Modeling the Q o site of crop pathogens in Saccharomyces cerevisiae cytochrome b Nicholas Fisher 1 , Amanda C. Brown 1 , Graham Sexton 2 , Alison Cook 2 , John Windass 2 and Brigitte Meunier 1 1 The Wolfson Institute for Biomedical Research, UCL, London, UK; 2 Syngenta, Jealott’s Hill International Research Centre, Bracknell, Berkshire, UK Saccharomyces cerevisiae has been used as a model system to characterize the effect of cytochrome b mutations found in fungal and oomycete plant pathogens resistant to Q o inhibitors (QoIs), including the strobilurins, now widely employed in agriculture to control such diseases. Specific residues in the Q o site of yeast cytochrome b were modified to obtain four new forms mimicking the Q o binding site of Erysiphe graminis, Venturia inaequalis, Sphaerotheca fuligi- nea and Phytophthora megasperma. These modified versions of cytochrome b were then used to study the impact of the introduction of the G143A mutation on bc 1 complex activ- ity. In addition, the effects of two other mutations F129L and L275F, which also confer levels of QoI insensitivity, were also studied. The G143A mutation caused a high level of resistance to QoI compounds such as myxothiazol, axoxystrobin and pyraclostrobin, but not to stigmatellin. The pattern of resistance conferred by F129L and L275F was different. Interestingly G143A had a slightly deleterious effect on the bc 1 function in V. inaequalis, S. fuliginea and P. megasperma Q o site mimics but not in that for E. graminis. Thus small variations in the Q o site seem to affect the impact of the G143A mutation on bc 1 activity. Based on this observation in the yeast model, it might be anticipated that the G143A mutation might affect the fitness of pathogens differentially. If so, this could contribute to observed differences in the rates of evolution of QoI resist- ance in fungal and oomycete pathogens. Keywords:Q o inhibitors; bc 1 complex; cytochrome b; resistance; plant pathogens. The mitochondrial bc 1 complex is a membrane-bound enzyme that catalyzes the transfer of electrons from ubiquinol to cytochrome c and couples this electron transfer to the vectorial translocation of protons across the inner mitochondrial membrane. In eukaryotes it is comprised of 10 or 11 different polypeptides, and addi- tionally operates as a structural and functional dimer. Cytochrome b, cytochrome c 1 and the Rieske iron–sulfur protein (ISP) form the catalytic core of the enzyme. The catalytic mechanism, called the Q-cycle, requires two distinct quinone-binding sites (Q o , quinol oxidation site, and Q i , quinone reduction site), which are located on opposite sides of the membrane and linked by a trans- membrane electron-transfer pathway. The mitochondrially encoded cytochrome b subunit provides both the quinol and quinone binding pockets and the transmembrane electron pathway (via hemes b l and b h ). A number of quinol antagonists are known that inhibit bc 1 activity. These are either specific for the Q i site, such as antimycin, or for the Q o site, such as myxothiazol, stigmatellin, natural and synthetic strobilurins. Some of the latter Q o inhibitor compounds (QoIs) are now widely used in agriculture to control fungal and oomycete plant pathogens. Resistance to these inhibitors has, however, emerged in field populations of some such plant patho- gens. Two target site mutations in cytochrome b in particular appear to play a central role in the mechanism of resistance: G143A which has been reported in resistant isolates from various important pathogens ([1] and references within) and F129L which has been found in pathogens of turf grass, vines and potatoes. A143 is also found in the strobilurin-producing basidiomycete Mycena galopoda [2]. In E. graminis, the mutation G143A has spread widely and is