Auxin herbicides induce H 2 O 2 overproduction and tissue damage in cleavers ( Galium aparine L . )

The phytotoxic effects of auxin herbicides, including the quinoline carboxylic acids quinmerac and quinclorac, the benzoic acid dicamba and the pyridine carboxylic acid picloram, were studied in relation to changes in phytohormonal ethylene and abscisic acid (ABA) levels and the production of H2O2 in cleavers (Galium aparine). When plants were roottreated with 10 mM quinmerac, ethylene synthesis was stimulated in the shoot tissue, accompanied by increases in immunoreactive levels of ABA and its precursor xanthoxal. It has been demonstrated that auxin herbicide-stimulated ethylene triggers ABA biosynthesis. The time-course and dose-response of ABA accumulation closely correlated with reductions in stomatal aperture and CO2 assimilation and increased levels of hydrogen peroxide (H2O2), deoxyribonuclease (DNase) activity and chlorophyll loss. The latter parameters were used as sensitive indicators for the progression of tissue damage. On a shoot dry weight basis, DNase activity and H2O2 levels increased up to 3-fold, relative to the control. Corresponding effects were obtained using auxin herbicides from the other chemical classes or when ABA was applied exogenously. It is hypothesized, that auxin herbicides stimulate H2O2 generation which contributes to the induction of cell death in Galium leaves. This overproduction of H2O2 could be triggered by the decline of photosynthetic activity, due to ABA-mediated stomatal closure.


Introduction
For more than 50 years synthetic substances possessing auxin activity have been among the most successful herbicides used in agriculture (Cobb, 1992;Sterling and Hall, 1997;Grossmann, 2000a).The most important chemical classes of these so-called auxin herbicides include chlorophenoxy acids, benzoic acids (e.g.dicamba), pyridines (e.g.picloram), and quinoline carboxylic acids (e.g.quinmerac, quinclorac).They basically mimic the effects of supraoptimal endogenous auxin concentrations (Cobb, 1992;Sterling and Hall, 1997;Taiz and Zeiger, 1998;Grossmann, 2000a).The early effects in sensitive dicots are characterized by growth abnormalities, such as epinasty and growth inhibition with intensified green leaf pigmentation within 24 h.These phenomena are followed by chloroplast damage, leading to chlorosis and by the destruction of membrane and vascular system integrity, leading to desiccation, tissue necrosis and decay (Cobb, 1992;Sterling and Hall, 1997;Grossmann, 2000a).Recently, conclusive evidence has been presented that the early, growth-retarding effects are caused by auxininduced ethylene which triggers an increase in the biosynthesis of abscisic acid (ABA; Grossmann et al., 1996;Hansen and Grossmann, 2000).Subsequently, ABA is distributed within the plant and inhibits growth by closing stomata, which limits carbon assimilation and, consequently, biomass production (Scheltrup and Grossmann, 1995;Grossmann et al., 1996;Grossmann, 2000a, b;Hansen and Grossmann, 2000).Direct inhibitory effects of ABA on photosynthetic enzyme activity and cell division and expansion have also been reported (Trewavas and Jones, 1991).In addition, ABA is recognized as an important hormone which promotes leaf senescence (Taiz and Zeiger, 1998).Therefore, it has been deduced that, together with ethylene and its biosynthetic coproduct cyanide, ABA also contributes to the mode of action underlying the late auxin herbicide effects in sensitive dicots, especially the induction of tissue decay and cell death (Grossmann, 2000b).
Senescence in leaves is an endogenously programmed process that involves, among other changes, the overproduction of reactive oxygen species (ROS), such as hydrogen peroxide (H 2 O 2 ;Leshem, 1988;Dat et al., 2000).ROS have also been proposed as a central component of plant response to biotic and abiotic stresses (Dat et al., 2000).An accumulation of H 2 O 2 is generally considered to contribute to oxidative damage and probably process signalling (Del Rio et al., 1998;Dat et al., 2000).This raises the question of whether H 2 O 2 accumulates in plants during auxin-induced tissue damage and functions as a trigger of this process.Therefore, the aim of the present work was to elucidate the relationship between auxin effects on the evolution of H 2 O 2 and changes in the phytohormone levels of ethylene and ABA, stomatal behaviour and CO 2 assimilation.The studies were focused on the action of the auxin herbicide quinmerac in cleavers (Galium aparine), because this dicot weed is particularly sensitive to auxins.To evaluate whether the response is common to all auxin herbicides, several representatives of the known chemical classes, including quinclorac, dicamba and picloram, were also examined.These investigations revealed an overproduction of H 2 O 2 elicited by all auxin herbicides in Galium shoot tissue.As far as is known, this phenomenon is reported for the first time.

