Biochimie (1992) 74, 875-882 © Soci6t6 franqaise de biochimie et biologie mol6culaire / Elsevier, Pads

875

Effect of glyphosate on plant cell metabolism, alp and 13C NMR studies E Gou0, R Bligny 2, P Genix 3, M Tissut 2, R Douce 2,3. ILaboratoire de R~sonance Magndtique en Biologie et M~decine; 2Laboratoire de Physiologie Cellulaire Vdgdtale, D~partement de Biologie moldculaire et Structurale, Centre d'Etudes Nucldaires et Universitd Joseph Fourier, 85X, 38041 Grenoble Cedex; -~Unitd mixte CNRS/RhOne Poulenc, RhOne Poulenc Agrochimie, 14-20 Rue Pierre Baizet, 69263 Lyon Cedex 09, France

(Received 27 May 1992; accepted 2 July 1992)

Summary m The effect of glyphosate (N-phosphonomethyl glycine; the active ingredient of Roundup herbicide) on plant cells metabolism was analysed by 31p and 13C NMR using suspension-cultured sycamore (Acer pseudoplatanus L) cells. Cells were compressed in the NMR tube and perfused with an original arrangement enabling a tight control of the circulating nutrient medium. Addition of 1 mM glyphosate to the nutrient medium triggered the accumulation of shikimate (20-30 lamol g-1 cell wet weight within 50 h) and shikimate 3-phosphate (1-1.5 lamol g-! cell wet weight within 50 h). From in vivo spectra it was demonstrated that these two compounds were accumulated in the cytoplasm where their concentrations reached potentially lethal levels. On the other hand, glyphosate present in the cytoplasmic compartment was extensively metabolized to yield aminomethylphosphonic acid which also accumulated in the cytoplasm. Finally, the results presented in this paper indicate that although the cell growth was stopped by glyphosate the cell respiration rates and the level of energy metabolism intermediates remained unchanged. glyphosate / shikimate / metabolism / herbicides / plant cells / 31p-13C NMR

Introduction The enzyme 5-enol-pyruvoylshikimate 3-phosphate synthase (EC 2.5.1.19) localized within the plastids [1], catalyzes the formation of 5-enol-pyruvoylshikimate 3-phosphate (EPSP) and phosphate from shikimate 3-phosphate and phosphoenolpyruvate in an unusual carboxyvinyl transfer reaction [2]. This enzyme, involved in the multibranched shikimic acid pathway leading to the synthesis of numerous aromatic compounds including phenylalanine, tyrosine and tryptophan [1], has been extensively studied because it is the target for N-phosphonomethyl glycine (glyphosate) [3], the active ingredient of Roundup herbicide, widely used for weed and vegetation control [4]. Glyphosate is absorbed and translocated in the phloem in both annual and perennial weeds [4]. Schulz et al [5] have shown that glyphosate induced in leaf mesophyll cells of tomato and spinach plants a marked accumulation of shikimate and shikimate 3phosphate. They also indicated that shikimate was preferentially localized in the vacuole whereas shikimate 3-phosphate accumulated in the chloroplasts. These interesting results prompted us to examine the *Correspondence and reprints

effect of glyphosate on the general metabolism of plant cells using 13C and 31p NMR.

Materials a n d m e t h o d s Material

Cell suspensions were chosen with a preference for dense tissues in order to improve the homogeneity of the incubation conditions (particularly the extra-cellular pH and the oxygen supply). Furthermore, with dense tissues such as maize root tips, external medium does not readily has access to the "nternal cells. In other words, the free diffusion of glyphosat¢ is hampered by the compactness of the tissue. The strain of sycamore (Acer pseudoplatanus L) used in the present study was grown as a suspension in a liquid nutrient medium according to Bligny [6] except Mn 2÷, excluded to prevent excessive broadening of the resonance of vacuolar compounds. The cell suspensions were maintained in exponential growth by frequent subcultures. The cell's weight was measured after straining culture aliquots on a glass fiber filter. Shikimate-3-phosphate was obtained from Transgene (Strasbourg, France). Petz'hloric extract

For perchloric acid extraction, cells (9 g wet weight) were quickly frozen in liquid nitrogen and ground to a fine powder with a mortar and pestle with 1 ml of 70% (v/v) perchloric acid. The frozen powder was then placed at -10°C and thawed.

