This is the first report of variability in sensitivity of Phaeosphaeria nodorum to the fungicide azoxystrobin, and also reports on sensitivity to propiconazole, prothioconazole and cyprodinil. An in vitro sensitivity test of 42 isolates from five Swedish winter wheat fields, collected in 2003–2005, was performed on malt extract agar amended with six concentrations of each fungicide. Four isolates collected during the early 1990s, before azoxystrobin had been commercially used in agriculture, were used as references. Fragments of DNA from 231 isolates, including the reference isolates, were sequenced for the genes of cytochrome b and CYP51 in search for amino acid substitutions known to cause loss of sensitivity to strobilurins and triazoles, respectively. The majority of the P. nodorum isolates possessed the amino acid substitution G143A, associated with loss of sensitivity in fungi to strobilurins, except in one field where only half of the isolates had the substitution. The EC50 values varied between 0·66 mg L−1 to estimations far above 1000 mg L−1, with an estimated median value of 366 mg L−1. The EC50 values of the reference isolates ranged from 0·02 to 80·72 mg L−1. The P. nodorum population is still sensitive to propiconazole, prothioconazole and cyprodinil, even though some isolates varied in sensitivity to triazoles. Part of the CYP51 gene, a possible target for triazole sensitivity, was sequenced but no nonsynonymous substitutions were found.
Various kinds of fungicides have been used for several decades against many economically important diseases including foliar diseases of wheat. Triazoles have been one of the major substances applied to wheat since the 1970s, and the strobilurins from the late 1990s to the present. During recent years, farmers have experienced a decline in efficiency of these substances, since leaf spots have continued to develop despite fungicide application. Studies have shown that two important leaf spotting fungi, Mycosphaerella graminicola (anamorph Septoria tritici) and Pyrenophora tritici-repentis (anamorph Drechslera tritici-repentis), causing septoria leaf blotch and tan spot respectively, have become less sensitive to fungicides based on strobilurins (Fraaije et al., 2005; Sierotzki et al., 2007) and triazoles (Cools et al., 2006; Jørgensen, 2007). So far, only one report on sensitivity to propiconazole in Phaeosphaeria nodorum (anamorph Stagonospora nodorum), causing stagonospora nodorum blotch on wheat, has been published (Peever et al., 1994).
Triazoles and strobilurins have different modes of action. Triazoles are demethylation inhibitors (DMIs) since they inhibit the enzyme 14-α-sterol demethylase encoded by the CYP51 gene, a member of the cytochrome P450 family (Siegel, 1981; Yoshida & Aoyama, 1987). This enzyme catalyses the production of ergosterol, an important part of the cell membrane (Mercer, 1984; Bossche et al., 1987). Inhibition of 14-α-sterol demethylase leads to accumulation of eburicol to toxic levels which halts the growth of the fungus. Several important pathogenic fungi have become less sensitive to triazoles and the mechanism is associated with nucleotide changes leading to amino acid substitutions in the CYP51 gene (Zhan et al., 2006; Fraaije et al., 2007; Leroux et al., 2007; Brunner et al., 2008).
The mode of action of strobilurins, such as azoxystrobin, is to inhibit mitochondrial respiration by binding to the Qo site (outer quinine oxidizing pocket) at the mitochondrial membrane protein cytochrome b. The electron transfer process between cytochrome b and cytochrome c1 is blocked, leading to disruption of ATP production needed for metabolism (Bartlett et al., 2002). Decreased metabolism prevents spore germination and inhibits mycelial growth during the latency period, which are two energy demanding processes (Wong & Wilcox, 2001).
Reduced sensitivity to strobilurins in pathogenic fungi is associated with three amino acid substitutions in the cytochrome b gene, referred to as F129L, G137R and G143A (Grasso et al., 2006; Sierotzki et al., 2007). The substitution G143A caused the highest EC50 values and is the most common one found among all species that have lost their sensitivity to strobilurins.
Cyprodinil belongs to the fungicide group anilinopyrimidines (Knauf-Beiter et al., 1995). The biological mode of action includes inhibition of methionine biosynthesis in the pathogen and the elongation of hyphae. The chemical mode of action is still unknown but research has been performed on resistant strains of Botrytis cinerea in order to find the candidate gene for the loss of sensitivity (Fritz et al., 2003).