Experiments with plants in hydroponics
Seedlings of Galium aparine L. at the first whorl stage (15 d after sowing) were raised to the third or fourth whorl stage, as described previously (Scheltrup and Grossmann, 1995).Uniformly developed plants were transferred into 320 ml glass vessels in half-strength Linsmaier-Skoog medium in 16u8 h lightudark cycles at 25u20 8C and 75% relative humidity (three plants per vessel, five replications) (Linsmaier and Skoog, 1964).Light (530 mmol m À2 s À1 , 400-750 nm) was provided by Osram Powerstar HQI-R 250 WuNDL and Osram Krypton 100 W lamps.The solution was aerated throughout the experiments.After 1 d of adaption, quinmerac, quinclorac, dicamba or picloram were added to the medium in acetone solution (0.1% final concentration of acetone).Controls received corresponding amounts of acetone alone, with no adverse effect on the growth of the plants.At various times after treatment, growth parameters and ethylene formation were measured.Shoots from parallel vessels were harvested, immediately frozen in solid CO 2 and stored at À80 8C.Plant material was powdered under liquid nitrogen.For characterization of the damage process, the plant material was analysed for total chlorophyll and for the enzyme activity of deoxyribonuclease (DNase) in three replicates as described previously (Grossmann and Jung, 1982).Stomatal behaviour of plants in parallel vessels was determined by measuring the diffusive resistance to water vapour transport with a diffusion porometer (Li-65, Li-Cor Autoporometer, Bachhofer, Reutlingen, Germany) with a horizontal sensor (LI-20S).Five leaves per plant were removed from the first and second whorls and carefully stuck to filter paper with a small strip of doublesided adhesive tape.Care was taken that the leaves with the abaxial side upwards fitted exactly over the aperture.The values of diffusion resistance were the means of six replications.All experiments were repeated at least twice and proved to be reproducible.The results of a representative experiment are shown.

Determination of ethylene formation
After treatment in hydroponic vessels, detached shoots of plants were transferred to 100 ml glass cylinders containing 10 ml of half-strength Linsmaier-Skoog medium (one shoot per cylinder; six replications) (according to Scheltrup and Grossmann, 1995).The cylinders were sealed with rubber caps.After incubation for a further 3 h under light, a 1 ml gas sample of the head space was withdrawn and ethylene was measured by GC.

Determination of ACC
Samples of powdered plant material (100 mg; three replications) were extracted with 70% aqueous ethanol.Following oxidative conversion, the ACC content was assayed as ethylene by GC (Lizada and Yang, 1979;Scheltrup and Grossmann, 1995).

Determination of tissue hydrogen cyanide
Tissue HCN was determined as previously described (Grossmann and Kwiatkowski, 2000).A sample of powdered plant material (1.5 g, three replications) was transferred to a glass vial (25 mm in diameter, 38 mm in height), the lid of which contained a filter paper disc to which 350 ml 1.5 M NaOH had been applied.Sealing the vial with the lid positioned the filter over the plant material.Then, 2.5 ml 5% H 2 SO 4 was injected into the sample through a perforation in the side of the vial, which was subsequently sealed.The resultant acidified brei was stirred at 22 8C for 5 h to allow evolved HCN to be trapped by the filter.The filter was then eluted with NaOH for 1 h and the cyanide content was determined using the colorimetric method developed previously (Lambert et al., 1975).

