Sdh sequence variation among M.graminicola field strains

DNA sequence comparison between IPO323, IRE30 and five other strains (BC1, BC4, R35‐3, R39‐1 and V212‐2) showed that the Sdh subunits were generally well conserved (GenBank accession numbers JF916683–916700). For SdhB, only changes resulting in the replacement of lysine (aag) by arginine (agg) at position 48 (B‐K48R) and cysteine (tgc) by arginine (cgc) at codon 276 (B‐C276R) were found in strains BC1 and V212‐2, respectively. Strain R35‐3 had a proline (ccc) residue [instead of serine (tcc)] at position 51 in SdhC (C‐S51P), and isolate R39‐1 had two changes in this subunit, a valine (gtc) residue [instead of isoleucine (atc)] at codon 29 (C‐I29V) and a glycine (ggc) residue [instead of arginine (cgc)] at codon 54 (C‐R54G). Two SdhC changes were also found in strain V212‐2, with two asparagine (aac) residues replaced by threonine (acc) at codons 33 (C‐N33T) and 34 (C‐N34T). For SdhD, strain R35‐3 carried a glycine (ggt) residue [instead of alanine (gct)] at codon 10 (D‐A10G), whereas IRE30 harboured a unique proline (ccg) residue [instead of arginine (cgg)] at codon 47 (D‐R47P). SdhC was most variable with regard to additional nucleotide changes leading to synonymous substitutions and sequence variation within intron regions (data not shown).

Docking of carboxin and other SDHIs in the quinone binding site of the M.graminicola Sdh complex

Docking studies (​(2,2, ​,3)3) showed carboxin positioned in a similar manner as in the crystal structure of the avian Sdh complex (Huang etal., 2006). SdhA did not form part of the quinone binding site. SdhB residues P220, S221, W223, W224, H267 and I269 form the major part of the binding pocket. SdhC (L71, W80, S83, A84 and R87) and SdhD (D129 and Y130) form the remainder of the fungicide binding cavity. The docking study also indicates that the binding energy for the carboxin–Sdh complex is relatively high, suggesting relatively poor binding when compared with the other fungicides (Table2). This probably reflects the relatively small size of this molecule, reducing extensive protein–fungicide interactions. Boscalid and bixafen contain an additional phenyl ring compared with carboxin, which is reflected in the lower binding energies. The predicted binding energies of carboxin, boscalid, bixafen and isopyrazam (Table2) are perfectly in line with the respective baseline fungicide sensitivity levels (Table1). The chlorinated phenyl rings of boscalid and bixafen occupy the same binding pocket as the phenyl moiety of docked carboxin (Fig.3A–C). However, the amide moiety is bound in another position, as is the second phenyl moiety of these fungicides, which is located at the surface of the protein. The pyrazole and pyridine moieties of bixafen and boscalid are found in the same pocket as the nonaromatic ring of carboxin. Binding seems to be stronger for bixafen, suggesting that the pyrazole moiety interacts favourably when compared with the pyridine moiety of boscalid, as both fungicides adopt similar binding modes when docked into the Sdh complex. Docking of isopyrazam (Fig.3D) and penthiopyrad (Fig.3E) yields similar binding energies when compared with bixafen (Table2). However, the binding mode more closely resembles the binding of carboxin. The pyrazole moieties occupy the same pocket as the phenyl group of carboxin, whereas their amide and hydrophobic moieties bind in the same location as the amide and the nonaromatic ring of carboxin. To compare the effectiveness of fluopyram insilico, we also docked this novel fungicide. The predicted fluopyram binding is largely similar to the binding of carboxin, as the phenyl ring and the amide moiety are in the same location as the nonaromatic ring and amide of carboxin (Fig.3F). However, the other moiety of fluopyram, a benzamide group, is separated from the amide bond by two carbons. This facilitates extended binding of this fungicide, as the benzamide moiety is favourably stacked in between two tryptophans, SdhB W223 and SdhC W80 (Fig.2). There is no such spacer between the amide and the analogous phenyl group in carboxin. This suggests very tight binding for fluopyram, as the binding energy, −138.6kJ/mol, is the lowest of all the docked SDHIs (Table2).