The overall goal of this study was to investigate the sensitivity profile of the Swedish population of P. nodorum to the active substances azoxystrobin, propiconazole, prothioconazole and cyprodinil, which are frequently used in fungicides. The following specific hypotheses were tested: i) the current population of P. nodorum is less sensitive to the active substances compared to the isolates collected during the 1990s; ii) amino acid substitutions in cytochrome b have led to strains of P. nodorum which are less sensitive to strobilurins; iii) mutations and deletions in CYP51 have led to strains of P. nodorum which are less sensitive to triazoles; iv) there is cross resistance against triazoles; and v) there are no correlations between mating type and sensitivity to fungicides.
Materials and methods
The active substances azoxystrobin, prothioconazole, propiconazole and cyprodinil were dissolved in methanol and blended in autoclaved 3% malt extract agar (MEA)(14 g agar L−1). A pilot study was first performed with four isolates and twelve concentrations of each fungicide to set the concentrations for the sensitivity test. The concentrations used in the sensitivity test were as follows: azoxystrobin: 1, 5, 10, 50 and 100 mg L−1; propiconazole: 0·1, 0·5, 1, 5 and 10 mg L−1; prothioconazole: 0·1, 1, 5, 10 and 50 mg L−1 and cyprodinil: 0·01, 0·05, 0·1, 0·5 and 1·0 mg L−1.
Forty-four conidial and two ascospore isolates of P. nodorum were used in this sensitivity test. Forty-two of the isolates, including the ascospore isolates, originated from field collections and will hereafter be referred to as the field isolates. Collections of diseased winter wheat leaves were made during mid July in 2003, 2004 and 2005 in five fields (located in three different regions in Sweden), with plots (20 × 20 m) that were not treated with fungicides (Blixt et al., 2008). The ascospores were caught in 2005. The mating types of the isolates were identified earlier in a genetic study (Blixt et al., 2008). Two isolates (Sn2-90 and Sn3-90) were collected in 1990 outside of Uppsala, and were used as reference isolates since they had not been exposed to strobilurins. Two other reference isolates, Sn 120 and Sn 415, were collected during the mid 1990s and were only used in the test with azoxystrobin. Four Danish isolates of M. graminicola were used on agar plates amended with prothioconazole and azoxystrobin to confirm the methodology since two of them were known to have lost their sensitivity to strobilurins (Jørgensen, 2007).
Winter wheat seeds were sterilized in hydrogen peroxide for 2 min, rinsed in sterile water and left to germinate in an autoclaved beaker filled with 3 cm of Murashige & Skoog agar. The beaker, sealed with parafilm, was placed in a growth chamber for 10 days. The seedling leaves formed were cut in c. 1 cm pieces, washed in 70% ethanol for two minutes, rinsed in sterile water for two minutes and placed on 3% MEA plates. A small piece of P. nodorum mycelium (c. 1 mm2) grown in vitro was put on top of the leaf and the fungus was grown under controlled conditions (continuous near ultraviolet light, 18°C) for three weeks. Agar plugs (4 mm diam.) were then cut from the edge of the cultures and put upside down on top of the intercept of the two perpendicular lines drawn on the back of agar plates. The plates were sealed and placed in a growth cabinet under the same conditions as above. The radial growth of the fungi was marked along the two perpendicular lines after 4, 8, 11, 15, 18 and 22 days. The experiment was performed on two occasions; propiconazole and cyprodinil were used in the first round whereas prothioconazole and azoxystrobin were used in the second round. Three replicates were used per fungicide concentration. Five control plates were used in the first round and four control plates were used in the second.
DNA fragments within the mitochondrial gene encoding cytochrome b and the gene CYP51 were sequenced for 227 P. nodorum isolates originating from the collections mentioned previously. The DNA extractions originated from the study presented in Blixt et al. (2008). The four reference isolates were also sequenced after DNA extraction as mentioned above.
