Author for correspondence: John L. Harwood Tel: +44 (0)29 20874108 Fax: +44 (0)29 20874116 Email: Harwood@Cardiff.ac.uk
• Several pesticide classes have been used for control of the notorious plant pathogen, Phytophthora infestans. Some of these alter lipids, suggesting that lipid metabolism may be a target site. Here, we investigate the action of a new, active, mandelamide compound, SX 623509 (N-2′-(4″-ethoxy-3″-methoxy) phenylethyl-3, 4-dichloromandelamide), on lipid metabolism.
• Phytophthora infestans cultured in pea-broth or minimal media was exposed to different concentrations of SX 623509. Lipid metabolism was followed with radiolabelled acetate, choline or ethanolamine. Products were analysed following separation by thin-layer chromatography and gas–liquid chromatography.
• SX 623509 reduced growth and lipid labelling from [14C]acetate in both media. The inhibition in lipid labelling was not caused merely by a reduction in uptake of the radiolabelled precursor. There were changes in labelling patterns, particularly reductions in phosphatidylcholine and triacylglycerol and increases in phosphatidate and diacylglycerol. The effect of SX 623509 on phosphatidylcholine labelling was followed in more detail.
• We conclude that the mandelamide pesticide SX 623509 reduces lipid synthesis at concentrations similar to those inhibiting growth. Specific effects on lipid labelling patterns are probably caused by inhibition of cholinephosphotransferase, which may be a future target site for pesticide development.
Phytophthora infestans is probably the most notorious of all plant pathogens, being responsible for the Irish potato famines of the nineteenth century (Griffith et al., 1992; Erwin & Ribeiro, 1996). A reflection of its recognized importance, is the high percentage of total pesticide expenditure which is spent in efforts to control P. infestans (Schwinn & Staub, 1995). However, the persistence of such Oomycetes and unique features of the biochemistry of the Pythiaceae means that many compounds such as fungicides are ineffective (Griffith et al., 1992). This has made P. infestans difficult to control efficiently.
A number of systemic pesticides have been used against Oomycetes such as P. infestans. These include phenylamides, carbonates, cyano-oxime, phosphonates, dimethomorph and strobilurins (Staub & Hubele, 1981; Schwinn, 1983; Griffith et al., 1992; Copping & Hewitt, 1998). Of the carbonates, propamocarb and prothiocarb cause increased membrane permeability (Papavizas et al., 1978) and it was suggested that this could be related to alterations in the fatty acid composition which are caused by low concentrations of propamocarb in both Pythium and Phytophthora species (Burden et al., 1988). In other, metabolic, studies Reiter et al. (1996) reported a number of significant effects of propamocarb on the lipid metabolism of P. infestans. These observations suggest that acyl lipid metabolism could be a target for pesticides in Phytophthora species and might offer alternative biochemical targets for agricultural control. This is especially relevant because the biochemistry of Phytophthora means that many otherwise useful pesticides which either bind to sterols in membranes or which interfere with their biosynthesis are ineffective (Griffith et al., 1992).
In an effort to develop new types of pesticides, a number of mandelamide compounds have been synthesized. Several of these are active towards Plasmopara and Phytophthora species and, although a mode of action was not elucidated, the most significant metabolic effects were towards lipids (AgrEvo (UK) Ltd, unpublished data). One alteration was an increased labelling of the free (nonesterified) fatty acid (NEFA) fraction. Because a build-up of NEFAs is lethal in some fungi (Weete, 1983) and may contribute to the action of triarimol on Ustilago maydis (Ragsdale & Sisler, 1972) as well as the activity of C14 demethylase inhibitors generally (Weete, 1983), this observation could be pertinent to the pesticidal properties of mandelamides. Accordingly, we investigated the biochemical effects of an active mandelamide compound, SX 623509 (Fig. 1), on lipid metabolism in P. infestans. At pesticide concentrations which cause growth inhibitions, a number of specific effects were found on lipid metabolism. Such specific effects would not be expected if lipid metabolism was being altered as a consequence of changes in growth and may therefore be related to the pesticide's preliminary mode of action.
Materials and Methods
Radiochemicals ([1-14C]acetate (1.85–2.29 GBq mmol−1), [methyl-14C]choline chloride (1.85–2.29 GBq mmol−1), [2-14C]ethan-1-ol-2-amine hydrochloride (1.85–2.29 GBq mmol−1)) were obtained from Amersham Pharmacia Biotech UK Ltd. (Amersham, UK). Fatty acid methyl esters and fatty acid standards were from Nu-Check Prep. Inc. (Elysian, MN, USA) and lipids from Sigma Chemical Co. Ltd (Poole, UK). Agar bacteriological (agar No. 1) and malt extract were from Oxoid Ltd (Basingstoke, UK). Other chemicals and solvents were of analytical or best available grades from Sigma or from Fisher Scientific (UK) Ltd. (Leicester, UK).