Open in a separate window

Figure 2

Docking of carboxin and fluopyram in the quinone binding site. The succinate dehydrogenase (Sdh) structural model is shown with the carbon atoms in green, with the residues forming the binding pocket with the carbon atoms in cyan. Carboxin and fluopyram are shown with the carbon atoms in magenta and yellow, respectively. Hydrogen bonds are indicated as broken lines.

Open in a separate window

Figure 3

Binding of succinate dehydrogenase inhibitors (SDHIs) to the quinone binding site of the succinate dehydrogenase (Sdh) complex. Carboxin (A), bixafen (B), boscalid (C), isopyrazam (D), penthiopyrad (E) and fluopyram (F). The electrostatic potential of the protein surface is coloured from red (negative charge) to blue (positive charge).

Table 2

Lowest binding energies obtained by docking of fungicides in a structural model of the wild‐type (WT) succinate dehydrogenase (Sdh) complex from Mycosphaerella graminicola and fungicide resistance‐related Sdh variants.

Inhibitor*	Binding energies for each complex (kJ/mol)†
	WT	S221P SdhB	R265P SdhB	H267L SdhB	H267Y SdhB	H267N SdhB	I269V SdhB	N86K SdhC	I127V SdhD	D129E SdhD	D129G SdhD
Boscalid	−113.0	−114.9	−114.2	−111.8	−112.7	−112.5	−112.8	−114.8	−114.6	−116.1	−110.5
Carboxin	−92.0	−92.5	−90.2	−90.0	−93.1	−89.1	−88.7	−90.3	−90.5	−91.1	−84.0
Bixafen	−121.9	−129.4	−122.6	−121.8	−122.2	−125.1	−120.0	−124.8	−122.7	−121.7	−122.1
Isopyrazam	−122.8	−115.5	−119.8	−115.2	−113.0	−126.4	−123.4	−118.9	−123.4	−118.6	−136.2
Penthiopyrad	−122.4	−120.5	−122.2	−120.2	−119.2	−122.6	−121.1	−124.1	−122.2	−121.1	−125.6

Open in a separate window

^* The S‐enantiomer of penthiopyrad and the (C6 S, C10 R, C21 S)‐enantiomer of isopyrazam were used for docking studies.

^† Higher binding energies in comparison with the wild‐type (WT) Sdh variant indicate a weaker binding of the inhibitor to the target. Values with a significant increase (>1kJ/mol) are presented in bold.

Identification of Sdh mutations in carboxin‐resistant UV mutants

In total 124 mutants, 98 derived from IPO323 (Table3) and 26 from IRE30 (Table4), were isolated and characterized. Mutants in the collection had a wide range of nonsynonymous mutations (Fig.4) and corresponding SDHI sensitivity profiles. Some rare mutants had multiple mutations in the same or in different Sdh subunit encoding genes. Figure4 shows the position of nine key amino acids that were substituted in SDHI‐insensitive mutants of M.graminicola, and how conserved are these residues in SdhB, SdhC and SdhD sequences from A.alternata, As.oryzae, B.cinerea, C.cassiicola and chicken (Gallus gallus). SdhB‐C137R is unique for M.graminicola and no mutations have been found at or in close proximity to this position in SDHI‐resistant strains of other fungi. This residue is normally part of an FeS cluster, relatively far from the quinone binding site, and it is expected that such a mutation would result in a nonfunctional protein. SdhB codon S221 (S221P/T) is one position away from codon 225 in B.cinerea, where substitutions P225L/F/T have been reported for field strains (Leroux etal., 2010; Stammler etal., 2008). Codon 230 in B.cinerea, where N230I has been linked to SDHI resistance (Leroux etal., 2010), is positioned next to W224 in M.graminicola, a residue predicted to form part of the quinone binding site (Fig.2). SdhB codons 265 (R265P), 267 (H267F/L/N/Y) and 269 (I269V) are clustered together, and mutations resulting in H267Y/L/N have also been reported in a range of other fungi (Avenot and Michailides, 2010). SdhC N80K, reported for Coprinus cinereus (Ito etal., 2004), is identical to N86K in M.graminicola. In addition, SdhC L85P and N86S were also found in M.graminicola mutants. SdhC S73P and T90I, reported for C.cassiicola (Miyamoto etal., 2010) and As.oryzae (Shima etal., 2009), respectively, are within three positions of codons 85 and 86 in M.graminicola. SdhD codons 127 (I127V) and 129 (D129E/G/T) are clustered together. SdhD D89G and D124E, reported in Paracoccus denitrificans (Matsson etal., 1998) and As.oryzae (Shima etal., 2009), respectively, are equivalent to D129G and D129E in M.graminicola. G109V in C.cassiicola (Miyamoto etal., 2010), H133R in A.alternata (Avenot etal., 2009) and H132R in B.cinerea (Leroux etal., 2010) are part of the α‐helix that positions the two crucial quinone binding site residues in M.graminicola, SdhD D129 and Y130. However, they are too far from the quinone binding site (approximately 12Å) to have a direct effect on fungicide binding.

