induction of phenylalanine ammonia-lyase activity in dark-grown maize roots. Plant Sci. Lett. 1978, 11, 185-190. 45b. Duke, S.O.; Hoagland, RE. Effects of glyphosate on the metabolism of phenolic compounds: VII. Root-fed amino acids and glyphosate toxicity in soybean (Glycine max) seedlings. Weed Sci. 1981, 29, 297-302.


Page 31 of 44


31 46. Plaxton, W.C. The organization and regulation of plant glycolysis. Ann. Rev. Plant Biol. 1996, 47, 185-214. 47. Smyth, D.A.; Min-Xian, W.; Clanton, C.B. Pyrophosphate and fructose 2,6-bisphosphate effects on glycolysis in pea seed extracts. Plant Physiol. 1984, 76, 316-320. 48. Orcaray, L.; Zulet, A.; Zabalza, A.; Royuela, M. Impairment of carbon metabolism induced by the herbicide glyphosate. J. Plant Physiol. 2012,169, 27-33. 49. Colombo, S.L.; Andreo, C.S.; Chollet, R. The interaction of shikimic acid and protein phosphorylation with pep carboxylase from the C4 dicot Amaranthus viridis. Phytochemistry 1998, 48, 55-59. 50. Gill, S.S.; Tuteja, N. Reactive oxygen species and antioxidant machinery in abiotic stress tolerance in crop plants. Plant Physiol. Biochem. 2010, 48, 909-930. 51. Unver, T.; Bakar, M.; Shearman, R.C.; Budak, H. (2010). Genome-wide profiling and analysis of Festuca arundinacea miRNAs and transcriptomes in response to foliar glyphosate application. Mol. Genet. Genomics 2010, 283, 397-413. 52. Zhu, J.; Patzoldt, W.L.; Shealy, R.T.; Vodkin, L.O.; Clough, S.J., Tranel, P.J. Transcriptome response to glyphosate in sensitive and resistant soybean. J. Agric. Food Chem. 2008, 56, 6355-6363. 53. Hoagland, R.E.; Duke, S.O.; Elmore, C.D. Effects of glyphosate on metabolism of phenolic compounds. III. Phenylalanine ammonia-lyase activity, free amino acids, soluble protein and hydroxyphenolic compounds in axes of dark-grown soybeans. Physiol. Plant. 1979, 46, 357366.



Page 32 of 44

32 54. Duke, S.O., Hoagland, R.E.; Elmore, C.D. Effects of glyphosate on metabolism of phenolic compounds. IV. Phenylalanine ammonia-lyase activity, free amino acids and soluble hydroxyphenolic compounds in axes of light-grown soybeans. Physiol. Plant. 1979, 46, 307-317. 55. Stintzing, F.C.; Kammerer, D.; Schieber, A.; Adama, H.; Nacoulma, O.G.; Carle, R. Betacyanins and phenolic compounds from Amaranthus spinosus L. and Boerhavia erecta L. Z. Naturforsch C. 2004, 59, 1-8. 56. Reddy, K.N.; Huang, Y.; Lee, M.A.; Nandula, V.K.; Fletcher, R.S.; Thomson, S.J.; Zhao. F. Glyphoste-resistant and glyphosate-susceptible Palmer amaranth (Amaranthus palmeri S. Wats.): hyperspectral reflectance properties of plant and potential classification. Pest Manag. Sci. 2014, 70, 1910-1917. Figure Captions Figure 1: PLS-DA Loading plots of water- and glyphosate-treated S- and R-biotypes at A) 8 HAT and B) 80 HAT and the corresponding score plots at C) 8 HAT and D) 80 HAT. The permutation cross validation of PLS-DA model for 8 and 80 HAT had P =0.0005 over 2000 iterations. The ellipse represent 95% confidence region. Figure 2: Heatmap and two-way hierarchical clustering of metabolites at A) 8 HAT and B) 80 HAT. Algorithm for Heatmap clustering was based on Euclidean distance measure for similarity and Ward linkage method for biotype clustering. ANOVA results comparing means of each metabolite across treatment groups, and false discovery rate (FDR) for multiple testing corrections are given in Supporting Information Table 3. Enlarged dendrogram is given in Figure S1. Figure 3: Shikimic acid abundance. The data represent the mean±SD of shikimic acid abundance in glyphosate-treated plants normalized with respect to ribitol abundance and