Determination of ABA and xanthoxal
Powdered plant material (1 g) was extracted with 80% aqueous methanol (three replicate extractions) and the extracts were passed through a C 18 -reversed phase prepacked column (SEPPAK; Waters, Ko ¨nigstein, Germany), as described previously (Weiler et al., 1986;Grossmann et al., 1987;Hansen and Grossmann, 2000).For ABA, the effluent was concentrated in vacuo, dissolved in 3 ml double-distilled water, acidified to pH 2.5 with 1 M HCl and partitioned three times into ethyl acetate (3 ml).The organic solvent was evaporated to dryness under a N 2 stream and samples were redissolved in 2 ml of 5% methanol in 0.1 M acetic acid.Separation of ABA in a 1 ml aliquot of the extract was performed by HPLC on a reverse-phase Nucleosil 120-5 mm C 18 column (250 3 10 mm, Machery-Nagel, Du ¨ren, Germany) using a linear gradient from 5% methanol in 0.1 M acetic acid to 95% methanol.The fractions containing ABA were collected and the quantitative determination was performed by enzyme-immunoassay (ELISA).Monoclonal antibodies for ABA (Mertens et al., 1983) were used for analyses according to a standard procedure described earlier (Weiler et al., 1986).
Separation of xanthoxal in an aliquot of 1 ml of the extract was performed by HPLC as described above.In this case, the gradient did not contain acetic acid, because the epoxy-group of xanthoxal is sensitive to acids.The fraction containing cis-and trans-xanthoxal were collected analysed by ELISA, as described previously (Feyerabend and Weiler, 1988a, b).Monoclonal antibodies for xanthoxal were used (Feyerabend and Weiler, 1988a, b;Hansen and Grossmann, 2000).The antibodies for ABA and xanthoxal were kindly provided by Professor Elmar W Weiler (University of Bochum, Germany).All samples were assayed at least in triplicate.Internal performance controls of assay accuracy and reliability were carried out as described earlier (Weiler et al., 1986;Grossmann et al., 1987).Recovery of ABA and xanthoxal, as verified by internal radiolabelled standards added to the methanolic extracts of the ground tissues, were above 90% in both cases.Confirmation of the identities of ABA and xanthoxal in the immunoreactive HPLC fractions of the plant extracts was obtained by GC-MS (Finnigan 4600, Finnigan, San Jose, California, USA) and LC-MS (Sciex API III, Sciex, Concord, Ontario, Canada), respectively.

Measurement of levels of hydrogen peroxide
The H 2 O 2 level was usually measured by the modified colorimetrical method of Okuda et al. (Okuda et al., 1991).Powdered plant material (100 mg, three replications) was extracted at 4 8C for 5 min in 1 ml of 0.2 N HClO 4 under shaking.The extract was centrifuged at 10 000 g for 7 min.The supernatant was adjusted to pH 7.5 with 2 N NaOH and an aliquot of 100 ml passed through a 0.4 ml column of Dowex AG-1-X2.The column was washed twice with 750 ml 0.375 M phosphate buffer (pH 6.5) and the eluate was used for the assay of H 2 O 2 .The reaction mixture contained 1.6 ml of the eluate, 600 ml of 12.5 mM 3-(dimethylamino)benzoic acid in phosphate buffer, 20 ml of 3-methyl-2-benzothiazoline hydrazone and 10 ml of peroxidase (13 units).After 15 min, the increase in absorbance at 590 nm was measured.The production of ROS was calculated using a calibration curve established in the presence of H 2 O 2 .Recovery was up to 90%, as determined using various amounts of added H 2 O 2 to the plant extracts as internal standard.Treatment of the sample and the standard with catalase (500 units) completely neutralized increases in absorbance.For verification of H 2 O 2 measurements in the plant material, levels were additionally determined using the colorimetric method with titanium sulphate as described previously (Chang and Kao, 1998).

Gas-exchange measurements
For determination of CO 2 uptake as a parameter for CO 2 assimilation, Galium plants at the second whorl stage were cultivated hydroponically in illuminated glass chambers (four plants per chamber, three replications) that received a constant stream of air (Scheltrup and Grossmann, 1995).After root treatment, the amount of CO 2 assimilated per unit time was determined continuously from the difference between the CO 2 contents of the inflowing and outflowing air streams.