Table 3

Overview of key mutations and associated amino acid substitutions in succinate dehydrogenase (Sdh) complexes of IPO323‐derived mutants (n= 98) and their succinate dehydrogenase inhibitor (SDHI) sensitivity profiles.

Strains/mutant Sdh variants	Corresponding codon changes	Mutant number	Fungicide sensitivity range (EC50 in µg/mL)
			Carboxin†	Boscalid	Bixafen	Isopyrazam
IPO323			3.50	0.718	0.216	0.316
B‐C137R	tgt > cgt	1	159	>12.5	2.01	1.93
B‐S221P (M152)*	tcc > ccc	1	11.8	0.344	0.014	<0.001
B‐S221P*	tcc > ccc	5	102–190	>12.5	0.761–1.91	0.889–2.34
B‐S221T	tcc > act	2	81.7–113	0.959–2.0	0.465–0.583	0.742–1.11
B‐R265P	cga > cca	6	133–175	3.26–5.0	0.235–1.02	0.644–1.83
B‐H267F	cac > ttc	1	58.1	>12.5	2.2	3.13
B‐H267L	cac > ctt/ctc	9	>200	>12.5	>12.5	10.0–>12.5
B‐H267N	cac > aac	5	81.0–131	0.662–1.33	0.284–0.787	0.451–2.09
B‐H267Y	cac > tac	48	67.7–193	>12.5	0.482–3.05	0.232–4.6
B‐I269V	att > gtt	7	63.5–143	1.1–2.86	0.397–1.03	1.15–1.86
B‐S221P, C‐R54G	tcc > ccc, cgc > ggc	1	102	2.65	0.462	1.08
B‐H267Y, C‐N86S	cac > tac, aac > agc	1	153	>12.5	1.46	2.11
B‐D166G, D‐D129G	gac > ggc, gac > gga	1	>200	>12.5	4.62	5.13
B‐I13V, D‐D129E, D‐V154A	att > gtt, gac > gaa, gtt > gct	1	40.0	3.08	0.969	1.44
C‐L85P	ctc > ccc	1	85.1	1.64	0.792	0.669
C‐N86K	aac > aaa	1	112	>12.5	>12.5	>12.5
D‐I127V	atc > gtc	1	85.1	1.86	0.433	1.27
D‐D129E	gac > gag/gaa	5	85.1–148	11.6–>12.5	0.412–2.15	0.666–4.07
D‐D129T	gac > acc	1	>200	3.41	3.49	3.6

Open in a separate window

^* Mutants with B‐S221P showed two different phenotypes. One mutant derived from IPO323 (M152) was extremely sensitive to bixafen and isopyrazam. The other B‐S221P mutants (n= 5) were more resistant to all SDHIs tested.

^† The carboxin sensitivity reported for strain IPO323 was higher than that reported in Table4 because of the wider range of fungicide concentrations tested to measure the high resistance levels in the mutants.

Table 4

Overview of key mutations and associated amino acid substitutions in succinate dehydrogenase (Sdh) complexes of IRE30‐derived mutants (n= 26) and their succinate dehydrogenase inhibitor (SDHI) sensitivity profiles.