PCR primers for the target regions (cytochrome b: GenBank Accession No. EU053989 amino acid position 46-218; cytochrome P450: GenBank Accession No. XM001794202.1 amino acid position 364–512) were selected using Primer3 (http://primer3.sourceforge.net). Cytochrome b forward, 5′-CAGGTGTAACATTAGCGATGC-3′, reverse, 5′-TGATTGTCCCCATGTCAATG-3′, and cytochrome P450 (CYP51): forward 5′-AAGGAGACTCTTCGCATCCA-3′, reverse, 5′-GCGGTCTCGAGAACAGACTT-3′.
The PCR solution of 50 µL contained 0·25 ng DNA µL−1, 0·75 mm MgCl2 (final concentration), 0·2 mm dNTP, 0·2 µm of each primer, 0·03 U µL−1 ThermoRed DNA polymerase (Saveen & Werner AB) and corresponding reaction buffer Y. The PCR conditions were 96ºC for 2 min, 30 cycles of 30 s at 96°C, 30 s at 58°C and 30 s at 72ºC, followed by a 5 min extension (RoboCycler® Gradient 96 Temperature Cycler, Stratagene). Negative controls were included in every PCR. The PCR products were separated on 1% agarose gel (Agarose D-1, Conda), half strength TBE, at 3·3 V cm−1 for 60 min. The gels were stained with ethidium bromide and visually analysed under UV light (GelDoc, Bio-Rad Laboratories).
The PCR products were purified using the Agencourt® AMPure® Protocol 000601v024. The analysis of the sequences was performed by Macrogen Inc. or with a Beckman Coulter CEQTM 8000 Genetic Analysis System according to the protocol from the manufacturer. In the latter system the PCR products for sequencing were purified using Sephadex G-50 Medium (GE Healthcare Bio-science).
The radial growth of the isolates was calculated as the colony diameter minus the diameter of the inoculation plug and then divided by two. To avoid the influence of growth variations between individual isolates, the radial growth of each isolate and treatment was calculated as a proportion of the average radial growth on the control plates on day 18. EC50 values were estimated using the equation for the trend line in a chart made from the proportion of growth and the logarithm of the concentration of each fungicide. The resistance factor (RF) was calculated by dividing the EC50 value of each isolate by the average EC50 value of two of the reference isolates (Sn2-90 and Sn3-90) for propiconazole, prothioconazole and cyprodinil. For azoxystrobin the average EC50 value of the four reference isolates (Sn2-90, Sn3-90, Sn 120 and Sn 415) was used for comparisons with the field isolates. No RF values were calculated for azoxystrobin due to variation of the EC50 values of the reference isolates.
The growth rate of each isolate was calculated from the average radial growth between day 8 and day 18 divided by the time interval (10 days). Statistical analyses of the results obtained were performed using the GLM procedure in SAS V9·1 (SAS Institute Inc., 2004).
The P. nodorum isolates were able to grow on agar amended with either of the four fungicides but with varying growth rates and degrees of inhibition. The majority of the isolates had a lag phase between inoculation and day 8, but grew thereafter at constant growth rate throughout the experiment. The growth rate on the control plates varied between 0·9 and 2·4 mm per day, with an average of 1·6 mm per day. Some of the cultures reached the edge of the control plates prior to day 22 and thus the radial growth reading at day 18 was used for calculations of EC50 values and growth rates.
Sensitivity to propiconazole and prothioconazole
The radial growth of P. nodorum was inhibited by propiconazole and prothioconazole (Fig. 1a,b). The isolates are ranked in descending order according to their growth rates on the fungicide-free control plates of the propiconazole test. The majority of the isolates were not able to grow on MEA amended with the two highest concentrations of triazoles or they just formed a small mycelial ball over the inoculation plug. One isolate grew faster (> 0·1 mm per day) on MEA amended with 0·1 mg L−1 propiconazole and seven isolates grew faster on MEA amended with 0·1 mg L−1 prothioconazole than on their respective control plates.