SX 623509 (N-2′-(4″-ethoxy-3″-methoxy) phenylethyl-3, 4-dichloromandelamide), a mandelamide compound (Fig. 1) (RS-mixture) was prepared by, and was a kind gift from Aventis Crop Science Ltd (Lyon, France). It was dissolved in dimethylsulphoxide (DMSO) and the solvent concentration kept below 0.25% (v : v) for experiments. This concentration of DMSO did not affect growth, lipid composition or metabolism (data not shown) but the control always contained the same DMSO concentration as test samples in the experiments.
Cultures of Phytophthora infestans were maintained on pea medium (Griffiths et al., 2003). Growth in liquid cultures and in minimal media (such as Henniger) was as described in Griffiths et al. (2003). For radial growth inhibition studies, the mandelamide compound was incorporated into the appropriate agar medium while still warm after autoclaving. Mycelial plugs were cut from sporulating cultures grown on agar media using a cork borer and transferred to the centre of the agar plates containing pesticide. Inhibition of growth was estimated by measuring growth from the edge of the mycelial plug to the edge of the growing culture over a number of days.
Liquid cultures (10 ml) of P. infestans were set up and incubated in appropriate media for 1–2 wk (depending on the medium used) until the mycelia nearly covered the surface of the medium (Griffiths et al., 2003). Cultures were preincubated for 1 h with pesticide (as required) before addition of radiolabelled substrate (37 kBq in 10 ml media, unless otherwise stated). Incubations were carried out for up to 18 h at 20°C in continuous light. At the end of the incubations, cultures were filtered through Whatman No. 1 filter paper (Whatman International Ltd, Maidstone, UK) under suction and washed three times with fresh medium. An aliquot of the waste media was taken for scintillation counting to check the amount of radiolabel taken up.
Samples were counted for radioactivity using Opti-fluor scintillant (Packard Bioscience BV, Groningen, The Netherlands) in a Rackbeta liquid scintillation counter (Wallac Oy, Turku, Finland). Quench correction was made automatically by the external standard channels ratio method.
Lipid extraction and analysis
This was carried out as described in Griffiths et al. (2003). Briefly, it involved an immediate inactivation of all enzymes at the end of incubation with 2-propanol, followed by two-phase extraction and separation of lipid classes by thin-layer chromatography (TLC).
Analysis of aqueous intermediates of phosphatidylcholine synthesis
The aqueous fraction from lipid extractions was analysed by TLC by the method of Vance et al. (1981). Intermediates were identified by co-chromatography with authentic choline-containing standards (Sigma) and by spraying with Dragendorff reagent (Kates, 1986).
Growth inhibition by mandelamide pesticide
Inhibition of radial growth of P. infestans by compound SX 623509 in two different liquid growth media is shown in Fig. 2. In both media, 0.5 µm pesticide caused severe inhibition of growth, which was statistically significant in the pea-agar medium by 3 d (Fig. 2a). Inhibition of growth was maintained throughout the experimental period (up to 20 d in the case of CaCl2-deficient Henniger medium: Fig. 2b). No growth was detected with 5 µm SX 623509 in pea-agar medium and none up to 14 d for 1 µm SX 623509 in CaCl2-deficient Henniger (Fig. 2b). In other minimal media, such as Henniger, inhibition of growth was significant at 0.5 µm fungicide (data not shown). Approximate I50 values for SX 623509 in the pea medium (0.35 µm) compared with the various minimal media (0.31 µm) were rather similar, although growth was at least twice as fast in the pea medium.
Effects on total lipid labelling
[1-14C]Acetate was used as a general precursor of lipids. Previous studies in a number of systems have justified its use for such purposes (Roughan & Slack, 1982) and, for acyl lipids, this radiolabelled precursor has an advantage that it labels almost exclusively the acyl portions. When P. infestans was incubated with [1-14C]acetate in pea medium about half of the available radiolabel was taken up by 3 h and around 80% by 18 h (Fig. 3a). At the same times, 18% and 32% of the total available [1-14C]acetate (i.e. that taken up) was incorporated into lipids (Fig. 3b).