Strains/mutant Sdh variants*	Corresponding codon changes	Mutant number	Fungicide sensitivity range (EC50 in µg/mL)†
			Carboxin	Boscalid	Bixafen	Isopyrazam
IRE30			2.82	0.389	0.070	0.088
B‐S221P	tcc > ccc	5	57.8–108	0.447–1.09	0.058–0.243	0.096–0.338
B‐H267N	cac > aac	4	64.7–99.3	1.11–1.88	0.284–0.81	0.398–2.09
B‐H267Y	cac > tac	14	69.0–155	>12.5	0.777–2.6	0.882–4.04
B‐P155L, B‐H267Y	cca > cta, cac > tac	1	23.6	0.878	0.682	0.486
C‐L85P, D‐V96A	ctc > ccc, gtc > gcc	1	53.1	3.33	0.874	1.75
D‐D129E	gac > gag	1	149	>12.5	1.25	2.4

Open in a separate window

^* Strain IRE30 and related mutants all carry D‐R47P in comparison with strain IPO323.

^† The carboxin sensitivity reported for strain IRE30 was higher than that reported in Table4 because of the wider range of fungicide concentrations tested to measure the high resistance levels in the mutants.

Open in a separate window

Figure 4

Positions of key amino acid alterations in different succinate dehydrogenase (Sdh) subunits from succinate dehydrogenase inhibitor (SDHI)‐resistant mutants of Mycosphaerella graminicola. Partial Sdh subunit sequences are presented for M.graminicola (Mg), Alternaria alternata (Aa), Aspergillus oryzae (Ao), Botrytis cinerea (Bc), Corynespora cassiicola (Cc) and Gallus gallus (Gg). Positions of key amino acid residues are marked with a star. Shaded residues are amino acid residues that have been altered and linked with SDHI resistance. Bold amino acids are conserved residues.

Relationship between mutant Sdh variants and SDHI sensitivity

​3,3, ​,44 show the relationship between the different Sdh variants and the sensitivity ranges measured for the different SDHI fungicides. Mutants carrying B‐S221P are clustered into three different phenotypic groups depending on the parental strain. In comparison with IRE30 mutants, B‐S221P mutants derived from IPO323 have higher levels of resistance to all SDHIs tested, in particular for boscalid with EC₅₀ levels above 12.5µg/mL. For one IPO323‐derived mutant, M152, carrying S221P with no other mutations detected in the SdhB, SdhC and SdhD genes, the sensitivity to bixafen and isopyrazam was much lower than for IPO323. For all other Sdh variants tested, there was no clear strain‐dependent difference in SDHI sensitivity. Most mutants with multiple mutations in Sdh genes (B‐P155L/B‐H267Y, B‐S221P/C‐R54G and B‐I13V/D‐D129E/D‐V154A) were more sensitive to most SDHIs tested relative to mutants carrying single amino acid substitutions (B‐S221P, B‐H267Y and D‐D129E). Mutants carrying variant B‐H267L were highly resistant to all four SDHIs tested. High levels of boscalid resistance were measured for variants carrying B‐C137R, B‐S221P (only IPO323‐derived mutants), B‐H267F, H267Y, C‐N86K, D‐D129E and D‐D129G. Sensitivity levels to bixafen and isopyrazam were similar for mutants carrying different Sdh variants, with the exception of B‐I269V and B‐S221T mutants that were inhibited more strongly with bixafen than isopyrazam. Interestingly, high levels of resistance to both bixafen and isopyrazam were only measured for mutants carrying B‐H267L or C‐N86K.

Generation of SDHI‐resistant UV mutants

Ten thousand spores of M.graminicola strains IPO323 or IRE30 were plated out onto YEPD agar amended with carboxin at 50µg/mL. Plates were exposed to UV light at 500 or 600J/m² using a Stratagene UV cross‐linker (model 1800; Stratagene Corporation, La Jolla, CA, USA), which resulted in approximately 20% survival. Putative carboxin‐resistant mutants were isolated from growing colonies after 14 days of incubation at 21°C in the dark. These mutants were grown for another generation on carboxin‐amended YEPD agar, harvested in sterile distilled water and stored in 80% (v/v) glycerol at −80°C for further use.

