35, 244-250 (1989)
A Rapid Method for the Analysis of the Mode of Action of Bleaching Herbicides A.J. Department of Biochemistry,
BRITTON, AND D. MUSKER
University of Liverpool, P.O. Box 147, Liverpool L69 3BX, United Kingdom Received April 21, 1989; accepted July 11, 1989
A method is described for the analysis of the action of bleaching herbicides in higher plants. A rapid, sensitive method using high-performance liquid chromatography has been devised to determine the pigment (chlorophyll and carotenoid) composition of herbicide-treated plants. This system differentiates between those compounds which cause bleaching through inhibition of carotenoid biosynthesis and those which destroy the plant’s existing pigments through photooxidative events. More importantly, careful and detailed analysis of the carotenoids from these plants will provide information on the precise nature of these bleaching events. It is therefore possible to differentiate between herbicides the action of which may result in very similar visible symptoms. The data obtained concerning the pigment composition also provide much information concerning the carotenoid biosynthetic pathway and the role of antioxidants in the plant cell. 8 1989 Academic Press, Inc.
Chlorosis of plant tissue is a common visible symptom following application of herbicides. It may result as a consequence of either (i) photooxidative destruction of existing pigments, an effect generally seen with compounds which interfere with photosynthetic electron transport (e.g., paraquat, monuron), or (ii) inhibition of pigment biosynthesis as seen with preemergent herbicides such as notflurazon’ and diflufenican. It is important to be able to differentiate between these two effects and, where possible, determine the exact nature of the injury. In the first case the existing green tissue is bleached by the action of a number of oxidative species (lo*, O;, H202, and ‘OH) generated in the chloroplast itself as a consequence of herbicide action. This re’ Abbreviations used: amitrole, lH-1,2,4triazol-3-amine; CPTA, 2-(4-chlorophenylthio)triethylamine hydrochloride; diflufenican, N-(2,4ditluorophenyl)-2-(3-trifluoromethylphenoxy)-3-pyridine carboxamide; diuron, IV’-(3,4-dichlorophenyl)-IV,,Ndimethylurea; monuron, N’-(4-chlorophenyl)N,Ndimethylurea; norflurazon (SAN 9789), 4-chloro%(methylamino)-2 (3-(trifluoromethyl)phenyl)-3(2H)pyridazhsone; paraquat, I,1 ‘d%nethyl(-4,4’-bipyridinium ion); SAN 6706, 4-chloro-5-(dimethylamino-2-a,a,atrithtoro-m-tolyl)-3-(2H)-pyridazinone. 244 0048-3575189 $3.00 Cowkht All rights
8 1989 by Academic Press, Inc. of reproduction in any form reserved.
sults for each particular herbicide, in a wellordered pattern of pigment loss. The carotenoid p-carotene is highly susceptible to such attack and in general, the carotenoids are lost more rapidly than the chlorophylls. In some cases, for example, with certain of the nitrodiphenyl-ether herbicides (A. J. Young and G. Britton, unpublished results), the carotenoids may react with free fatty acids, formed as a result of lipid peroxidation, to form carotenoid acyl esters. These compounds are relatively stable and the plant tissue may appear slightly yellow as the chlorophylls are destroyed. The action of herbicides which interfere with the biosynthesis of carotenoids is well described (1, 2). The application of such preemergence herbicides results in the formation of bleached tissue lacking the carotenoids and chlorophylls normally found. Although a wide range of compounds will inhibit carotenoid biosynthesis, this inhibition usually results in the accumulation of only one or two carotenoid intermediates, namely phytoene (with, for example, norflurazon) or c-carotene (with, for example, 5334). In cases where phytoene is accumulated, traces of phytofluene are usually found, although the relative amount will
vary from herbicide to herbicide. Neurosporene is, however, rarely detected. These carotenoid intermediates are not normally present in untreated photosynthetic tissue. A method is described for the analysis of chlorophylls and carotenoids by highperformance liquid chromatography (HPLC). Several examples are given to show the versatility of this technique in analysing herbicidal modes of action. Particularly important is the identification of carotenoid intermediates or evidence for the photooxidative destruction of carotenoids in treated tissues. This is particularly useful in understanding the mechanism by which a wide variety of herbicidal structures function. MATERIALS