876 The thick suspension thus obtained was centrifuged at 10 000 g for 10 rain to remove particulate matter, and the supernatant was neutralized with 2 M KHCO3 to about pH 6.5. The supernatant was then centrifuged at 10 000 g for 10 min to remove KCIO4; the resulting supematant was lyophilized and stored in liquid nitrogen. For the NMR measurements, this freeze-dried material was redissolved in 2.5 ml water containing 10% D20 (perchloric acid extract) and neutralized at pH 7.5 in the presence of 1 mM trans-l,2 diaminocyciohexane-N,N,N',N'tetraacetic acid (CDTA) for t3C NMR spectra and about 60 mM CDTA for 3tp NMR spectra. The laC and 3tp NMR spectra of neutralized perchloric acid extracts were measured on a Bruker NMR spectrometer (AMX 400, narrow bore) equipped with a 10-mm multinuclear probe. For t3C NMR spectra (100.6 MHz) acquisition used 18-~ pulses (60 °) at 4-s intervals (spectra of carboxyl groups were unchanged when a pulse interval of 10 s was employed). Two levels of proton decoupling were used: 2.5 W during the data acquisition (0.54 s) and 0.4 W during the delay period (3.46 s). Spectra were acquired over a period of 2 h (1800 scans). Free induction decays were accumulated using 16 K data points and zero-filled to 32 K prior to Fourier transformation. A l-Hz line broadening was applied. Chemical shifts were obtained by reference to the hexamethyldisiloxane (HDMSO) resonance at 2.7 ppm. Spectra of standard solutions of known carbon compounds at pH 7.5 were compared with that of a perchloric acid extract of sycamore cells. For 31p NMR spectra (161.93 MHz) acquisition used 10-1as pulses (50 °) at 1.8-s intervals. The deuterium resonance of D,O was used as a lock signal, and the spectra were recorded over a 2-h period under conditions of broad-band proton decoupling. 4048 scans (repetition time 1.8 s) were taken at a sweep-width of 6000 Hz (acquisition time 0.68 s). An exponential multiplication (0.5 Hz line width) was used to increase signal-to-noise ratio. Perchloric acid extract are referenced to methylphosphonic acid (MeP) (pH 8.9) at 23.7 ppm. The definitive assignments were made alter running a series of spectra obtained by addition of the authentic compounds to the perchloric acid extracts. in vivo NMR measurements In order to get a better signal-to-noise ratio an experimental arrangement was realized to analyze the maximum cell volume and to optimize the homogeneity of the cell incubation conditions. Cells (9 g wet wt) were slightly compressed by hand between two polymer filters to a volume of 17-18 ml [7] and perfused under slight pressure at a flow rate of 50 ml min-t with a well oxygenated (Oz bubbling) nutrient medium (Mn - free culture medium containing 100 laM phosphate) circulated via a 4-1 reservoir [71. The pH of this circulating medium (pile) was controlled after the passage through the compressed cells and maintained constant in the reservoir using a pH-stat coupled to a titrimeter (Urectron 6, Tacussel, France) monitoring the addition of registered amounts of HCI or KOH. The laC and 3tp NMR spectra were measured on a Bruker spectrometer (AMX 400, wide bore) equipped with a dual t3C and alp 25-mm probe tuned at 100.62 MHz for I3C and 161.93 MHz for 31p. For 13C NMR spectra acquisition used 30-gs pulses (60 °) at 4-s intervals (spectra of carboxyi groups were not significantly changed when longer intervals were chosen). Two levels of decoupling (Waltz pulse sequence) were used: 9.0 W during the data acquisition (0.38 s) and 0.5 W during the delay period (3.64 s). Spectra were acquired during periods in the 2-h range. Free induction decays were accumulated using 16 K data points and zero-filled to 32 K prior Fourier transformation. A

5 Hz line broadening was applied. Chemical shifts were obtained by reference to HDMSO resonance at 2.7 ppm. Spectra of standard solutions of known carbon compounds at pH 7 were compared with that of sycamore cells as described above. The definitive assignments were made after running a series of spectra obtained by addition of the authentic compounds to the perchloric acid extracts [8]. For 31p NMR spectra acquisition used 30-Iris pulses (50 °) at 0.6-s intervals. A 20-s recycling time was used to obtain fully relaxed spectra necessary for quantitative measurements but for qualitative purposes relative variations of Pi or phosphate esters were measured from free induction decays (FIDs) recorded with 4 K data points and zero-filled to 8 K prior to Fourier transformation. Spectra were referenced at 16.32 ppm to a solution of 50 mM methylene diphosphonic acid (pH 8.9 in 30 mM Tris) contained in a 0.8-mm capillary, itself inserted inside the inlet tube along the symmetry axis of the tissue sample (see [7]; for the introduction to high-resolution NMR spectroscopy and its application to in vivo and in vitro studies, see [9, 101), Chemical shifts were obtained by reference to MeP resonance at 23.7 ppm. The assignment of Pi, phosphate esters and nucleotides to specific peaks was carded out according to Navon et al [l 1], Evans and Kaplan [121 and Roberts and Jardetzky [9] and from spectra of the perchloric acid extracts that contained the soluble low molecular weight constituents. Finally, when given, intracellular concentrations were calculated on the following basis: 1 g cell wet weight corresponds to 1 ml cell volume and roughly 0.16 ml cytoplasm and 0.8 ml vacuole.