without any apparent fitness penalty. In other pathogens, such as V. inaequalis, G143A has thus far been detected only in a localized geographical area. Still other pathogens have, however, not yet shown QoI resistance despite their exposure to Q o I fungicides ([1]). Several mechanisms might explain differences in the emergence of such resistance. One factor may be subtle variations in the structure and function of the Q o binding domain of the pathogens. In this work, the resistance mutations, in particular G143A were investigated in the context of yeast bc 1 structures. Yeast was used as a model system to construct several forms of the Q o domain, mimicking distinctive plant pathogen derived forms of this region based on both primary and tertiary structure comparisons, and to study the effect of the introduction of the QoI resistance mutation G143A on enzyme activity. Some of these distinctive changes in the Q o domain have been found to affect the impact of the resistance mutation on enzyme activity. Correspondence to B. Meunier, the Wolfson Institute for Biomedical Research, UCL, Gower Street, London, WC1E 6BT, UK. E-mail: b.meunier@ucl.ac.uk Abbreviations: ISP, iron–sulfur protein; PMSF, phenylmethylsulfonyl fluoride; QoI, Q o inhibitor. (Received 27 February 2004, revised 5 April 2004, accepted 16 April 2004) Eur. J. Biochem. 271, 2264–2271 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04169.x Experimental procedures Media and chemicals The following media were used for the growth of yeast: YPD [1% (w/v) yeast extract, 2% (w/v) peptone, 3% (w/v) glucose], YPG [1% (w/v) yeast extract, 2% (w/v) peptone, 3% (w/v) glycerol], transformation medium [0.7% (w/v) yeast nitrogen base, 3% (w/v) glucose, 2% (w/v) agar, 1 M sorbitol, and 0.8 gÆL )1 of a complete supplement mixture minus uracil; Anachem]. Decyl ubiquinone and myxo- thiazol were purchased from Sigma. Stigmatellin was purchased from Fluka. Generation of the yeast mutant strains Plasmid pBM5, carrying the wild-type intron-free version of the CYTB gene, was constructed by blunt end cloning of a PCR product of CYTB into the pCRscript vector (Strata- gene). Site directed mutageneses were performed using the Quickchange Site-Directed Mutagenesis Kit (Stratagene) according to the manufacturer’s recommendations. After verification of the sequence, plasmids carrying the intended mutant genes were used for microprojectile bombardment mediated mitochondrial transformation of yeast as des- cribed in [3]. Preparation of decylubiquinol Ten milligrams of 2,3-dimethoxy-5-methyl n-decyl-1,4-ben- zoquinone (decylubiquinone, Sigma), an analogue of ubi- quinone was dissolved in 0.4 mL nitrogen-saturated hexane. An equal volume of aqueous 1.15 M sodium dithionite was added, and the mixture shaken vigorously until colorless. The upper, organic phase was collected, and the decyl- ubiquinol recovered by evaporating off the hexane under nitrogen. The decylubiquinol was dissolved in 100 lL 96% (v/v) EtOH (acidified with 10 m M HCl)andstoredin aliquots at )80 °C. The concentration of decylubiquinol was determined spectrophotometrically from absolute spectra, using e 288)320 ¼ 4.14 m M )1 Æcm )1 . Preparation of crude mitochondrial membranes and measurement of cytochrome c reductase activity Wild-type and mutant yeast strains were grown to stationary phase (48 h) in 200 mL YPD cultures at 28 °C. The cells (approximately 2 g wet weight per culture) were then harvested by centrifugation at 4000 g for 10 min. Cell pellets were then washed by resuspension in 40 mL 50 m M potas- sium phosphate, 2 m M EDTA (pH 7.5) and centrifuged as before. The harvested cells were resuspended in 10 mL 50 m M potassium phosphate, 2 m M EDTA (pH 7.5) sup- plemented with 0.2 m M phenylmethylsulfonyl fluoride (PMSF) and 0.05% (w/v) bovine