Histochemical determination of starch grains in chloroplasts
After formalin fixation of Galium leaves and embedding in paraffin, 5 mm longitudinal sections were stained using the PAS reaction with carbohydrates as described (Ruzin, 1999).The procedure involved first the oxidation of carbohydrates with periodic acid followed by reaction with the fluorescent Schiff 's reagent pararosanilin.

Results
Following hydroponic treatment of Galium plants at the third whorl stage with 10 mM quinmerac, epinastic stem and leaf curvature developed within 4 h and subsequently inhibition of shoot growth, as measured by fresh weight, commenced during the next 20 h (Fig. 1).The plants remained stunted, internode elongation and leaf area were reduced.During this period of the early, growth-retarding effects of quinmerac, ACC and ethylene production were stimulated in the shoot tissue, accompanied by increases in immunoreactive levels of ABA and its precursor xanthoxal within 8 h (Fig. 2).In contrast, levels of cyanide, which is generated as a co-product of ethylene biosynthesis (Grossmann, 2000a), were not changed by Phytotoxic effects of auxin herbicides 1813 quinmerac treatment in Galium shoot tissue (Fig. 2).After exposure of Galium plants to quinmerac, the onset and time-course of ABA accumulation in the shoot tissue closely correlated with increasing stomatal closure, as determined by water vapour diffusive resistance (Fig. 2).Concomitantly, photosynthetic CO 2 assimilation by the plants, as evaluated by the kinetics of CO 2 uptake, was progressively reduced compared with controls (Fig. 2).After 24 h, increases in ABA levels in the shoot tissue of 18-fold, relative to controls, coincided with reductions in CO 2 uptake by 32% (Fig. 2).No starch grains in the chloroplasts of the treated leaf tissue were found which indicates a reduced formation (Fig. 3).These alterations were accompanied by an accumulation of H 2 O 2 and progressive tissue damage, characterized by an increase in DNase activity in the shoot tissue (Fig. 1).The increase of DNase activity which has been shown to be a sensitive indicator for senescence progression (Grossmann and Jung, 1982) was detectable as early as 24 h after quinmerac application, before chlorophyll loss could be detected (Fig. 1).During 72 h of treatment, the timecourse of DNase activity closely correlated with the rise in H 2 O 2 levels in the shoot tissue (Fig. 1).The result was confirmed by using the titanium method for the determination of H 2 O 2 levels (data not shown).This method showed identical qualitative changes at 4 3 higher absolute H 2 O 2 levels in the tissue, compared to the peroxidase method.Similar correlations between increases in ABA and H 2 O 2 levels and DNase activity were observed in dose-response experiments with quinmerac (Fig. 4).On a shoot dry weight basis, DNase activity and H 2 O 2 levels in the shoot tissue increased up to 3-fold, relative to the control (Fig. 4).Comparing the effects of auxin herbicides from different chemical classes, such as quinclorac, dicamba and picloram, it was demonstrated that the concomitant induction of ABA, H 2 O 2 and tissue damage are effects common to all auxin herbicides in Galium   (Fig. 5).In addition, exogenous application of ABA to Galium plants via the root for 72 h led to H 2 O 2 accumulation in the shoot in a dose-dependent manner (Fig. 6).In contrast, the ethylene-releasing compound ethephon did not change H 2 O 2 production at a concentration of 100 mM (Fig. 6).