Results Growth of sycamore cells was 90% inhibited by glyphosate concentrations of 1 mM. However, growth was not impaired at 0.2 m M (not shown). This suggests an I~0 for growth inhibition of sycamore cells of between 0.2 and ! raM. We have, therefore, examined the metabolic consequences of impaired aromatic amino acid biosynthesis in sycamore cells treated with 1 m M glyphosate. t~C and 31p NMR o f neutralized perchloric acid extracts Figure 1 illustrates the changes that occur in the 3~p N M R spectra o f perchloric extracts obtained from sycamore cells treated with 1 m M glyphosate. In this experiment the cells were preincubated in their culture medium in the presence of 1 m M glyphosate during 50 h. The perchloric extracts were prepared from 9 g of glyphosate-treated and non-treated sycamore cells as described in Materials a n d methods. Under these conditions the observed resonance peaks were sharp, The major peaks were characterized by the addition o f known amounts of phosphorylated compounds to the extracts including shikimate 3-phosphate. The assignment of each resonance is given in the legend to figure 1. The peak at 4.4 ppm close to that o f mannose 6-phosphate (4.5 ppm) and glucose 6-phosphate (4.6 ppm) and present only in glyphosate-treated cells was

877 assigned to shikimate 3-phosphate. We must point out that the exact position of this resonance peak was highly sensitive to pH and to the chemical environment, making its correct identification difficult. Since shikimate 3-phosphate was partially obscured by the resonances of glucose 6-phosphate and mannose 6phosphate we repeated the same experiment using sucrose-starved cells instead of normal cells. In this experiment cells were incubated in 200 ml of culture medium devoid of sucrose and containing 1 mM glyphosate for 50 h. As expected (see [7]) after 50 h of sucrose starvation, the cells showed small vestigial glucose-6-P and UDP-glucose resonances whereas cytoplasmic phosphorylcholine increased considerably (fig 1B, B1). Indeed, it has previously been shown that when almost all the intracellular carbohydrate pools had disappeared the cell fatty acids deriving from membrane polar lipids such as phosphatidylcholine were utilized as oxidizable substrates for ATP production, whereas cytoplasmic phosphorylcholine increased symmetrically and was not further metabolized (for a review see Douce et al [13]). However, under these conditions, the resonance of shikimate 3-phosphate molecules was clearly distinguishable (fig IB) (shikimate 3-phosphate first became detectable in the perchloric extract only from 0.2 mM glyphosate). The intracellular level of shiMmate 3-P raised up to 1---1.5 lamol g-! cell wet weight within 50 h in sucrose-depleted cells. Surprisingly, the nucleotide region including 7-, ¢t-, and 13-ATP; 7-, ~-, and [3-UPT; or-, and [3-UDPG; 0~- and 13-UDP-gal phosphorus resonance peaks, was not altered by the

Fig 1. Proton-decouplcd 31p NMR spectra of perchioric acM extracts of sycamore cells. Perchloric extracts were prepared from 9 g (wet weight) of oxygenated cells as described in the text. Divalent cations (particularly Mn2* and Mg2+) were chelated by the addition of sufficient amounts of CDTA (ranging from 50--150 lamol depending on the cultures conditions). The samples (2.5 ml) containing 250 ~ of 2H20 were analyzed for 2 h at 25°C with a 10-mm multinuclear probe tuned at 161.93 MHz and the spectra were obtained as described under Materials and methods. Before perchloric acid extraction cells were incubated during 50 h in the following nutrient solutions: standard culture medium (A); culture medh~m contaita~ng ! mM glyphosate (AI); sucrose-free :;ulture medium (B); sucrose-free culture medium containing 1 mM glyphosate (B1). The phosphomonoesters area is shown on an expanded scale. Note the accumulation of shikimate 3-P when glyphosate is present in the culture medium and the accumulation of phosphorylcholine under sucrose-deprived culture conditions (for an explanation see Roby et al [7]). Peak assignments: Shik-3-P, shikimate-3-P, GIc-6-P, glucose-6-P; P-choline, phosphorylcholine; Man-6-P, mannose-6-P; NTP, nucleotides triphosphate (mainly ATP and UTP); NDP, nucleotides diphosphate; UDPG, uddine-5'-diphosphate-~-o-glucose; X, aminomethylphosphonic acid; Y, glyphosate.

presence of 1 mM glyphosate in the culture medium (fig 1A, A1). Indeed, we have observed that the presence of 1 mM glyphosate did not affect the respiration of sycamore cells (not shown). The resonance at 7.8 ppm (peak Y) was attributable to glyphosate. A separate peak at 9.5 ppm (peak X), tentatively assigned to a degradation product of glyphosate, has also been observed (fig 1A1, B1). This peak, which was slightly downfield from the peak assigned to glyphosate, responded to alkalinization and to acidification of the perchloric extract and was attributable to aminomethylphosphonic acid. The basis for this assignment was that the NMR parameters agreed perfectPi

Shik-3-P

B1

Shik-3-P

Phytate 5

XY

B

P_choline

4

P-choline

f.-NTP~

A

GIc-6-P

UDPG

!t

I I

,

I

I

)vl A

I! [; I'-NDP \\1 U o-NDP

I

l

I

I

I

I

I

•

,,/~,L...~. 5 4 ppm I

I

0 Chemical shift (,9, ppm

I

I

-20

.