serum albumin prior to disruption in a Retsch MM300 glass bead mill operating at 30 Hz for 10 min at 4 °C. Membranes were separated from cell debris by centrifugation at 10 000 g for 20 min. The supernatant was centrifuged at 100 000 g for 90 min and the pelleted membranes resuspended in 1 mL of 50 m M potas- sium phosphate (pH 7.5), 2 m M EDTA containing 10% (v/v) glycerol. Resuspended membranes were stored in 0.1 mL aliquots at )80 °C. Cytochrome c reductase activity measurements were made in 50 m M potassium phosphate, pH 7.5, 2 m M EDTA, 10 m M KCN, 0.025% (w/v) lauryl maltoside and 30 l M equine cytochrome c at room tem- perature. Membranes were dilutedto 2.5 n M cytochrome bc 1 complex (determined from the reduced minus oxidized difference spectra, using e ¼ 28.5 m M )1 Æcm )1 at 562–575 nm [4]. Cytochrome c reductase activity was initiated by the addition of decylubiquinol (5–100 l M ). Reduction of cyto- chrome c was monitored in a Cary 4000 spectrophotometer at 550 vs. 542 nm over a 4 min time-course. Initial rates (computer-fitted as zero-order kinetics) were measured as a function of decylubiquinol concentration, and V m and K m values derived from Eadie–Hofstee (v vs. v/[S]) plots [5]. All rate measurements were performed in triplicate. Spectroscopic analysis of cytochromes in whole cells Spectra were generated by scanning cell suspensions with a single beam spectrophotometer built in-house and operating at room temperature. The cells, grown on YPD plates for 48 h, were resuspended at a concentration of 200 mg cells per milliliter and reduced by dithionite. The cytochrome concentration was estimated from the reduced spectra as described in [3]. Results and discussion Construction of yeast mutants with modified cytochrome b Q o sites The sequence of cytochrome b is highly conserved between species, especially in catalytic domains such as the Q o region. This site is actually a relatively large domain formed from components encompassing amino acid residues 120–150 and 260–280 of cytochrome b. The cavity consists of two lobes, a heme b l ÔproximalÕ lobe and a ÔdistalÕ lobe. The distal lobe is close to the surface region of cytochrome b and is involved in interactions with the peripheral domain of the iron–sulphur protein. The stigmatellin head-group binds in this distal lobe of the Q o site and is positioned in a pocket formed by amino acid tracts 122–131 (transmembrane helix C), 142–152 (helix cd1 and the cd1-cd2 linker), 268–280 (helix ef). The methoxyacrylamide moiety of myxothiazol, and methoxyacrylate moiety of strobilurin-related inhibi- tors, occupy the proximal domain, and are closely associ- ated (< 5 A ˚ separation) with the sidechains of residues F129 (transmembrane helix C), Y132 (ibid), G143 (helix cd1) and F275 (helix ef) [1,6]. Comparison of cytochrome b sequences around residues 129 and 143, involved in QoI resistance, showed some variations between pathogen species (Fig. 1). Firstly, S. cerevisiae, used as a model system in this work, has a unique feature: the CCV(133–135) sequence which, although also found in related yeast (Fig. 1A), is replaced by the sequence VLP(133–135) in most other organisms, including all plant pathogens we have analyzed and more distantly related species including mammals. To address the question whether the ÔCCVÕ sequence is essential to yeast bc 1 complex function or assembly, this sequence was replaced bythemorecommonÔVLPÕ sequence and the respiratory growth competence, the cytochrome b level and bc 1 activity Ó FEBS 2004 Modeling the Q o site of crop pathogens (Eur. J. Biochem. 