Page 33 of 44


33 corrected with water-treated shikimic acid abundance. Zero represents water-control abundance. Bars with the same letters are not significantly different at 0.05% confidence. Figure 4: Comparisons of aromatic amino acids across the biotypes and treatments at 8 and 80 HAT. The mean±SD was tested by Two-Way ANOVA with Tukey’s HSD comparing across all treatments and biotypes at a confidence level of 0.05%. Tukey’s comparison between the treatment means are given in Supporting Information Table 2. Figure 5: TCA cycle metabolite pools in water- and glyphosate-treated S- and R-biotype at A) 8 HAT and B) 80 HAT. The data represent the mean±SD of the metabolites. Tukey’s comparison between the treatment means are given in Supporting Information Table 4. Figure 6: Total sugar metabolites at A) 8 HAT and B) at 80 HAT. The data represent the mean±SD of the total sugar abundance normalized with respect to ribitol abundance. Bars with the same letters are not significantly different at 0.05% confidence. Figure 7: A) TBARS assay quantifying malondialdehyde equivalents in S- and R- biotypes across the treatments at 80 HAT. B) Glutathione reductase assay comparing difference in NADPH levels in glyphosate-treated S- and R- biotypes across 8 and 80 HAT. The data for TBARS assay represent the mean±SD of the assay and the data for GR assay represents water-control normalized mean±SD data of fold changes in NADPH levels. Bars with the same letters are not significantly different at 0.05% confidence. Figure 8: Phenylalanine ammonia lyase assay comparing difference in activities of the PAL enzyme in S- and R- biotypes across the treatments at A) 8 HAT and B) at 80 HAT. The data represent the mean±SD of the assays. Bars with the same letters are not significantly different at 0.05% confidence.



Page 34 of 44

34 Figure 9: BCA assay for total soluble protein (TSP) quantification in plants harvested at 80 HAT. The data represent the mean±SD of the assay. Bars with the same letters are not significantly different

at

0.05%


confidence.

Page 35 of 44


35

0.10

0.2

A

B

3-dehydroshikimate

0.05 0.1

Dehydroascorbic acid Glycine

-0.05

Succinate Benzoate

Fumarate

-0.10

Glycine α−ketoglutarate

Glutamine

-0.15 -0.20 -0.25

Glycolate

Glucose

Citrate

Porphine

Sucrose

Lysine Lactate Shikimate β−alanine

Xylitol Valine Malate Asparagine Urea Tryptophan Tagatose Leucine Glycerol Phosphoric acid Dihydroxyacetone Fructose Ferulic acid Pyruvate Stearate Lyxoslamine Phenylalanine Serine Proline Malonate t-aconitate Threonine Palmitate Tyrosine Adenine Glucose Gluconate Allantoin Caffeic acid Itaconate Alanine Guanidinobutyrate Oxalate Isoleucine Glycerate Aspartate Nicotinate Dehydroascorbate Methionine

0.0

Component 2

Component 2

0.00

Palmitate

Tyrosine Aspartate

Succinate Stearate Malate Citrate Shikimate Lactate Caffeic acid Asparagine Sucrose Glycolate Oxalate Trans-aconitate Porphine Threonine Dihydroxyacetone Isoleucine Nicotinate Phosphoric acid Serine Adenine Dehydroglutamate Lysine α−ketoglutarate Fructose Allantoin 3-dehydroshikimate Tagatose β−alanine Tryptophan Leucine Glycerol Urea Glutamine Alanine Guanidinobutyrate Proline Xylitol Valine Gluconate