~

878 Shik .Shik

I!G,o ....

Shikl

Shik

I I Glu

i"

IGlu

1

81

c,

I

I ma, Asr, I IL ILl

~

Asn

B

Shik

s, s,

c~t

~

/

Chemical shift

:'

,

c!t

'

A

ppm

Fig 2. Proton-decoupled 13C NMR spectra of perchloric acid extracts of sycamore cells. Perchloric extracts were prepared from 9 g (wet weight) of oxygenated cells as described in the text. The resolution of the carboxyl group was improved after adding 2 pmol CDTA to the extracts. The samples (2.5 ml) containing 250 l.tl of 2H20 were analyzed for 2 h at 25°C with a 10-mm multinuclear probe tuned at 100.6 MHz and the spectra were obtained as described under Materials and methods. Before perchloric acid extraction cells were incubated during 50 h under the conditions (A, AI, B, BI) given in the legend to figure 1. Note the accumulation of shikimate (Shik) and glutamine (Gin) when glyphosate is present in the culture medium and the accumulation of asparagine (Asn) under sucrose-deprived culture conditzons (far an explanation see Genix et al [8]). Peak assignments: Glu, glutamate; cit, citrate; mal, malate; s, sucrose; f, fructose; g, glucose.

ly with those we measured for aminomethylphosphonic acid. The increase in metabolism effect with time resulted in a higher proportion of aminomethylphosphonic acid at the expense of intact glyphosate (data not shown). 13C NMR spectra obtained from perchloric acid extract of sycamore cells (9 g wet weight) at pH 6.5 showed that the major resonances in the natural-abundance spectra corresponded to those of sucrose, citrate, and malate sequestered in the vacuole (see [8]) (fig 2A). The resonances of highest intensity corresponded to those of the glucosyl and fructosyl moieties of sucrose [14] and were estimated to correspond to an intracellular abundance of approximately 70 Imaol g-~ wet weight, in good agreement with biochemical determinations [15]. In addition, the spectra showed resonance peaks arising from the natural abundance of ~3C in D-glucose and fructose. Signals from citrate (intracellular concentration: 30 mM; centered at 182.3, 179.5, 76.4 and 45.5 ppm) and malate (intracellular concentration: 5 mM, centered at 181.7, 180.6, 71.2, and 43.4 ppm) were well characterized (fig 2). Signals from ),-amino butyrate centered at 35 and 40.5 ppm as previously observed in sycamore cells [8] did not emerge from the background noise. This compound, which derives from glutamate decarboxylation, accumulated during the first hour of anaerobiosis and appeared, therefore, systematically in the perchloric extract obtained from compact cells. On the other hand, following oxygenation the intracellular concentration of ),-amino butyrate dropped considerably (results not shown). Indeed there is much evidence that anoxia brings about marked changes in the levels of y-amino butyrate [ ! 6]. After 50 h of glyphosate (1 mM) treatment, several important changes occurred in the perchloric acid extract of the cells (fig 2A l). Of particular interest was the appearance of large amounts of shikimate. The definitive assignment of resonance peaks to shikimate centered at 175.9 (carboxyl group), 136.6 (C1 carbon atom), 131.2 (C2 carbon atom), 72.8 (C6 carbon atom), 67.6 (C3 carbon atom), 67.1 (C5 carbon atom), and 33.4 (C4 carbon atom) ppm was made after running a series of spectra obtained by addition of the authentic compound to the perchioric extracts at various pHs. There were was also signals from glutamine at 178.6, 174.9, 55.1, 31.6 and 27.0 ppm (not detectable in the control cells) and glutamate at 181.9, 175.3, 55.6, 34.2 and 27.8 ppm. Glutamine accumulation only occurred in glyphosate-treated cells and its final intracellular cell level was approximately 1015 larnol g-! wet weight. This spectrum also exhibited a multitude of resonance peaks emerging from the background noise (all the carbon compounds present in the perchloric acid extract at concentrations lower than 500 lalVl were not distinguishable from the back-

879 ground noise in our experimental conditions; see Materials a n d methods) which were ascribed to amino acids including aromatic amino acids (the presence of these amino acids is probably the result of protein turnover). Since shikimate peaks were partially obscured by the resonances of sucrose peaks we have also repeated the same experiment using sucrosestarved cells instead of normal cells (see [15]). In this experiment cells were first incubated in 200 ml of medium devoid of sucrose and containing 1 mM glyphosate for 50 h and then analysed by 31p NMR. Under these conditions the perchloric extract showed small vestigial sucrose (fig 2B, B l) and the resonances of all the carbon atoms of shikimate molecules were clearly distinguishable (fig 2B1). Of special interest after 50 h of sucrose-deprivation in the presence of glyphosate was the marked decrease in the amount of glutamine and the increase in the amount of asparagine centered at 35.6 and 52.3 ppm.