271) 2265 were monitored (Tables 1 and 2). No effect was observed, suggesting that the yeast enzyme can accommodate the VLP sequence without loss of function. This new form of Q o domain, with the common VLP(133–135) sequence, has therefore been used throughout the other studies reported here. The effect of other variations in the Q o binding domain on bc 1 function and inhibitor resistance was then investi- gated. Four plant pathogens were chosen for this study, E. graminis (Ascomycete, pathogen of wheat), V. inaequalis (Ascomycete, pathogen of apple), S. fuliginea (Ascomycete, pathogen of cucumber) and P. megasperma (Oomycete, causing root rot disease) based on comparison of their primary sequences. The cytochrome b sequences of these plant pathogens, either obtained from public databases or by targeted PCR amplification and sequencing of field isolates, showed only small but distinctive changes in the Q o site (Fig. 1). Three permutations at position 136: tyrosine, phenylalanine and tryptophan, and three permutations at position 141: histidine, leucine and phenylalanine were observed in the four pathogens. In addition, a change of residue 275 from leucine to phenylalanine is seen P. mega- sperma cytochrome b. This latter change has been also reported in Pneumocystis carinii resistant to atovaquone treatment [7] and is naturally present in the corresponding mammalian enzyme [8]. Appropriate changes in the yeast cytochrome b sequence were introduced in order to obtain four new forms of cytochrome b: E. graminis-like (AB1), Fig. 1. Comparison of cytochrome b sequences in a region comprising the Q o domain. (A) Aligned sequences from yeasts and, as a representative mammal, humans (residues 121–155, S. cerevisiae numbering). (B) Corresponding sequence comparison of S. cerevisiae with the four plant pathogens employed in this study. (C) The sequence of the 15 yeast variants constructed and analyzed in this work. The mutated residues are in bold. The sequences of E. graminis, V. inaequalis and P. megasperma are available from the EMBL database. The sequence of S. fuliginea was determined by targeted RT-PCR amplification as described in [11]. 2266 N. Fisher et al. (Eur. J. Biochem. 271) Ó FEBS 2004 V. inaequalis-like (AB4), S. fuliginea-like (AB7) and P. megasperma-like (AB9) mutants (Fig. 1C). These new forms of cytochrome b were also used to compare the impact of the introduction of the mutations G143A, F129L and L275F on bc 1 complex activity. To this end, we introduced these additional mutations into the Table 1. Respiratory growth competence, cytochrome b content and resistance to Q o inhibitors. To determine the doubling time, cells were inoculated in respiratory medium (YPG) and the optical density was monitored periodically at 600 nm. The cytochrome b (cyt b)contentwasdeterminedin whole cells by spectrophotometry as described in experimental procedures, using e ¼ 25 m M -1 .cm -1 at 562–575 nm. The cyt b concentration in the wild type cells was 5.7 nmol per gram of cells. The respiratory growth in presence of inhibitor was monitored on respiratory media (YPG) plus 1 or 10 l M inhibitor as described in Fig. 3. +++ indicates vigorous growth; ++ and +, weaker growth; – , no growth. Strains Mutations Doubling Time (hrs) Cyt b content (%) Growth on Myxothiazol Stigmatellin Azoxystrobin Pyraclostrobin 10 10 1 10 1 10 WT 4 100 – + – – – – Erysiphe graminis-like AB1 VLP 4 95 – – – – – – AB2 G143A 5 100 +++ – +++ +++ +++ +++ AB3 F129L 5 100 +++ + + - - - AB17 F129L, G143A 5 90 +++ + +++ +++ +++ +++ Venturia inaequalis-like AB4 H141L 4 100 – – – – – – AB13 H141L, G143A 5 100 +++ – +++ ++ +++ +++ AB5 F129L, H141L 6 100 + – – – – – AB18 F129L, H141L, G143A 10 100 ++ + ++ ++ ++ ++ Sphaerotheca fuliginea -like AB7 Y136F, H141L 4 100 – – – – – – AB8 Y136F, H141L, G143A 5 95 ++ – +++ ++ +++ +++ Phytophthora megasperma-like AB9 Y136W, H141F 4.5 85 – – – – – – AB10 Y136W, H141F, G143A 5 75 +++ – +++ +++ +++ +++ AB16 F129L, Y136W, H141F 5 90 +++ ++ ++ – – – AB11 Y136W, H141F, L275F 5 