-0.1

-0.2

Lyxosylamine

-0.3

-0.30

Ferulic acid

Malonate Itaconate Glycerate Pyruvate Benzoate

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

Component 1

Component 1

D

C

RC RG SC SG

Figure 1: PLS-DA Loading plots of water- and glyphosate-treated S- and R-biotypes at A) 8 HAT and B) 80 HAT and the corresponding score plots at C) 8 HAT and D) 80 HAT. The permutation cross validation of PLS-DA model for 8 and 80 HAT had P =0.0005 over 2000 iterations. The ellipse represent 95% confidence region.



Page 36 of 44

36 A

B

Figure 2: Heatmap and two-way hierarchical clustering of metabolites at A) 8 HAT and B) 80 HAT. Algorithm for Heatmap clustering was based on Euclidean distance measure for similarity and Ward linkage method for biotype clustering. ANOVA results comparing means of each metabolite across treatment groups, and false discovery rate (FDR) for multiple testing corrections are given in Supporting Information Table 3. Enlarged dendrogram is given in Figure S1.


Page 37 of 44


37

30

PTime < 0.001 PBiotype < 0.001 PInteraction < 0.001

20

8HAT 80HAT

10

0

B

A

S-Biotype

B

B

R-Biotype

Figure 3: Shikimic acid abundance. The data represent the mean±SD of shikimic acid abundance in glyphosate-treated plants normalized with respect to ribitol abundance and corrected with water-treated shikimic acid abundance. Zero represents watercontrol abundance. Bars with the same letters are not significantly different at 0.05% confidence.



38

Figure 4: Comparisons of aromatic amino acids across the biotypes and treatments at 8 and 80 HAT. The mean±SD was tested by Two-Way ANOVA with Tukey’s HSD comparing across all treatments and biotypes at a confidence level of 0.05%. Tukey’s comparison between the treatment means are given in Supporting Information Table 4.


Page 38 of 44

Page 39 of 44


39

Figure 5: TCA cycle metabolite pools in water- and glyphosate-treated S- and R-biotype at A) 8 HAT and B) 80 HAT. The data represent the mean±SD of the metabolites. Tukey’s comparison between the treatment means are given in Supporting Information Table 4.



Page 40 of 44

40

2000 1500

PTreatment = 0.016 PBiotype = 0.115 PInteraction = 0.005

80 HAT

**

1000 500 0

B

A

S-Biotype

B

B

R-Biotype

Figure 6: Total sugar metabolites at A) 8 HAT and B) at 80 HAT. The data represent the mean±SD of the total sugar abundance normalized with respect to ribitol abundance. Bars with the same letters are not significantly different at 0.05% confidence.


Page 41 of 44


41

A)

B)

Figure 7: A) TBARS assay quantifying malondialdehyde equivalents in S- and Rbiotypes across the treatments at 80 HAT. B) Glutathione reductase assay comparing difference in NADPH levels in glyphosate-treated S- and R- biotypes across 8 and 80 HAT. The data for TBARS assay represent the mean±SD of the assay and the data for GR assay represents water-control normalized mean±SD data of fold changes in NADPH levels. Bars with the same letters are not significantly different at 0.05% confidence.



Page 42 of 44

42

A)

B) 8h

80 h

Figure 8: Phenylalanine ammonia lyase assay comparing difference in activities of the PAL enzyme in S- and R- biotypes across the treatments at A) 8 HAT and B) at 80 HAT. The data represent the mean±SD of the assays. Bars with the same letters are not significantly different at 0.05% confidence.


Page 43 of 44


43

Figure 9: BCA assay for total soluble protein (TSP) quantification in plants harvested at 80 HAT. The data represent the mean±SD of the assay. Bars with the same letters are not significantly different at 0.05% confidence.



44

Table of Contents Graphic

RC RG SC SG


Page 44 of 44

Metabolic Profiling and Enzyme Analyses Indicate a Potential Role of

Recommend Documents