By contrast, the presence of SX 623509 reduced total uptake of [1-14C]acetate, which was statistically significant at 18 h (Fig. 3a). The radiolabelling of total lipids was also reduced significantly. Thus, by 3 h, it represented 11% of the [1-14C]acetate taken up (Fig. 3b). Total radiolabelling of lipids was 17% of the [1-14C]acetate taken up after 18 h.
Thus, the mandelamide pesticide had two effects on acetate incorporation. First, it reduced total uptake of the precursor, although this was only significant at 18 h. Second, it reduced the proportion of radioactive incorporation into lipids. The latter was seen even after short incubation times. Comparable experiments were also carried out with Merz minimal medium, with similar results (data not shown).
The experiments reported in Fig. 3 were carried out with a high (100 µm) concentration of SX 623509 in order to challenge the system severely, but total lipid labelling was reduced by concentrations as low as 0.5 µm (data not shown), in keeping with previous observations (Aventis Crop Science Ltd) and at a similar concentration to that needed for growth inhibition (Fig. 2).
Qualitative effects on lipid class labelling
The inhibition of lipid labelling by SX 623509 (Fig. 3) raised the obvious question as to whether this was a general action on all acyl lipids via fatty acid synthesis or whether the compound inhibited specific pathways of lipid metabolism. Moreover, because growth was inhibited by the mandelamide compound, it was possible that the reduction in lipid labelling was merely secondary to a reduced demand for membrane lipid formation.
Initially, we examined fatty acid labelling. For P. infestans grown in pea medium the major labelled acids at 3 h were palmitate (43%) myristate (16%), oleate (14%) and linoleate (13%). After 18 h the major products were palmitate (36%), linoleate (19%), oleate (14%) and myristate (10%). In Merz medium, palmitate represented 51% and 32% at 3 h and 18 h, respectively, while oleate was 25% and 27% at these times. Linoleate was labelled much less (3% and 7% at 3 h and 18 h, respectively) than in the pea medium. SX 623509 at concentrations up to 100 µm did not alter the fatty acid labelling patterns (data not shown).
We then examined labelling of intact lipid classes by separating them by TLC. Data for polar lipids is shown for labelling at 3 h and 18 h using pea medium in Fig. 4. Although the proportion of total labelling in neutrals compared with total polars was unchanged by SX 623509, there were alterations in the patterns of both groups. For polar lipids the most noticeable changes were a decrease in phosphatidylcholine (PtdCho) and an increase in phosphatidate (PtdOH). These alterations became more obvious by 18 h (Fig. 4b). It was of note that, while the radiolabelling of polar lipid classes in control incubations was generally in keeping with their abundance (Griffiths et al., 2003), SX 623509 caused the proportion of phosphatidate to increase markedly (nearly 30% of total polar lipids at 18 h: Fig. 4b). In Merz medium, PtdCho was again the most labelled polar lipid and SX 623509 reduced its labelling significantly (data not shown).
The labelled neutral lipids for P. infestans grown in pea (Fig. 5a) and in Merz media (Fig. 5b) were dominated by triacylglycerol (TAG) with smaller amounts of diacylglycerol (DAG), monoacylglycerol (MAG) and nonesterified fatty acids (NEFA). Incubation with SX 623509 reduced the proportional labelling of TAG with increases in MAG and DAG for growth in pea medium (Fig. 5a) and in DAG and NEFA for growth in Merz medium (Fig. 5b), which were significant after 18 h.
Concentration effects of the mandelamide
As discussed above, the effects of SX 623509 on P. infestans grown on pea or a minimal Merz medium were generally similar. However, one problem with the use of Merz medium is that, because of poor growth in that medium, cultures had to be started in pea medium before transfer. Thus, P. infestans used components that might not have been derived solely from a minimal medium and therefore undefined components from the pea medium, which might have unknown effects on fungicide action, could also be present (albeit in small quantities) in Merz.
Therefore, we also conducted experiments in a different minimal medium, Henniger (Henniger, 1959) as modified (Reich, 1994) which could promote growth from sporangial suspensions and produced enough mycelial material within 2 weeks compared to 1 week's growth in pea medium.
As with the minimal Merz medium, increasing amounts of SX 623509 caused a decrease in PtdCho labelling and an increase in that of PtdOH within the polar lipids for P. infestans grown in both Henniger and CaCl2-deficient Henniger media. The data for the latter are shown in Fig. 6a. Similarly, SX 623509 caused statistically significant increases in the labelling of MAG, DAG and NEFA (Fig. 6b), as had been seen for other media. All these changes were proportional to concentration. In the neutral lipids, TAG labelling appeared to be decreased at 1 µm and 10 µm SX 623509, but this was only significant (P < 0.05) for 1 µm concentrations.