calculated that within the receiving coil the cell volume comprised approximately 50% of the total volume). Since the chemical shift of the carboxyl group of shikimate is sensitive to pH below 6, ~3C NMR can discriminate the vacuolar shikimate pool at the acidic pH from the cytoplasmic shikimate pool at the slightly alkaline pH. A close examination of an expanded scale of the spectra between 174 and 184 ppm (fig 3A1, B l) and titration curves plotting chemical shift versus pH for shikimate in solution in crude cell extract (supernatant obtained after a 10-min centrifugation at 18 000 g following disruption of cells by freeze-thawing) (insert, fig 3) indicated that the posit~o;~ of carboxyl group corresponds to shiki-

Shik

pH ~

Shikimate

ref.

13C and 3tp N M R o f compressed cells

~3C NMR obtained from slightly compressed cells under aerobic conditions showed that the resonances of shikimate were clearly discernable. In these experiments the cells were maintained for more than 50 h in a continuously oxygenated circulating solution (pH 6.5) with (fig 3A, A1) and without (fig 3B, B l) sucrose and containing (fig 3A1, B 1) or not (fig 3A, B) 1 mM glyphosate. To determine accurately the internal ceP shikimate concentration a calibration of the peak intensity of the shikimate resonances with known amounts of external shikimate was first performed (not shown). The curve thus generated gave estimates of shikimate cell levels of 20--30 lamol g-~ wet weight after 50 h (in our experimental conditions, we have Fig 3. In vivo proton-decoupled 13C-MR spectra (100.6 MHz) of compressed sycamore cells. The cells (9 g wet weight) were packed in a 25-mm NMR tube as described under Materials and methods and continuously perfused with a well-aerated Mn-free culture medium maintained at pH 6.5. The cell volume comprised about 50% of the total (cell + perfusing medium) analyzed volume. Spectra recorded at 20°C with a 60 ° pulse angle and a 4-s repetition time are the result of 1800 transients. Data treatments include Gaussian multiplication and zero filling. Before NMR analysis cells were incubated during 50 h under the conditions (A, AI, B, B1) given in the legend to figure 1. The carboxyl group area is shown on an expanded scale. Insert: chemical shift of shikimate versus pH. Peak assignments: Shik, shikimate; Glu, glutamate; ref, reference; Asn, asparagin.:'; Cit, citrate; S, sucrose; mal, malate. The reference at 2.7 ppm corresponds to HDMSO. The reference at 37.5 ppm corresponds to the methyl groups of the reference used for in vivo (31p) spectra.

,

I I

180

I/

I | I

176 t ..> /Shik,

176

175

Chemical shilt Shik (ppm} / 1~ : I Glu

Shik . ] ret;iG lu

B It I 11~0

I r-

~

;

1"/6 I

cit Asn

i

cit

As.

I

, I I I111 I1Gin 1 -,&,ol

I

.....

'*°-i ,s.,. 8

f ~j r

~

'w S

l

A ,

, ~

180

,

cit

'/real

~

150

100

Chemical shift

50

ppm

880

I Vac--Pi

p.8t ~ Shikimate3-P]

"

g'

~'

$

"~i

Chemicalshift(ppm}I

6 I ,.-NTPI'Nw! UDPG, f/,-NTP _. I, ,I '

'

'

'

Chemical

shift

(,~),ppm-

'

i

Fig 4. in vivo 31p spectra (161.93 MHz) of compressed sycamore cells. The cells (9 g wet weight) were packed in a 25-mm NMR tube as described under Materials and methods and continuously peffused with a well-aerated Mnfree culture medium maintained at pH 6.5. The cell volume comprised about 50% of the total (cell + peffusing medium) volume. Spectra recorded at 20°C using a multinuclear NMR probe with a 50° analyzed pulse angle and a 0.6-s repetition time. Each spectrum is the result of 12 000 transients (2 h). Data treatments include Gaussian multiplication and zero filling. Before NMR analysis cells were incubated during 50 h under the conditions (A, AI, B, B1) given in the legend to figure 1. Insert: chemical shift of shikimate-3-P versus pH. Peak assignments: Cyt-Pi, cytoplasmic phosphate; Vac-Pi, vacuolar Pi; a, complex peak including phosphoglyceric acid, fructose-6-P, ribose-5-P, and phosphorylethanolamine; X, aminomethylphosphonate; Y, glyphosate. mate at pH above 6.5. In addition, following perfusion during 30 min with 1 mM MnCI2 we observed a severe line broadening and loss of intensity of the