Discussion
Restriction of CO 2 diffusion through stomatal closure appears to be responsible for the decline in CO 2 uptake and assimilation by leaves under stress conditions (Cornic, 2000).Limited CO 2 fixation leads to an enhanced accumulation of ROS, such as H 2 O 2 , which originates from increased thylakoid membrane electron leakage to O 2 in the chloroplasts (Dat et al., 2000).
A decline in photosynthesis below a certain threshold level isalso proposed to act as a senescence-inducing signal (Bleecker and Patterson, 1997).Hydrogen peroxide is known to react with superoxide radicals to form hydroxy radicals, which cause cellular damage through lipid peroxidation (Dat et al., 2000).In addition, H 2 O 2 might function as a signal molecule in senescence (Del Rio et al., 1998), plant stress responses (Dat et al., 2000) and in the ABA signal transduction pathway leading to gene expression (Guan et al., 2000).Stress-induced ABA biosynthesis is recognized as a major trigger for stomatal closure and senescence (Fig. 7; Taiz and Zeiger, 1998;Qin and Zeevaart, 1999).In addition, ethylene is involved in the regulation of leaf senescence (Fig. 7; Taiz and Zeiger, 1998).Studies have presented evidence that auxin herbicides stimulate the formation of ethylene via the induction of ACC synthase.Ethylene elicits ABA biosynthesis through increasing xanthophyll cleavage to the ABA precursor xanthoxal in sensitive dicots (Fig. 7; Hansen and Grossmann, 2000).Accumulated ABA then causes reductions in stomatal aperture and CO 2 assimilation, followed by growth inhibition and tissue senescence and decay (Fig. 7; Scheltrup and Grossmann, 1995;Grossmann et al., 1996;Hansen and Grossmann, 2000;Grossmann, 2000b).As presented here, as far as is known   Phytotoxic effects of auxin herbicides 1815 for the first time, the application of auxin herbicide also induces an accumulation of H 2 O 2 , which coincides with the time-course and dose-response of ABA increase and the progression of tissue damage in Galium.In accordance, exogenously applied ABA promoted H 2 O 2 generation.In sensitive grasses, auxin herbicide-induced cyanide, which is formed at damaging concentrations as a co-product of ethylene during the oxidation of ACC, is implicated in phytotoxic growth inhibition (Grossmann and Kwiatkowski, 2000;Grossmann, 2000a).In Galium, quinmerac did not affect cyanide levels in shoot tissue which points against a damaging role of cyanide.Therefore, it is hypothesized that the decline in photosynthetic CO 2 assimilation, due to ABA-mediated stomatal closure, leads to an overproduction of ROS, which are involved in the induction of tissue damage and cell death (Fig. 7).More detailed studies are required to elucidate whether these effects are in addition or part of the senescence syndrome known to be accelerated by ABA and ethylene (Fig. 7).

Fig. 1 .
Fig. 1.Time-course of the effects of quinmerac on shoot growth (measured by fresh weight), and levels of total chlorophyll, H 2 O 2 and DNase activity (related to tissue dry weight, DW) in shoots of Galium aparine.G. aparine plants were root-treated with compound in hydroponics.Vertical bars represent SE of the mean.

Fig. 2 .
Fig. 2. Time-course of the effects of quinmerac on leaf diffusion resistance and CO 2 uptake, ethylene formation and levels of ACC, cyanide and immunoreactive xanthoxal and ABA (related to tissue dry weight, DW) in shoots of Galium plants treated in hydroponics.Vertical bars represent SE of the mean.

Fig. 3 .
Fig. 3. Fluorescence micrographs of longitudinal sections from Galium leaf tissue after treatment of plants with 10 mM quinmerac for 24 h (B) and without treatment (A), showing the starch grains (indicated by arrows) in the chloroplasts.

Fig. 4 .
Fig. 4. Dose-response of the effects of quinmerac on levels of total chlorophyll, H 2 O 2 , immunoreactive ABA, and DNase activity in shoot tissue of Galium plants treated for 72 h hydroponically.Vertical bars represent SE of the mean.For control values "SE (100%) related to tissue dry weight see Figs 1 and 2.

Fig. 5 .
Fig. 5. Effects of selected auxin herbicides (quinclorac, QC; dicamba, DC; picloram, PC) on levels of total chlorophyll, H 2 O 2 , immunoreactive ABA, and DNase activity in shoot tissue of Galium plants treated with 10 mM of compounds for 72 h hydroponically.Vertical bars represent SE of the mean.For control values (100%) see Figs 1 and 2.

Fig. 6 .
Fig. 6.Effects of ABA or ethephon on H 2 O 2 production in shoot tissue of Galium plants treated for 72 h hydroponically.Vertical bars represent SE of the mean.For control values (100%) see Fig. 1.