60 – + – – – – AB12 F129L, Y136W, H141F, L275F 5 90 +++ ++ – – ++ + Table 2. QH 2 cytochrome c reductase activities. QH 2 cytochrome c reductase activity was assayed as described in experimental procedures. Strains Mutations bc 1 Complex activity Rates (s )1 ) at 50 l M QH 2 V m (s )1 ) K m (QH 2 l M ) WT 40 +/– 1.8 (100%) 80 18 Erysiphe graminis-like AB1 VLP 40 +/– 1.1 (100%) 82 17 AB2 G143A 35 +/– 1.8 (87%) 74 12 AB3 F129L 35 +/– 2.5 (87%) – – AB17 F129L, G143A 32 +/– 2.2 (80%) – – Venturia inaequalis-like AB4 H141L 28 +/– 1.1 (100%) 42 12 AB13 H141L, G143A 14 +/– 1.0 (50%) 25 6 AB5 F129L, H141L 26 +/– 0.5 (93%) – – AB18 F129L, H141L, G143A 12 +/– 1.7 (43%) – – Sphaerotheca fuliginea-like AB7 Y136F, H141L 39 +/– 2.1 (100%) 68 17 AB8 Y136F, H141L, G143A 26 +/– 1.2 (67%) 36 10 Phytophthora megasperma-like AB9 Y136W, H141F 27 +/– 0.6 (100%) 38 12 AB10 Y136W, H141F, G143A 11 +/– 1.0 (41%) 23 8 AB16 F129L, Y136W, H141F 12 +/– 1.1 (44%) – – AB11 Y136W, H141F, L275F 12 +/– 1.8 (44%) 27 12 AB12 F129L, Y136W, H141F, L275F 18 +/– 1.8 (67%) – – Ó FEBS 2004 Modeling the Q o site of crop pathogens (Eur. J. Biochem. 271) 2267 pathogen-like mutants. In total, 15 variants were constructed (Figs 1C and 2). These were generated by a biolistic trans- formation procedure, which produces homoplasmic yeast strains carrying only the variant cytochrome b sequence [3], andthenusedtomonitorrespiratoryfunctioninvariousways. Effects of mutations on respiratory growth and cytochrome b content All the variant cytochrome b yeast strains constructed were respiration competent. Their doubling times in nonferment- able medium (YPG) were 4–5 h, with the exception of strains AB5 and AB13 which showed doubling times of 6 and 10 h, respectively. This phenotype was not investi- gated further. In order to assess the effect of mutations on the assembly of the bc 1 complex, we also monitored the concentration of cytochromes in whole cells, as described in experimental procedures: changes introduced in the Q o domain had little effect on cytochrome b assembly. Cyto- chrome b content was between 90 and 100% of that of the wild-type, in the E. graminis-, V. inaequalis-andS. fuligi- nea-like constructs; though the changes introduced in the P. megasperma-like constructs seemed to hinder enzyme assembly slightly as judged by the decrease in cytochrome b content (Table 1). Lowest cytochrome b levels were observed in the strain harboring the three mutations Y136W, H141F and L275F (60% of the wild type). Interestingly, these three changes are naturally present in mammals. The introduction of a fourth mutation, F129L restored the cytochrome b content to near wild-type level (Table 1). It seems likely that the introduction of three bulky residues, Y136W, H141F and L275F, sterically hinders the folding of cytochrome b and the assembly of the complex. The replacement of phenylalanine at position 129 by a smaller residue leucine may then alleviate the hindrance and restore the proper folding of cytochrome b. Resistance to Q o inhibitors As mutations G143A and F129L had been found in plant pathogen isolates resistant to QoIs, we monitored the respiratory growth competence of the different constructs in the presence of stigmatellin, which binds in the distal lobe of the Q o site, and myxothiazol, azoxystrobin and pyra- clostrobin, which bind at the proximal lobe of the Q o site (Fig. 3 and Table 1). The control strains, AB1, AB4, AB7 and AB9 were all sensitive to myxothiazol, stigmatellin, azoxystrobin and pyraclostrobin. Introduction of G143A in all four Q o forms led to strong resistance to myxothiazol, azoxystrobin and pyraclostrobin: strains AB2, AB13, AB8 and AB10 grew on nonfermentable medium in presence of 10 l M of each of these compounds but were still sensitive to stigmatellin. Interestingly structural studies suggest that the Ca hydrogen Fig. 2. Structure of the Q o site. The cyto- chrome b a-carbon backbone is shown in orange. The location of residues altered to model the Q o -sites from the pathogenic fungi discussed in the text are shown in green. The VLP(133-135) region of cytochrome b is indi- catedinwhite.Q o -bound stigmatellin and hemes b l /b h are represented in cyan and red, respectively. This figure was prepared from the yeast bc 1 crystal structure coordinates 1KYO.pdb [12] using VISUAL MOLECULAR DYNAMICS software [13]. 