vacuolar Pi and citrate resonance peaks whereas under the same conditions, the cytoplasmic Pi and nucleotides resonance peaks were still apparent. Indeed, the strongly paramagnetic Mn2+ ions which are rapidly transported across the tonoplast, did not accumulate in sycamore cell cytoplasm [17]. Based on the chemical shift of the cytoplasmic and vacuolar Pi peaks, the pH values remained at 7.5 and 5.7, respectively. Since the line widths of the shikimate were essentially unchanged in the presence of Mn 2+ this result indicates that shikimate molecules did not accumulate in the vacuolar compartment. These surprising results suggest, therefore, that shikimate accumulated in the cytoplasmic compartment and not in the vacuole in contrast with the results of Schulz et al [5] on tomato and spinach plants after fractionation of freezestopped material in non-aqueous media. In considering the total intracellular shikimate concentration (approximately 30 gmol g-~ cell wet weight) and the ratio of vacuolar volume to cytoplasmic volume (roughly 5) [7], the final concentration attained in the cytoplasm was considerable: 100-150 mM including shikimate present in cytosol and various cell organelles. 31p NMR spectra obtained from compressed glyphosate-treated sycamore cells supplied with 02 and at pH 6.5 did not show a distinct peak of intracellular shikimate 3-phosphate. In fact, as shown above (fig 1), it is very likely that shikimate 3-phosphate is obscured by several resonance peaks in the region of the spectrum to highfield of glucose-6-P (fig 4A, A l, peak a). Since sucrose starvation led to a marked decline in the concentration of intracellular phosphate-ester including glucose-6-P and glycerate-3-P [7] we decided to follow shikimate 3-phosphate accumulation by sycamore cells during the course of sucrose starvation in the presence of I mM glyphosate. Figure 4B shows a typical 31p NMR spectrum of sucrose-starved sycamore cells (compare with fig 1). In this experiment the cells were maintained for more than 50 h in their nutrient medium devoid of sucrose and containing (fig 4B) or not 1 mM glyphosate (fig 4B1). Under these conditions, signal from shikimate 3-phosphate centered at 4.4 ppm was well characterized in glyphosate-treated cells (fig 4B l). Titration curves plotting chemical shift v e r s u s pH for shikimate 3-phosphate in solution in crude cell extract (insert, fig 4) indicated that the position of the peak at 4.4 ppm corresponds to shikimate 3-phosphate at pH 7.5. This suggests that glyphosate triggered the accumulation of shikimate 3phosphate in the cytoplasmic compartment. Likewise, titration curves plotting chemical shift v e r s u s pH for glyphosate and its degradation product aminomethylphosphonic acid in solutions of various composition indicated that the position of peaks X and Y (fig 4) corresponds to aminomethylphosphonic acid and glyphosate, respectively, at pH 7.5. This suggests that

881 glyphosate and its degradation product accumulated in the cytoplasmic compartment and are not expelled into the vacuole. The final concentration of aminomethylphosphonic acid attained in the cytoplasm is far from being negligible: to determine accurately the internal cytoplasmic aminomethylphosphonic concentration a calibration of the peak intensity of the aminomethylphosphonic acid resonance with known amounts of external aminomethylphosphonic acid was first performed (not shown; for the calculation see [7]). The curve thus generated gave estimates of cytoplasmic aminomethylphosphonic acid levels of 0.30.7 raM. Interestingly, we have observed that the addition of this compound to the perfusion medium did not trigger the accumulation of shikimate or shikimate 3-phosphate. Discussion

These results demonstrate that incubation of sycamore cells into flasks containing 1 mM glyphosate triggers the accumulation of shikimate 3-phosphate and shikimate in the cytoplasmic compartment and are consistent with the known mode of action of glyphosate, namely inhibition of EPSP synthase [3, 18]. Part of the difficulty in observing the visible pool of shikimate 3-phosphate in glyphosate-treated cells was the presence of nearby peaks (glucose 6-phosphate and possibly 3-phosphoglycerate) which partially overlap the cytoplasmic shikimate 3-phosphate resonance signals. The present study demonstrates that it is possible to better resolve shikimate 3-phosphate peak if 3~p NMR spectra are collected from sucrose-starved cells. The final concentration of shikimate attained in the cytoplasmic compartment was considerable and this surprising result raises the problem of the regulation of shikimic acid pathway in higher plants. The dramatic accumulation of shikimate is partly attributable to the fact that when mid-log phase plant cells are exposed to glyphosate, the specific activity of 2dehydro-3-deoxy-D-arabino-heptulonate 7-phosphate synthase (EC 4.1.2.15) increases several folds within 24 h [ 19]. Our results strongly suggest that 2-dehydro3-deoxy-D-arabino-heptulonate 7-phosphate synthase, 3-dehydroquinate synthase (EC 4.6.1.3), 3-dehydroquinate dehydratase (EC 4.2.1.10), and shikimate dehydrogenase (EC 1.1.1.25) involved in the shikimate pathway are not feedback-inhibited by shikimate and shikimate 3-phosphate. In other words, these two important metabolites are not involved in the control of carbon flow between phosphoenolpyruvate (+erythrose-4-phosphate) and shikimate 3-phosphate. The accumulation of shikimate-phosphate is ascribed to a direct inhibition of EPSP synthase, the penultimate enzyme of the chorismate pathway, by this potent