The sensitivity to propiconazole and prothioconazole varied between the field isolates. The EC50 values for propiconazole varied between 0·03 and 2·09 mg L−1 (Fig. 2a) and between 0·30 and 4·27 mg L−1 for prothioconazole (Fig. 2b). The average EC50 values of the two reference isolates (Sn2-90 and Sn3-90) were 0·19 mg L−1 for propiconazole and 0·37 mg L−1 for prothioconazole. Five of the isolates had RF scores above five for propiconazole and two other isolates had an RF above five for prothioconazole. These seven isolates originated from two fields.
No correlation was found between the EC50 values for the two triazoles (Fig. 3, R2 = 0·02). The EC50 values for propiconazole were in general lower than for prothioconazole with median EC50 values of 0·25 mg L−1 and 0·89 mg L−1 for propiconazole and prothioconazole, respectively. The reference isolate Sn2-90 and three field isolates, originating from different fields, had similar EC50 values for both substances. There were no significant differences between the EC50 values of the field isolates and the reference isolates for either propiconazole or prothioconazole.
Substitutions found in CYP51
The gene homologue of CYP51 associated with sensitivity to triazoles in M. graminicola was found in P. nodorum. Three synonymous substitutions were found in the analysed sequence of 445 bp of CYP51; TCA to TCG at amino acid position 375, ACC to ACT at position 379 and TAC to TAT at position 474. The two former substitutions were unique observations while the substitution at position 474 was found in three isolates. These isolates had EC50 values between 0·15 and 0·25 mg L−1 for propiconazole and between 1·10 and 2·31 mg L−1 for prothioconazole.
Sensitivity to cyprodinil
All isolates were inhibited by cyprodinil at the higher concentrations. The average growth rate of the field isolates was 0·25 mm per day on MEA amended with 1 mg L−1 cyprodinil, compared to 1·6 mm per day on average for the control cultures. Eight of the field isolates grew faster on MEA amended with 0·01 mg L−1 cyprodinil compared to control (Fig. 1c). Three of those eight isolates also grew faster on 0·05 mg L−1. One isolate grew only on the lowest concentration, 0·01 mg L−1 and it was not possible to determine the EC50 value. The EC50 values for cyprodinil varied between 0·02 and 0·43 mg L−1 with a median of 0·09 mg L−1 (Fig. 2c). The average EC50 value of the two reference isolates (Sn2-90 and Sn3-90) was 0·06 mg L−1. Four isolates had a resistance factor above five, one of which also had a resistance factor above five for propiconazole.
Sensitivity to azoxystrobin
All isolates grew on agar containing azoxystrobin at all concentrations but with varying growth rates (Fig. 1d). The average growth rate of the field isolates was 1·1 mm per day on MEA amended with 100 mg L−1 compared to the control growth rate (1·6 mm per day) during day 8 to 18. Six isolates grew faster on the two lowest concentrations than in the absence of azoxystrobin. The radial growth of the majority of the isolates stayed at a constant proportion compared to the control treatment from day 8 and onwards. The estimated EC50 values for azoxystrobin were between 0·66 mg L−1 and far above 1000 mg L−1, with a median value of 366 mg L−1 (Fig. 2d). The EC50 values for azoxystrobin were set to 400 during further calculations if the estimated value was > 400 mg L−1. The EC50 values of the reference isolates ranged from 0·02 mg L−1 to 80·72 mg L−1 with an average of 42·48 mg L−1; significantly lower than the EC50 values of the field isolates (P < 0·0001). The reference isolate Sn2-90 had a drastically reduced growth even at 1 mg L−1.
Substitutions found in cytochrome b
The gene homologue of cytochrome b associated with sensitivity to strobilurins in M. graminicola and P. tritici-repentis was found in P. nodorum. The substitution G143A was found among the P. nodorum isolates from the five fields. No other nucleotide substitution was found in the 518 bp long sequence. The frequencies of isolates with the substitution were 79–98% in four of the fields while only 50% of the isolates collected at Hagby in 2004 had the substitution (Fig. 4). None of the reference isolates had any nucleotide substitution in the sequenced region and are therefore not included in Fig. 4.