Biosynthesis of phosphatidylcholine
The consistent effects of SX 623509 in reducing PtdCho labelling and increasing that of DAG and PtdOH could have indicated some specific effects on PtdCho biosynthesis. For example, inhibition of cholinephosphotransferase would lead to reduced PtdCho labelling and increase that of DAG while PtdOH labelling would be increased by feedback in the Kennedy pathway (Gurr et al., 2002). Accordingly, we examined PtdCho biosynthesis in more detail.
We used [methyl-14C]choline as a precursor and, in untreated P. infestans cultures, about 92% of the radiolabel was taken up by 6 h. Of this label, 68% was incorporated into lipid, with PtdCho the only detectable compound. At 1 µm, SX 623509 decreased labelling by 44% with only a small (9%) decrease in uptake (data not shown). This fitted with the proposal (above) that the CDP-choline pathway to PtdCho was being inhibited, possibly at the level of the cholinophosphotransferase.
The water-soluble choline intermediates were separated by chromatography and Fig. 7 shows that increasing concentrations of SX 623509 caused a significant rise in the radioactivity associated with CDP-choline. This compound, which represented the smallest pool of intermediates in untreated P. infestans, increased about fourfold at 1 µm SX 623509 (Fig. 7). This was additional evidence that the mandelamide compound was inhibiting the cholinephosphotransferase reaction.
Biosynthesis of PtdCho by methylation
Phosphatidylcholine can be made by various methylation pathways (e.g. Williams & Harwood, 1994) in addition to the CDP-choline pathway (Gurr et al., 2002). In some organisms, the methylation route (especially from phosphatidylethanolamine) can be very important or the exclusive pathway. In lipid separations (Griffiths et al., 2003) as well as in [1–14C]acetate labelling experiments, we noted that methylated derivatives of PtdEtn were present in P. infestans. These results suggested that the methylation pathway for PtdCho formation was important in this fungus.
In order to gain more knowledge of this, we used [2–14C]ethanolamine labelling. In untreated P. infestans, about 50% of the total radiolabel was taken up by the oomycete within 16 h in Henniger medium (Table 1). Of the [2-14C]ethanolamine taken up, about 50% was incorporated into lipids (Table 1) of which nearly 70% was in PtdCho with the remainder distributed evenly between phosphatidylethanolamine (PtdEtn) and DiMePtdEtn (Dimethylphosphatidylethanolamine) with a small amount of MePtdEtn (Methylphosphatidylethanolamine) (data not shown). Incubation with SX 623509 caused a large decrease in [2-14C]ethanolamine uptake and a small reduction in incorporation at 1 µm (Table 1). These two effects were reflected in an inhibition of lipid labelling by about 65% (Table 1). There was no significant change in the relative amounts of PtdCho and PtdEtn radiolabelling, indicating that SX 623509 did not affect the methylation pathway directly. It did, however, decrease PtdCho labelling through methylation by virtue of the reduced labelling of the precursor PtdEtn. Although this inhibition seemed to mainly be at the level of [2-14C]ethanolamine uptake, an additional reduction of PtdEtn formation by the CDP-base pathway also seemed to take place. This would agree with the decreased labelling of PtdEtn from [1-14C]acetate caused by SX 623509 in some experiments (Fig. 4b).
Table 1. The SX623509 (N-2′-(4″-ethoxy-3″-methoxy) phenylethyl-3, 4-dichloromandelamide) pesticide reduces uptake of [2-14C] ethanolamine and its incorporation into lipids in Phytophthora infestans
Radiolabelling (dpm × 10−3)
Liquid cultures of P. infestans (see the Materials and Methods) were preincubated with SX 623509 for 1 h, followed by addition of radiolabel and incubation for 18 h. Results are mean ± SD (n = 3).
257.1 ± 5.9
126.1 ± 9.8
37.8 ± 3.5
85.8 ± 7.1
244.0 ± 9.9
120.9 ± 12.2
37.7 ± 3.4
80.6 ± 8.1
111.7 ± 10.6
44.4 ± 7.2
11.4 ± 2.6
30.6 ± 4.2
Compound SX 623509 was found to inhibit lipid biosynthesis in general but, in addition, it produced some specific qualitative changes in radiolabelling. Radial growth was inhibited at 0.5 µm and stopped at 1 µm fungicide (Fig. 2). Thus, because specific lipid effects were also found at such concentrations, we believe that they are not merely secondary to growth inhibition and therefore could be important for the mode of action of compound SX 623509.