non-selective postemergence herbicide. On the other hand, shikimate can derive either from the removal of the phosphate group of shikimate 3-phosphate by an unspecific phosphatase or from a direct effect of shikimate 3-phosphate on shikimate kinase (EC 2.7.1.71) leading to an accumulation of shikimate. Whether or not high concentrations of shikimate and shikimate 3phosphate in the cytoplasmic compartment may exert secondary growth inhibition remains to be explored. The data reported here also raise the problem of the stability of shikimate in plant cells. According to Tateoka [20, 21] a dehydroshikimate hydrolase may initiate a series of reacuons which recycle shikimate via dehydroshikimate into catabolic pathways ultimately yielding CO2. In contrast, our results indicate that in the presence of glyphosate shikimate exhibits a high metabolic inertness because in the absence of sucrose, the blockage by glyphosate of the carbon flux through the shikimate pathway also led to a marked accumulation of shikimate which was not further metabolized (parenthetically it is surprising to note that the absence of sucrose in the perfusion medium did not stop the flux of carbon through the shikimate pathway at least during the first hours of sucrose starvation). Under these conditions, elevated intracytoplasmic levels of shikimate triggered by glyphosate inhibition of EPSP synthase might be toxic. The metabolic mayhem caused by shikimate imbalances, therefore, merits study. The results presented in this article also indicate that the inhibition of EPSP synthase by glyphosate triggered a marked accumulation of intracellular glutamate and glutamine. Since glutamate is often the amino-donor substrate in biosynthetic transaminations leading to the synthesis of numerous amino acids [22] it might be expected that the concentration of glutamate attained under steady state conditions is strongly dependent on the respective activities of enzymes of glutamate formation (glutamate synthase and glutamine synthetase) and aminotransferases involved in glutamate for protein synthesis. It is thus reasonable to expect that the inhibition of aromatic amino acid synthesis might elicit a marked increase in the intracellular concentration of glutamate and glutamine because no net gain of protein (and hence growth) could occur under these circumstances. In support of this suggestion, preliminary experiments carded out in our laboratory indicate that the inhibition of branchedchain amino acids by a sulfonylurea (chlorsulfuron) also leads to an accumulation of intracellular glutamate and glutamine (see also [23]). The results presented in this article also demonstrate that glyphosate present in the cytoplasmic compartments is extensively metabolised to yield aminomethylphosphonic acid. These results confirmed the findings of previous studies with Equisetum arvense

882

[24] and Eiymus repens [25] but contrasted with the results obtained from other plant species in which [~4C]glyphosate remained intact over short periods [26, 27] or was metabolised to only a limited extent [28, 29]. The mechanism involved in the degradation of glyphosate is unclear. It is possible, however, that aminomethyl phosphonic acid along with glyoxylate are the products of the reaction catalysed by monoamine oxidase (amine:oxygen oxidoreductase(deaminating) a flavoprotein (FAD) acting on glyphosate (a secondary amine). In support of this suggestion we have observed that O_, is required for the metabolism of glyphosate. Further research is required to examine the metabolic fate of glyphosate once it is incorporated into the cytoplasmic compartment. Finally our results highlight the importance of nuclear magnetic resonance spectroscopy to study the primary biochemical alterations induced by the blockage of an important metabolic pathway and the modes by which various herbicides inhibit cellular metabolism.

References I

2 3

4

5

6 7

8 9 10

Gilchrist DG, Kosuge T (1980) Aromatic amino acid biosynthesis and its regulation. In: The Biochemistry of Plants, vol 5. Amino acids and derivatives (Miflin BJ, ed) 507-531, Academic Press, New York Levin JG, Springson DB (1964) The enzymatic formation and isolation of 3-enolpyruvylshikimate 5-phophate. J Bioi Chem 239, 1142-1150 Steinrticken HC, Amrhein N (1980) The herbicide glyphosate is a potent inhibitor of 5-enolpyruvylshikimic acid 3phosphate synthase. Biochem Biophys Res Commun 94, 1207-1212 Sprankle P, Meggitt WE Penner D (1975) Absorption, translocation, and metabolism of 2,4-D and glyphosate in hemp dogbane (Apocynum cannabinium) Weed Sci 23, 235-240 Schulz A, Munder T, Hollander-Czytko H, Amrhein N (1990) Giyphosate transport and early effects on shikimate metabolism and its compartmentation in sink eaves of tomato and spinach plants. Z Natmforsch 45c, 529-534 Bligny R (1977) Growth of suspension-cultured Acer pseudoplatamts L cells in automatic units of large volume. Plant Physio160, 675--679 Roby C, Martin JB, Biigny R, Douce R (1987) Biochemical changes during sucrose deprivation in higher plant cells II. Phosphorus-31 nuclear magnetic resonance studies. J Biol Chem 262, 5000-5007 Genix P, Bligny R, Martin JB, Douce R (1990) Transient accumulation of asparagine in sycamore cells after a long periode of sucrose starvation. Plant Physio194, 7 ! 7-722 Roberts JKM, Jardetzky O (1981) Monitoring of cellular metabolism by NMR. Biochim Biophys Acta 639, 53-76 Pfeffer PE, Gerasimowicz WV (1989) Introduction of high-resolution NMR spectroscopy and its applications to in vivo studies of agricultural systems. In: Nuclear Magnetic Resonance in Agriculture (Pfeffer PE, Gerasimowicz WV, eds) CRC, Press, Boca Raton, FL, 3-70