Isolation of fungal genomic DNA, PCR and DNA sequencing

Genomic DNA for PCR was extracted and quantified according to the protocol of Rudd etal. (2010). To amplify genes encoding for SdhB, SdhC and SdhD, specific primers for each subunit (see Table6) were designed based on the available genome sequence for strain IPO323 downloaded from the Joint Genome Institute (JGI) M.graminicola genome website (URL: http://genome.jgi‐psf.org/Mycgr3/Mycgr3.home.html). PCR was carried out in an Applied Biosystems (Foster City, CA, USA) Veriti 96‐well thermal cycler in a final volume of 30µL containing 2.5 units of Easy‐A high‐fidelity PCR cloning enzyme (Stratagene Corporation), 3µL of 10 × Easy‐A reaction buffer, 200µm of each deoxynucleoside triphosphate (dNTP), 1.0µm of each primer and 30ng of fungal template DNA. Amplification conditions consisted of 94°C for 2min, followed by 35 cycles at 94°C for 20s, 57°C (SdhB), 66°C (SdhC) or 58°C (SdhD) for 30s, and 72°C for 2min, with a final DNA extension at 72°C for 8min. PCR products were separated in ethidium bromide‐stained 1% agarose gels run in 1 × tris(hydroxymethyl)aminomethane (Tris)–borate–ethylenediaminetetraacetic acid (EDTA) buffer and exposed to UV light to visualize DNA fragments. PCR products were cloned into pGEM‐T easy vector (Promega Corporation, Madison, WI, USA) according to the manufacturer's instructions, and subsequently sequenced with M13 sequence primers using a dideoxy chain termination reaction method. Sequences were assembled and further aligned with Vector NTI and ClustalW software.

Phenotyping and genotyping of carboxin‐resistant UV mutants and field strains

Wells of flat‐bottomed microtitre plates (TPP92696, Techno Plastic Products, Trasadingen, Switzerland) were filled with 100µL of double‐strength Sabouraud dextrose liquid medium (SDLM; Oxoid, Basingstoke, Hampshire, UK) without fungicide or amended with different fungicide concentrations. The following concentrations were used to test field strains: bixafen, boscalid and isopyrazam (0.00023, 0.00092, 0.0037, 0.0146, 0.0586, 0.234, 0.938, 1.875, 3.75, 7.5 and 15µg/mL), carboxin (0.0977, 0.195, 0.391, 0.781, 1.56, 3.13, 6.25, 12.5, 25, 50 and 100µg/mL), chlorothalonil (0.0195, 0.0391, 0.0781, 0.156, 0.313, 0.625, 1.25, 2.5, 5, 10 and 20µg/mL), epoxiconazole (0.00014, 0.00051, 0.00192, 0.0072, 0.0270, 0.101, 0.379, 1.42, 5.33, 20 and 75µg/mL) and prothioconazole‐desthio (Parker etal., 2011), a metabolite of prothioconazole (0.00034, 0.0010, 0.0030, 0.0091, 0.0274, 0.0823, 0.247, 0.741, 2.22, 6.67 and 20µg/mL). For the mutants, higher concentrations of SDHIs were used: bixafen, boscalid and isopyrazam (0.00076, 0.0031, 0.0122, 0.0488, 0.195, 0.781, 3.13, 12.5 and 25µg/mL) and carboxin (0.488, 0.977, 1.95, 3.91, 7.81, 15.6, 31.3, 62.5, 125, 250 and 500µg/mL). Before dilution in media, chemicals were dissolved in acetone, with the exception of chlorothalonil, which was dissolved in dimethylsulphoxide (DMSO). One hundred microlitres of spore suspension (2.5 × 10⁴ spores/mL) of M.graminicola strains or UV mutants were added to each well. Plates were incubated for 4 days at 23°C, and growth was measured at 630nm using a FLUOstar OPTIMA microplate reader (BMG Labtech, Offenburg, Germany) in well‐scanning mode with a 2 × 2 matrix of scanning points within a 3‐mm diameter. Fungicide sensitivities were determined as the 50% effective concentration to inhibit growth (EC₅₀ in µg/mL) using a dose–response relationship according to OPTIMA software. Based on the phenotyping results, the SdhB, SdhC and SdhD genes of mutants with differences in SDHI sensitivity were sequenced. Sequence analysis identified putative resistance‐conferring mutations. Subsequently, pyrosequencing assays were developed to screen the remaining mutants for the presence of these mutations. The SdhB, SdhC and SdhD genes of mutants testing negative by pyrosequencing for the identified mutations were also sequenced. This was repeated until all mutants were genotyped.