For preemergence treatment, seeds of vuigare (L.) or Raphanus sativus (L.) were soaked overnight in aqueous or aqueous/ethanolic (ethanol CO. I%, v/v) solutions of the appropriate herbicide. The seeds were sown in John Innes No. I potting compost and grown for 6-g days under continuous light (150 pmol m - *set - ‘). For controls, seeds were soaked overnight in aqueous/ethanolic solutions and grown under identical greenhouse conditions. All treatments were performed at least three times. Hordeum
Known amounts of leaf material were homogenised in ethanol, the homogenate was filtered through a cotton wool plug, and the solvent was evaporated under a stream of N,. The residue of pigment-containing lipid material was redissolved in diethyl ether, transferred to a clean vial, and again evaporated under N, ready for HPLC analysis. High Performance Liquid Chromatography
a Zorbax (DuPont Ltd.) ODS
(Octadecylsilyl) 5-pm reversed-phase column (25.0 cm x 4.6 mm) was employed with a solvent gradient of O-100% ethyl acetate in acetonitrile/H,O (9/l) at 1 ml min-‘. Triethylamine (O.l%, v/v) was added to the acetonitrile/H,O. All solvents used were HPLC grade (Rathbums Ltd, UK). The HPLC equipment used incorporated a uv/vis diode array detector (Hewlett Packard HP1040A) which allowed simultaneous monitoring and integration at up to eight separate wavelengths. In a single run the wavelengths monitored were those of the max of the individual carotenoids and the Soret peaks of the chlorophylls: phytoene and monohydroxyphytoene (287 nm); phytofluene (350 nm); c-carotene (401 nm); chlorophyll a (431 nm); neoxanthin (439 nm); antheraxanthin, lutein, and p-carotene-5,depoxide (447 nm); p-carotene and chlorophyll b (455 nm); &carotene (461 nm); y-carotene (465 nm); lycopene (475 nm). Compounds were identified by their electronic absorption spectra determined on-line in the HPLC solvent, retention times (compared to external standards), mass spectra, and, where necessary, ‘H NMR spectra. Carotenoid standards were kind gifts from Hoffmann La Roche (Switzerland) . RESULTS
Although several different approaches are available for the study of herbicide action (e.g., cell-free carotenogenic enzyme systems and plant-cell tissue culture), there are many advantages in determining the effect of herbicidal compounds directly on plants. Analysis of the chloroplast pigments, particularly the carotenoids, provides a method that will immediately distinguish between those herbicides which act by inhibiting carotenoid biosynthesis and those which interfere with photosynthetic electron transport. It is also very important to be able to identify those compounds which may have more than a single effect, i.e., those which may block carotenoid biosyn-
thesis at more than one site, or those that may affect both carotenoid biosynthesis and photosynthetic electron transport. Analysis of the carotenoid composition of treated plants will allow this. Screening of compounds for herbicide action will often involve the analysis of large numbers of samples, and frequently only small amounts of plant material are available for analysis. In such cases it may have only been possible to note whether bleaching of tissue has occurred or not and detailed investigations are not possible. It is necessary therefore to have a procedure for a full and rapid analysis of carotenoids and chlorophylls from a small amount of plant material . The pigment compositions of untreated 6-day-old seedlings of H. vulgare or R. sativus are given in Table 1. The carotenoid composition of higher plant photosynthetic tissue is fairly uniform and the data given in Table 1 are typical of most plants. The xanthophyll lutein and the hydrocarbon p-carotene are the most abundant carotenoids, while neoxanthin and violaxanthin each make up approximately 15% of the total carotenoid. Traces of antheraxanthin and lutein-5,6-epoxide (particularly in R. sativus) are also detected. TABLE 1 Pigment Composition of Untreated (Control) H. vulgare and R. sativus (7-Day-Old Seedlings) H. w&are Neoxanthin Violaxanthin Lutein-5,6-epoxide Antheraxanthin Lutein Zeaxanthin P-Carotene-5,6-epoxide @Carotene Phytoene OH-phytoene Phytofluene Chlorophyll a:b ratio Carotenoid:chlorophyll ratio Note. N.D., not detected.