Correlations between sensitivity tests and molecular data
The field isolates were sorted by increasing EC50 value and according to the amino acid at position 143 (Fig. 5). The average EC50 value of reference isolates is included in the figure for comparison of the EC50 values and the occurrence of the substitution. The number of field isolates with and without amino acid substitution in comparison to the EC50 value for azoxystrobin is presented in Table 1. The estimated EC50 values of the 11 field isolates without the substitution ranged from 15·53 mg L−1 to > 400 mg L−1, of which 10 had an EC50 value higher than those of the reference isolates. Most of these 10 isolates had different morphology compared to the control plates when grown on high concentration of azoxystrobin. The mycelium on the control agar plates was thick and colourful (Fig. 6a), while it was pale and thin on MEA amended with azoxystrobin (Fig. 6b). Thirty-one of the isolates had the substitution G143A and all, except for three isolates, had an EC50 value higher than the average value of the reference isolates. The morphology of the isolates with the substitution and an EC50 value higher than the reference isolates was similar on MEA plates with and without azoxystrobin although the growth rates were reduced (Fig. 6c,d).
Table 1. Number of Phaeosphaeria nodorum field isolatesa with EC50 values higher and lower than the average of the reference isolates, subdivided by the amino acid at position 143 in the gene encoding cytochrome b
The sensitivity of the isolates to the four substances was found to be independent of the mating type. The distribution of the mating types for the 227 isolates sorted by the amino acid at position 143 in cytochrome b was found to be in a 1:1 ratio (Fig. 4). One inconsistency was found: 21 out of 28 isolates from Hagby 2004 with the substitution had mating type MAT1-2, while the isolates without the substitution had a 1:1 mating type ratio (P < 0·05). Among those 21 isolates 15 different genotypes were identified. The mating types of the reference isolates were not determined.
This is the first report of the substitution G143A in the cytochrome b gene in P. nodorum. This gene mutation, which is known to influence sensitivity to azoxystrobin in fungi, is widespread throughout the populations of P. nodorum in Sweden. There are strong indications that azoxystrobin would be ineffective in controlling P. nodorum, in agreement with earlier studies on other pathogenic fungi (Grasso et al., 2006; Sierotzki et al., 2007).
The triazoles and cyprodinil were, in general, effective in inhibiting fungal growth. The EC50 values for cyprodinil of the majority of the P. nodorum isolates were lower than the EC50 values needed for germ-tube elongation in M. graminicola (Leroux et al., 2007). The fungal populations tested in this survey had EC50 values for propiconazole in the same range as reported by Peever et al. (1994) but some isolates with higher EC50 values were also found, indicating variation of sensitivity in the fungal population. Seven isolates had a resistance factor (RF) above five for the triazoles and they originated from two fields, Yxstad 2004 and Husby 2005, which were repeatedly treated with triazoles prior to the collection. Reduction of sensitivity to triazoles has been reported to be associated with variations in the ABC transporters for M. graminicola (Stergiopoulos et al., 2003) and P. tritici-repentis (Reimann & Deising, 2005) and may also be the case for the isolates of P. nodorum that showed higher EC50 values reported here. The slightly higher EC50 values for prothioconazole compared to propiconazole could be due to the sulphhydryl-group in prothioconazole that binds to media-components and plastics leading to less active substance in the media (A. Mehl, personal communication).
Reduction of sensitivity to triazoles is probably a complex of different functions since there was no correlation between the EC50 values for propiconazole and prothioconazole (Fig. 3). This may reduce the possibilities for cross-resistance between the two triazoles. The synonymous mutations found at position 474 in the CYP51 region had no impact on the sensitivity of P. nodorum to triazoles, since no correlation between RF values and mutations were found. The hypotheses of a complex of mechanisms and no cross resistance are supported by sensitivity tests performed on Blumeria graminis, M. graminicola and P. tritici-repentis (Wyand & Brown, 2005; Jørgensen, 2007; Jørgensen & Skov Jensen, 2007).
There was a large variation in EC50 values for azoxystrobin among the isolates (Fig. 5). The field isolates in general showed reduced growth rate in the presence of azoxystrobin, but the majority had EC50 values far above the average value of the reference isolates. Many of the isolates without the substitution G143A also had high EC50 values suggesting that there may be other mechanisms that can reduce the efficacy of azoxystrobin.