2268 N. Fisher et al. (Eur. J. Biochem. 271) Ó FEBS 2004 atom of G143 approaches within 3.5 A ˚ of the methoxy- acrylamide moiety of myxothiazol and hence mutation to the bulkier residue alanine is likely to abolish the binding of this class of Q o antagonist [1,6]. A similarly close interaction with the benzene ring ÔlinkerÕ region of azoxystrobin and pyraclostrobin could explain resistance to these compounds. The pattern of resistance induced by F129L was different. Strains AB3 and AB16 were resistant to myxothiazol and stigmatellin. They also show limited cross-resistance to azoxystrobin as growth was observed at 1 l M azoxystrobin but not at 10 l M . Yeast cells carrying this mutation were rather more sensitive to pyraclostrobin: no growth was observed at 1 l M pyraclostrobin. The sidechain of F129 approaches within 3 A ˚ of the myxothiazol methoxyacryl- amide moiety. By contrast, F129 has a closest approach of 4A ˚ with the hydrophobic tail of stigmatellin. The likely mechanism of F129L stigmatellin resistance is therefore not clear, but it could be due to a subtle alteration of the backbone fold at Q o , or a change in accessibility for the antagonist to the Q o site. The slight variance in sensitivity to azoxystrobin and pyraclostrobin is likely to be due to the difference in pharmacophore structure between these two compounds, as discussed in more detail below. As men- tioned above, strain AB5 showed a weaker growth that could explain the apparent sensitivity. Interestingly AB12, which combined F129L with L275F, was sensitive to azoxystrobin but resistant to pyraclostro- bin. In this case it is likely that the two changes have slightly modified the structure of the Q o site, which can now accommodate azoxystrobin but not pyraclostrobin. The sidechain of F275 in chicken bc 1 complex is involved in a stabilizing ring–stacking hydrophobic interaction with the phenyl group of MOA-stilbene [6], a Q o inhibitor closely related to strobilurin. This may explain why strain AB11 (Y136W, H141F, L275F) retains sensitivity to the strobilu- rin-related inhibitor azoxystrobin. Strain AB12 (Y136W, H141F, L275F + F129L) demonstrated resistance to both myxothiazol and pyraclostrobin, but remained sensitive to azoxystrobin. Pyraclostrobin and Azoxystrobin differ in pharmacophore structure; the former contains an alkoxy- amino moiety, whereas the latter is methoxyacrylate based (Fig. 4). Significantly, the pharmacophore of pyraclostrobin occupies a smaller volume than that of azoxystrobin, and might have a greater degree of rotational freedom due to the Fig. 4. Structure of Q o inhibitors azoxystrobin and pyraclostrobin [1]. Pharmacophore groups are indicated by boxes. Fig. 3. Sensitivity to Q o inhibitor exposure. The name and position of the strains are shown in the right-hand panel. A drop of each strain was inoculated on a nonfermentable medium plate (YPG) with or without 10 l M inhibitor and incubated for 3–4 days. Ó FEBS 2004 Modeling the Q o site of crop pathogens (Eur. J. Biochem. 