Figure 1 shows the reversed-phase HPLC separation of these carotenoids and of chlorophyll a and b from untreated seedlings of H. vulgare. On reversed-phase chromatography, the more polar compounds, i.e., the xanthophylls (neoxanthin, violaxanthin, antheraxanthin, lutein5,6-epoxide, lutein), are eluted first followed by the chlorophylls and the hydrocarbons (mainly p-carotene). Biosynthetic intermediates and precursors of carotenoids are completely absent from photosynthetic tissue. In contrast, Fig. 2 shows the pigment profile for norflurazon-treated seedlings. In this example, using 10 @f norflurazon, bleaching is nearly complete and only trace amounts of the normal cyclic carotenoids (neoxanthin, violaxanthin, lutein, and p-carotene) and chlorophylls are present. At a lower herbicide concentration (1 k&f), inhibition of carotenoid biosynthesis is less effective and the pigment composition is similar to that of the controls (Table 2). The biosynthetic intermediates, phytoene and monohydroxyphytoene, are present in both herbicide treatments. Traces of phytofluene, however, can only be detected with low concentrations of norflurazon. It is only in the presence of an inhibitor of carotenoid biosynthesis that these intermediates are detected in photosynthetic tissue. Treatment with other known inhibitors of
% carotenoid 13.0 12.9 17.1 16.6 N.D. 2.3 0.2 0.7 40.3 40.9 N.D. N.D. N.D. N.D. 29.4 26.4 N.D. N.D. N.D. N.D. N.D. N.D. 1.63:1 1.53:l 0.48: 1
FIG. 1. Typical reversed-phase HPLC separation of chlorophylls and carotenoids from untreated barley seedlings. Peak identifications (monitored at 447 nm): A, neoxanthin; B, violaxanthin; C, lutein-.5,6-epoxide; D, antheraxanthin; E, lutein; FF’, chlorophyll b; GG’, chlorophyll a; I, $-carotene.
10 TIME (MN)
FIG. 2. Typical reversed-phase HPLC separation of pigments extracted from norflurazon-treated barley. Monitored at (a) 287 nm and(b) 447 nm. Peak identifications: A, lutein; B, chlorophyll b; C, chlorophyll a; D, monohydroxyphytoene; E, phytoene.
phytoene desaturation also results in partially or wholly bleached plant tissue. The carotenoid composition from such treated cotyledons reveals much about the specific mode of action of this type of herbicide. For example, pre- and postemergence application of metflurazon, SAN 6706 (10 FLM to 1 mM) results in carotenoid and chlorophyll compositions that are indistinguishable from similar treatments with norflurazon. On the other hand, preemergence treatment with the herbicide diflufenican results in a significant decrease in the relative p-carotene content, particularly at high herbicide concentrations (Table 2). The specific loss of p-carotene following postemergence treatment with ditlufenican has also been noted previously (3). Again, this is not seen following the application of norfIurazon or SAN 6706, indicating that diflufenican has an additional function which results in the photooxidative destruction of p-carotene. The accumulation of a number of hydroxylated derivatives of the series of biosyn-
thesis intermediates has been observed with a range of phytoene-desaturase inhibitors. Monohydroxyphytoene in particular is of interest as it occurs in relatively large amounts, while others include dihydroxyphytoene, monohydroxyphytofluene, and, possibly, a dihydroxydidehydrophytoene (3, 4). In addition to this, in treatments of higher plants with compounds that serve to inhibit cyclisation (for example, amitrole), a range of hydroxylated derivatives of IX-, p-, y-, and Z-carotene can be detected by HPLC (unpublished results). Similarly, treatment with CPTA will result in the accumulation of lycopene and traces of OHlycopene. The significance of these hydroxylated derivatives is not fully understood. Their role in relation to the desaturation reactions in carotenoid biosynthesis and with respect to the mode of action of these herbicides is currently under investigation in this laboratory. It is evident, however, that subtle differences in the occurrence of these hydroxy compounds are found with different treatments. Treatment of R. sativus, for example, with norflurazon or SAN 6706, over a range of concentrations, will result in a phytoene:monohydroxyphytoene ratio twice that measured in similar treatments of H. vulgare. In contrast, diflufenican treatment of R. sativus will result in a phytoene:monohydroxyphytoene ratio much lower than those measured in the above treatments. Again, the significance of this information is not fully understood, although such differences between both species and herbicide treatments may be of importance in evaluating herbicidal mode of action. Amitrole is a compound that is known to partially inhibit both carotenoid desaturation and cydisation (5), hence, a number of biosynthetic intermediates may be present. Bleaching of tissue is incomplete even at high herbicide concentrations, resulting in a distinctive pattern of bleaching in the seedlings. Tips are bleached white with increasing pigmentation towards the base. The
Pigment Composition of Norflurazon-
TABLE 2 and Dijlufenican-Treated
H. vu&are and R. sativus Seedlings
H. vulgare Norflurazon (1 p,M) Neoxanthin Violaxanthin Lutein-5,6-epoxide Antheraxanthin Lutein Zeaxanthin 8-Carotene-5,6-epoxide p-carotene Phytoene:OH-phytoene Phytofluene Chlorophyll a.% ratio Carotenoid:chlorophyll
% carotenoids 8.3 13.1 N.D. 4.7 41.6 4.4 N.D. 27.9 1.8:1” 5.5:lb N.D. 2.44:1 0.380: 1
N.D. 21.7 N.D. N.D. 49.7 N.D. N.D. 28.6 2.8:1” 5.2: lb tr. 2.8:1 0.524: 1
Diflufenican 1 mM 13.1 10.7 N.D. N.D. 66.7 9.5 N.D. N.D. 3.6:1 N.D. 3.70:1 0.238: 1
Note. N.D., not detected; tr., trace amounts. n H. vulgare. b R. sativus.