SX 623509 inhibited uptake of [14C]acetate at longer incubation times (Fig. 3a) but this could have been due, at least in part, to the decreased growth and viability of cultures. However, significant effects were seen on lipid labelling even at 3 h (Fig. 3b). The inhibition of incorporation of radioactivity from [1-14C]acetate into lipids included some specific effects. Thus, while the major membrane lipid, PtdCho, was consistently less labelled (Figs 4a and 6a), the proportion of label in phosphatidate was markedly increased (Figs 4b and 6b). Phosphatidate, an intermediate of the biosynthetic (Kennedy) pathway for complex acyl lipids does not accumulate in healthy tissues. Its accumulation pointed to specific inhibitions caused by SX 623509 in the later stages of lipid formation. The simplest explanation of the decrease in PtdCho (and PtdEtn) labelling was that the final enzyme of the CDP-base pathway was being inhibited. This would account for the accumulation of DAG (Figs 5 and 6b) and, because of feedback inhibitions in the Kennedy pathway, for an increased labelling of phosphatidate. Abnormal accumulation of DAG might also provide an opportunity for lipase (and acylhydrolase) catabolism, which could have led to the rise in MAG and NEFA labelling (Figs 5 and 6b). Any increase in such degradative enzymes would also be likely to reduce TAG labelling.
By following the incorporation of radioactivity from [methyl-14C]choline into PtdCho we obtained independent confirmation that the synthesis of PtdCho de novo by the CDP-base pathway was being inhibited. Furthermore, the specific accumulation of CDP-choline (Fig. 7) was entirely consistent with cholinephosphotransferase being a site of action for the fungicide.
In the work described here, the inhibition of PtdCho labelling caused by SX 623509 was usually seen for PtdEtn also (Figs 4b and 6a, Table 1). This is consistent with an inhibition of ethanolaminephosphotransferase in addition to cholinephosphotransferase. In animals, separate phosphotransferases have been reported for the CDP-choline and CDP-ethanolamine pathways (Pelech & Vance, 1984; Cornell, 1992). In yeast (Saccharomyces cerevisiae) two genes have been reported, one coding for an enzyme that was highly selective for CDP-choline and one that could use either substrate (Hjelmstad & Bell, 1991). In plants, while indirect evidence suggested that two selective enzymes were present (Dykes et al., 1976), several studies with crude enzyme preparations indicated that the same enzyme may use both CDP-choline and CDP-ethanolamine (Lord, 1975; Justin et al., 1985). Two genes characterized in Arabidopsis thaliana both encoded enzymes with choline- and ethanolaminephosphotransferase activity. However, one of the enzymes had a greater preference for CDP-choline over CDP-ethanolamine (Goode & Dewey, 1999). Although the inhibition of PtdCho labelling by SX 623509 was generally mimicked by a reduction of that of PtdEtn, the lack of consistency (Fig. 4a) may suggest that there is more than one phosphotransferase in P. infestans. However, such a conclusion needs further work to substantiate it.
An inhibition of PtdCho synthesis has been reported as a target of the commercial phosphorothiolate fungicides, epifenphos and iprobenphos against Pyricularia oryzae (rice blast) (Akatsuka et al., 1977; Kodama et al., 1979, 1980; Copping & Hewitt, 1998). However, these fungicides appear to target the methylation pathway of PtdCho biosynthesis rather than the CDP-base route. In P. infestans, the similar percentage inhibition of PtdEtn and PtdCho labelling from [2-14C]ethanolomine (Table 1) suggests that methylation is not significantly affected by SX 623509. Nevertheless, the importance of methylation as a route for PtdCho biosynthesis in P. infestans may mean that other pesticide classes which target this pathway could be efficacious in treating infections.
Previous studies with mandelamide compounds on P. infestans (AgrEvo (UK) Ltd, Chesterford Park, UK: unpubl. data) failed to detect any significant effect on nucleic acid or protein synthesis or on carbohydrate metabolism. The inhibition of lipid formation at similar concentrations of mandelamides that caused growth effects suggested that the former could be a possible mode of action. Moreover, the selective effects of SX 236509 on lipid metabolism suggest that such activity is not just a secondary reaction to growth inhibition. Whether the inhibition of formation of major membrane components such as PtdCho alone is enough to explain the pesticidal activity of mandelamides against P. infestans needs further research. However, the ability of commercially successful herbicides to target primary pathways of lipid metabolism (Harwood, 1991) suggests that, in future, other pesticides will be developed which have similar sites of action.
We are grateful to the BBSRC and AgrEvo (UK) Ltd. for a CASE studentship to R. G. G.