271) 2269 lack of methoxyacrylate p-bonded structure. Mutation of both F129 and L275 to leucine and phenylalanine, respect- ively, are required to inhibit pyraclostrobin binding. As expected, strains AB17 and AB18 harboring both G143A and F129L combined resistance to myxothiazol, azoxystrobin and pyraclostrobin with resistance to stig- matellin. In order to quantify the level of resistance induced by G143A, bc 1 complex sensitivity to myxothiazol and stig- matellin was monitored in membranes from strain AB2 and its control AB1. QH 2 cytochrome c reductase activity (using 2.5 n M bc 1 complex), as in Table 2, was measured in presence of increasing concentration of inhibitors. The concentration of stigmatellin required for 50% decrease of activity (I 50 ) was around 2.5 n M for AB1 and AB2, whereas the I 50 for myxothiazol was 2.5 n M for AB1 and 18 l M for AB2: a 7500-fold increase. This is in good agreement with previous results. The G143A mutation was first reported in mammalian cells after selection in presence of myxothiazol, conferring > 7000-fold resistance to the inhibitor [9]. Effect of mutations on bc 1 complex activity In order to study possible effects of the mutations on bc 1 function, mitochondrial membranes were prepared from the different strains and cytochrome c reductase activity was monitored spectrophotometrically as described in experi- mental procedures. As shown in Table 1, the replacement of the yeast sequence CCV(133-135) by the much more common sequence VLP, in the E. graminis-like strain had no effect on enzyme activity. In the V. inaequalis-like strain (AB4), histidine 141 was replaced by leucine. The activity of the resultant enzyme was then decreased by 30% compared to the wild-type yeast. Activity was however, restored to near wild-type levels by the introduction of a second change, Y136F, in the S. fuliginea-like strain (AB7). The P. mega- sperma-like enzyme (in strain AB9), which harbored Y136W and H141F also showed a 30% decrease in bc 1 activity. The introduction of G143A, F129L or both changes together (though this has not been seen in any natural isolate to our knowledge) in the E. graminis-like Q o site had little effect on bc 1 activity (80–87% of wild-type rate). Thus this Q o site can accommodate the G143A and F129L mutations without loss of function. This is consistent with previous observations with E. graminis itself, which showed that the isolates carrying the G143A mutation did not suffer any fitness penalty [10]. In the V. inaequalis-like strains, the situation was different. As mentioned above, the control strain (AB4) harboring the change H141L showed a lower activity than the wild-type yeast strain (turnover number 28 s )1 vs. 40 s )1 ). Interestingly the introduction of the G143A mutation in this Q o site further decreased the bc 1 activity to 14 s )1 (50% of the control AB4). In contrast, F129L had no effect. In AB18, which combined G143A and F129F, the enzyme activity was 43% of the control. It seems therefore that the V. inaequalis-likeenzymecannotaccom- modate the G143A mutation without reduction of function. Similar results were obtained with the P. megasperma-and the S. fuliginea-like Q o sites. The introduction of G143A caused, respectively, a 60% and 33% decrease of the bc 1 activity compared to the controls. We have also used the P. megasperma-like form to monitor the effect of F129L and L275F. The mutation L275F is naturally occurring in Phytophothora sp. The introduction of these mutations decreased the bc 1 activity to 44% of the control AB9. Their combination in AB12 restored the activity to 67% of the control AB9. Thus the introduction of L275F in the P. megasperma-like Q o caused a decrease both in bc 1 content and activity, while F129L partially compensated the defect. To gain further information on the effect of the mutation G143A, steady-state cytochrome c reductase activity was monitored as a function of decylubiquinol (QH 2 ) concen- tration. The apparent V m and K m for QH 2 were calculated from initial rate measurements