pigment compositions of separate areas of treated H. vulgare seedlings are given in Table 3. As the detection levels for HPLC are very low (typically 2-3 ng for carotenoids), separate areas of tissue can be
analysed from a single leaf. The pigment profile of amitrole-treated plants, determined by reversed-phase HPLC, is shown in Fig. 3. Comparison with the control plants shows that a number of carotenoid
TABLE 3 Pigment Composition of Amitrole-Treated
Neoxanthin Violaxanthin Antheraxanthin Lutein-5,6-epoxide Lutein Zeaxanthin B-Carotene-5,6-epoxide B-Carotene Lycopene and a-carotene Phytoene Phytofluene OH-phytoene Chlorophyll a.% ratio Carotenoid:chlorophyll ratio Lutein:B-carotene ratio Coloured:noncoloured carotenoid ratio
H. vulgare Seedlings
Light-green tissue immediately behind tip
Dark-green tissue at base
6.7 13.1 10.3 N.D. 23.7 12.5 N.D. 23.2 10.5 N.D. 76.4 5.3 18.3 2.067: 1 1.236: 1 1.385:1 4.4: 1
% carotenoid 13.8 14.4 5.2 N.D. 36.2 2.7 N.D. 13.4 11.1 3.2 68.2 31.8 N.D. 1.577: 1 0.471:1 2.708: 1 61.6:1
13.4 16.7 1.5 N.D. 38.7 N.D. N.D. 27.9 1.8 N.D. N.D. N.D. N.D. 1.873:1 0.398: 1 1.387:1 -
tissue may also result from the action of those herbicides which interfere with photosynthetic electron transport. This results in photooxidative destruction of the carotenoids and chlorophylls normally present in photosynthetic tissue. In the presence of the herbicides diuron and paraquat, for example, selective destruction of some pigments over others is seen and a characteristic pattern emerges. In general, p-carotene and neoxanthin are particularly susceptible to photooxidative damage while lutein, violaxanthin, and the chlorophylls FIG. 3. Typical reversed-phase HPLC separation of are more resistant. Loss of violaxanthin is, a saponified pigment extract from amitrole-treated however, complicated by the induction of barley seedlings. Monitored at (a) 287 nm and (b) 447 cycle (6). This generally nm. Peak identifications: AA’, neoxanthin; BB’, vio- the violaxanthin results in a light-driven deepoxidation of vilaxanthin; C, antheraxanthin; D, lutein, E, zeaxanto zeaxanthin via antheraxanthin; F, monohydroxylycopene; G, monohydroxy de- olaxanthin rivatives of 6- and y-carotene; H, lycopene; I, thin and a subsequent dark-driven epoxida&carotene; J, y-carotene; K, p-carotene. tion back to violaxanthin. The deepoxidation part of the cycle is often promoted by short (for example, paraquat) or long term biosynthetic intermediates are present. It is exposure (for example, monuron) to postevident that desaturation of phytoene is in- emergent herbicides but it is also evident in hibited, resulting in the accumulation of a the treatments with low concentrations of the preemergent herbicides norflurazon and number of acyclic carotenoids: phytoene, amitrole. Treatment with these postemerOH-phytoene, and traces of phytofluene. However, the inhibition of desaturation is gent herbicides does not result in the accuof biosynthetic intermediates incomplete, and unlike nortlurazon or di- mulation flufenican, other intermediates are found; (phytoene and lycopene, for example) seen lycopene is also present, often in large with herbicides like nofflurazon. The application of compounds that interamounts, indicating that partial blockage of cyclisation takes place. The HPLC can also fere with photosynthetic electron transport detect trace amounts of the monocyclic ca- and, in particular, those that block electron flow at PSI1 (diuron and monuron, for exrotenoids y- and b-carotene, intermediates of between lycopene and p-carotene and ample) will result in the accumulation a-carotene, respectively. Cyclisation of small amounts of P-carotene-5,6-epoxide. both E and p end groups appears to be This compound is formed as a direct consequence of the photooxidative breakdown equally, inhibited by the action of