One possible explanation to the loss of sensitivity to azoxystrobin may be the alternative oxidase pathway. This has been found to reduce the efficacy of azoxystrobin in strains of M. graminicola and Botrytis cinerea which lacked any amino acid substitutions (Ziogas et al., 1997; Wood & Hollomon, 2003; Miguez et al., 2004). Alternative oxidase (AOX) is an enzyme that gives rise to a small amount of electron transportation resulting in sufficient ATP production for the fungus to survive and grow in vitro despite exposure to strobilurins (Wood & Hollomon, 2003). Salicylichydoxyamic (SHAM) has been reported to inhibit the activity of AOX (Ziogas et al., 1997) and addition of SHAM to the agar might have influenced the effect of AOX activity during the sensitivity test.
Three isolates with the G143A substitution had an EC50 value lower than the average of the reference isolates. This may be due to the natural variability of the effect of azoxystrobin on fungi with the G143A substitution. This has been reported in studies of M. graminicola (Lardinois et al., 2006) and P. tritici-repentis (Sierotzki et al., 2007).
The four fields (Yxstad 2003 & 2004, Skofteby and Husby) where the substitution G143A dominated had been treated regularly with strobilurins (azoxystrobin or pyraclostrobin) in the years prior to isolate collection. These populations of P. nodorum have therefore experienced selection for survival in the presence of this group of fungicides. The population at Hagby 2004, with 50% substitutions, had not experienced the same level of selection since this, and the surrounding fields had rarely been treated with fungicides during the last decade.
Substitutions of amino acids leading to loss of sensitivity to fungicides are most likely occurring randomly and the isolates with such substitutions will be selected for during strobilurin treatment. The G143A substitution occurring within the European populations of M. graminicola originates from four independent mutation events based on the genetic variation in the mitochondrial DNA (Torriani et al., 2009). The genetic study also showed that less sensitive strains of M. graminicola migrated via wind borne ascospores within Europe in a west-to-east direction. The population of the less sensitive strains of P. nodorum within a field may have been founded by incoming ascospores carrying the G143A substitution, since a previous study concluded that P. nodorum reproduces sexually in Sweden (Blixt et al., 2008). Ascospores could increase the spread of the substitution among fields but the consequences for the P. nodorum population are unclear. The mating type distribution within fields was not correlated to the sensitivity to fungicides. However, the majority of isolates with the G143A substitution from the field at Hagby, 2004 were MAT1-2 (Fig. 2a–d, Fig. 4).
This study indicates that the P. nodorum population in Sweden is not efficiently inhibited in vitro by azoxystrobin although the fungal population is still efficiently inhibited by propiconazole, prothioconazole and cyprodinil. The shift towards isolates less sensitive to strobilurins within the populations sampled could be due to the high frequency of isolates with the G143A substitution. The reduced sensitivity to triazoles in some isolates could not be explained by nonsynonymous substitutions found in CYP51. No cross resistance was found between either of the fungicides. While the use of strobilurins alone for protection of wheat against P. nodorum may not be advisable, triazoles and anilinopyrimidines can probably still provide useful protection, although reliance solely on fungicides for disease control is probably not advisable in modern agriculture.
The active substances were kindly provided by Andreas Mehl (Bayer CropScience), Helge Sierotzki and Maria Rosén (Syngenta) and Kristina Forsberg (Makhteshim-Agan Industries). The authors wish to thank Ingrid Happstadius (Swalöf Weibull AB) for providing the seed, Lise Nistrup-Jørgensen, University of Aarhus, for providing the M. graminicola isolates and useful information and Andreas Mehl and Helge Sierotzki for useful comments on the manuscript. The experimental work could not have been performed without the assistance of Katarina Ihrmark and the youngsters helping out with the sensitivity test. We are also grateful to the Swedish Plant Protection Centres and the farmers for providing information from the sampling locations. The project was financed by the Swedish University of Agricultural Sciences (SLU) and the postgraduate school IMOP (Interactions between Micro-Organisms and Plants) at SLU.