Pyrosequencing assays

Primers (forward primer, reverse primer and sequence primer) were designed with Pyrosequencing Assay Design Software (Version 1.0.6; Biotage, Uppsala, Sweden) targeting different regions of SdhB, SdhC and SdhD in which SDHI resistance‐conferring mutations were identified (Table7). Polymerase chain reactions (PCRs) were carried out in a Biometra T3 thermocycler (Biotron, Göttingen, Germany) with 0.2 units of Red Hot DNA polymerase (ABgene, Epsom, Surrey, UK) using 20mm (NH₄)SO₄, 75mm Tris‐HCl (pH 9.0), 0.01% Tween‐20, 1.5mm MgCl₂, 150µm of each dNTP, 0.5µm of each primer and 20ng template DNA in a final volume of 20µL. Cycling conditions were as follows: 94°C for 2min, followed by 40 cycles at 94°C for 10s, 61°C (SdhB codon 221) or 65°C (other targets) for 20s and 72°C for 30s, with a final DNA extension at 72°C for 5min. Single‐stranded biotinylated PCR products were prepared using the Pyrosequencing Vacuum Prep Tool (Biotage). Three microlitres of Streptavidin Sepharose HP beads (GE Healthcare, Uppsala. Sweden) were added to 40µL of binding buffer (10mm Tris‐HCl, pH 7.6, 2 m NaCl, 1mm EDTA, 0.1% Tween‐20) and mixed with 15µL of PCR product and 25µL of sterile distilled water for 10min at room temperature using an Orbis plate shaker (Mikura, Lower Beeding, West Sussex, UK). Beads containing the captured DNA product were aspirated onto filters after applying a vacuum, washed with 70% (v/v) ethanol for 5s, rendered single stranded with denaturation solution (0.2 m NaOH) for 10s and neutralized with washing buffer (10mm Tris‐acetate, pH 7.6) for 5s. The vacuum was released and beads with biotinylated single‐stranded DNA were transferred into a PSQ 96‐well plate (Biotage) containing 45µL of annealing buffer (20mm Tris‐acetate, 2mm magnesium acetate, pH 7.6) and 0.5mm sequence primer. Pyrosequencing reactions were performed and analysed on a PyroMark Q96ID according to the manufacturer's instructions using a PyroMark Gold Q96 QSA Reagent Kit (Biotage), and the nucleotide dispensation orders are shown in Table7.

Insilico studies on the binding of SDHIs to the Sdh complex of M.graminicola

For docking studies of fungicides in the M.graminicola Sdh complex, the crystallographic coordinates of the avian Sdh complex were taken from the Protein Data Bank (PDB). The Sdh complex of chicken (PDB code 2FBW) is similar to that of M.graminicola (70% sequence identity) and also contains a bound carboxin molecule. By replacing all residues within a 14‐Å radius of the bound carboxin with the corresponding residues of the M.graminicola Sdh complex using PyMOL, a structural model of the fungal quinone binding site was prepared. All fungicides tested were also prepared in PyMOL. The Molegro Virtual Docker (MVD; Molegro ApS, Aarhus, Denmark) was used for docking simulations. For docking simulations, the structural model and all fungicides were first subjected to energy minimization in MVD. For each resistance‐related mutation in one of the subunits, the corresponding mutant protein was built and optimized. Fungicides were docked within a sphere (radius, 12Å) that encompassed the quinone binding site. For docking, the default settings were used, except for the grid resolution which was set to 0.2, and a total of 25 runs was performed with a maximum of 500 iterations. The best superimposed positions with lowest MolDock energy scores were used in this study. For the visualization of the docked fungicides, PyMOL was used.

Part of this work was supported by the Department of Environment, Food and Rural Affairs (DEFRA) through the Sustainable Arable LINK programme, project LK09133, in collaboration with Home‐Grown Cereals Authority (HGCA), BASF, Bayer CropScience, DuPont, Syngenta, Velcourt, Scottish Agricultural College (SAC) and Agricultural Development Advisory Service (ADAS). Rothamsted Research receives grant‐aided support from the Biotechnology and Biological Sciences Research Council (BBSRC) of the UK. Thanks are also due to Gert Kema, Andreas Mehl, Mike Shaw, Gerd Stammler and Suvi Viljanen‐Rollinson for supplying M.graminicola strains.