amitrole. of p-carotene in the thylakoid membrane. This is in contrast to the action of CPTA, Its presence in plant tissue is indicative of also a desaturase and cyclase inhibitor, photooxidative damage to the photosynwhich accumulates y-carotene in much larger quantities than a-carotene indicating thetic apparatus of higher plants (unpubthat cyclisation of the P-end group is pref- lished results). Figure 4 shows the effect of paraquat on erentially inhibited (unpublished results). of H. vulgare Preemergence treatment of H. vulgare or the pigment composition seedlings. The pigment profile is very difR. sativus with norflurazon or amitrole illustrates the effect of such inhibitors on the ferent from that found with inhibitors of carotenoid biosynthesis (cf. Figs. 2 and 3). It composition of bleached tissue. Bleached
FIG. 4. Typical reversed-phase HPLC separation of pigments extracted from paraquat-treated barley seedlings. Peak identifications (monitored at 447 nm): A, chlorophyll breakdown product; B, neoxanthin; C, violaxanthin; D, lutein; E, chlorophyll b; F, chlorophyll a; G, P-carotene-5,6-epoxide; H, p-carotene.
can clearly be seen that pigment destruction has taken place rather than inhibition of carotenoid biosynthesis. Clear signs of this are (i) the presence of P-carotene-5,6epoxide and of chlorophyll breakdown products, (ii) the overall pigment composition indicating the selective destruction of some pigments, (iii) the very much increased chlorophyll:carotenoid molar ratio, and, (iv) most importantly, the complete absence of biosynthetic intermediates. CONCLUSIONS
A rapid and very sensitive technique that will provide information concerning the action of herbicides directly on plants rather than on model systems was developed. Full analysis by HPLC can be completed in approximately 1 hr (including extraction of pigments from treated plants), thereby allowing a large number of samples to be examined over a short period of time. HPLC analysis, particularly in conjunction with a diode-array detector, is a very powerful technique. It will allow minor compounds to be detected, the presence of
which can reveal a great deal of information concerning the mode of action of herbicides. The versatility of the diode-array detector will, in addition, allow for a complete, qualitative and quantitative analysis of a single extract of a small amount of plant material to be made at any wavelength from 210 to 610 nm. This permits simultaneous detection of the normal cyclic carotenoids and their acyclic biosynthetic intermediates, many of which have a A,, far removed from that of the coloured carotenoids (for example, phytoene h,,, 287 nm, lutein A,,, 447 nm). ACKNOWLEDGMENTS
Financial support for this research was provided by the Science and Engineering Research Council of Great Britain. REFERENCES
1. S. M. Ridiey, Interaction of chloroplasts with inhibitors: Induction of chlorosis by diuron during prolonged illumination in vitro, Plant Physiol. 59, 724 (1977). 2. G. Britton, P. Barry, and A. J. Young, Carotenoids and chlorophylls: Herbicidal inhibition of carotenoid biosynthesis. in: “Herbicides and Plant Metabolism” (A. D. Dodge, Ed.), Cambridge Univ. Press, Cambridge, in press. 3. G. B&ton, P. Barry, and A. J. Young, The mode of action of diflufenican: Its evaluation by HPLC, in “Proceedings of the British Crop Protection Conference (Weeds),” Vol. 3, pp. 1015-1022. 4. P. Barry, “Factors Affecting Carotenoid Content and Composition of Higher Plant Tissues,” Ph.D. thesis, University of Liverpool, England, 1987. 5. S. M. Ridley, Carotenoids and herbicide action, in “Carotenoid Chemistry and Biochemistry” (G. Britton and T. W. Goodwin, Eds.), pp. 353369, Pergamon Press, Oxford, 1982. 6. H. Y. Yamarnoto, The biochemistry of the violaxanthin cycle in higher plants, Pure Appl. Chem. 51, 639