11 Navon G, Shulman RG, Yamane T, Eccleshall TR, Lam KB, Baronofsky JJ, Marmur J (1979) Phosphorus-31 nuclear magnetic resonance studies of wild-type and glycolyti,: pathway mutants of Saccharomyces cerevisiae. Biochemisov 18, 4487---4499 12 Evans FE, Kaplan NO (1977) 31p Nuclear magnetic resonance studies of HeLa cells. Proc Nail Acad Sci USA 74, 4909-4913 13 Douce R, Bligny R, Brown D, Dome AJ, Genix P, Roby C (1991) Autophagy triggered by sucrose deprivation in sycamore (Acer pseudoplatanus) cells. In: Compartmentation of plant Metabolism in Non-Photosynthetic Tissues (Emes J, ed) Society for Experimental Biology, Cambridge, UK 127-145 14 Thomas TH, Ratcliffe RG (1985) 13C NMR determination of sugar levels in storage roots of carrot (Daucus carota). Physiol Plant 63, 284-290 15 Joumet E, Bligny R, Douce R(1986) Biochemical changes during sucrose deprivation in higher plant cells. J Biol Chem 261, 3193-3199 16 Stewart GR, Larher F (1980) Accumulation of aminoacids and related compounds in relation to environmental stress. Amino acids and derivatives. In: The Biochemistry of Plants (Miflin BJ, ed) vol 5, Academic Press, New York 609-635 17 Roby C, Bligny R, Douce R, Tu Shu-I, Pfeffer PE (1988) Facilitated transport of Mn 2+ in sycamore (Acer pseudo° platanus) cells and excised maize root tips. A comparative in vivo 31p NMR study. Biochem J 252, 401--408 18 Kishore GH, Shah DM (1988) Amino acid biosynthesis inhibitors as herbicides. Annu Rev Biochem 57, 627-663 19 Pinto JEBP, Dyer WE, Weller SC, Herrmann KM (1988) Glyphosate induces 3-deoxy-D-arabino-heptulosonate 7phosphate synthase in potato (Solarium tuberosum L) cells grown in suspension culture. Plant Physio187, 891-893 20 Tateoka TN (1970) Studies on catabolic pathway of protocatechuic acid in mung bean seedlings. Bot Magn 83, 49-59 21 Tateoka TN (1972) 3-Dehydroshikimate hydrolase of Phaseolus mungo seedlings. Bot Mag 85, 285-291 22 Givan CV (1980) Aminotransferases in higher plants. Amino acids and derivatives. In: The Biochemistry of Plants (Miflin BJ, ed) vol 5, Academic Press, New York 329, 357 23 Rhodes, D, Hogan AL, Deal L, Jamieson GC, Haworth P (1987) Amino acid metabolism of Lemna minor L. II. Responses to chlorsuifuron. Plant Physio184, 775-780 24 Marshall G, Kirkwood RC, Martin DJ (1987) Studies on the mode of action of asulam ammoniotriazole and glyphosate in Equisetum arvense (field horsetail). 2. The metabolism of (C-14) asulam, (C-14) aminotriazole and (C- 14) glyphosate. Pest Sci 18, 65-77 25 Coupland D (1984) The effect of temperature on the activity and metabolism of glyphosate applied to rhizome fragments of Elymus repens (= Agropyron repens). Pest Sci 15,226-234 26 Schultz ME, Burnside OC (1980) Absorption, translocation, and metabolism of 2,4-D and glyphosate in hemp dogbane (Apocynum cannabinium). Weed Sci 28, 13-20 27 Marquis LY, Comes RD, Yang CP (1979) Selectivity of glyphosate in creeping red fescue and reed canarygrass. Weed Res 19, 335-342 28 Putman AR (1976) Fate of glyphosate in deciduous fruit trees. Weed Sci 24, 425-430 29 Sandberg C, Meggitt WF, Penner D (1980) Absorption, translocation and metabolism of glyphosate-C 14 in several weed species. Weed Res 20, 195-200