using derived Eadie–Hofstee plots (Table 1). The mutation G143A appeared to decrease both the V m and the K m for quinol in AB13, AB8 and AB10. It might therefore be that this mutation slightly affects the structure of the Q o site which, as a result, becomes saturated with substrate more rapidly than the control due to lower electron transfer, or alternatively it may reflect a decreased ÔonÕ rate for quinol binding. The replacement of glycine by alanine is a relatively conservative structural change, and unlikely to disrupt the fold of the cd1 helix. The introduced methyl group may sterically hinder interactions with the quinol headgroup, or unfavorably alter the conformation of bound quinol such that electron transfer or deprotonation rates are decreased. Thus variations in the Q o domain seem to affect the impact of the QoI resistance mutation G143A on cyto- chrome bc 1 activity. In some cytochrome b forms, the introduction of G143A decreases the QH 2 cytochrome c activity of the complex. Under standard laboratory condi- tions in S. cerevisiae, this decrease has no effect on cell growth as little as 20% of bc 1 complex activity is enough to support respiratory growth. Therefore a decline in respir- atory growth will only be seen when the complex is severely inhibited. However in other organisms, such as plant pathogens, when the energetic demands are higher, this decrease might affect the fitness of the cells. In combination with other factors, this could explain the differences in the evolution of QoI resistance in fungal and oomycete pathogens. Interestingly the characteristic Q o site features of E. graminis, one of the pathogens which showed field resistance to Q o I fungicides particularly quickly, seem to be most functionally accommodating of the resistance-associ- ated G143A mutation. Acknowledgements This work was supported by Syngenta. The authors acknowledge the contributions made by our colleagues, Ms Carole Stanger and Ms. Judith Burbidge, to the analysis of cytochrome b gene and/or mRNA sequences from plant pathogen isolates. We would also particularly wish to recognize the interest, enthusiasm and insight in initiating these studies shown by our late colleague Steve Heaney, and this paper is dedicated to his memory. References 1. Gisi, U., Sierotzki, H., Cook, A. & McCaffery, A. (2002) Mechanisms influencing the evolution of resistance to Qo inhibitor fungicides. Pest Manag. Sci. 58, 859–867. 2270 N. Fisher et al. (Eur. J. Biochem. 271) Ó FEBS 2004 2. Kraiczy, P., Haase, U., Gencic, S., Flindt, S., Anke, T., Brandt, U. & von Jagow, G. (1996) The molecular basis for the natural resistance of the cytochrome bc 1 complex from strobilurin-pro- ducing basidiomycetes to center Q P inhibitors. Eur. J. Biochem. 235, 54–63. 3. Hill, P., Kessl, J.J., Meshnick, S.R., Trumpower, B.L. & Meunier, B. (2003) Recapitulation in Saccharomyces cerevisiae of cyto- chrome b mutations conferring resistance to atovaquone in Pneumocystis jiroveci. Antimicrob. Agents Chemother. 47, 2725– 2731. 4. Vanneste, W.H. (1966) Molecular proportion of the fixed cyto- chrome components of the respiratory chain of Keilin-Hartree particles and beef heart mitochondria. Biochim. Biophys. Acta 113, 175–178. 5. Cornish-Bowden, A. (1995) Fundamentals of Enzyme Kinetics. Portland Press, London. 6. Zhang,Z.,Huang,L.,Shulmeister,V.M.,Chi,Y I.,Kim,K.K., Hung, L W., Crofts, A.R., Berry, E.A. & Kim, S H. (1998) Electron transfer by domain movement in cytochrome bc 1 . Nature 392, 677–684. 7. Kazanjian, P., Armstrong, W., Hossler, P.A., Huang, L., Beard, C.B.,Carter,J.,Crane,L.,Duchin,J.,Burman,W.,Richardson,J. & Meshnick, S.R. (2001) Pneumocystis carinii cytochrome b mutations are associated with atovaquone exposure in patients with AIDS. J. Infect. Dis. 183, 819–822. 8. Kessl, J.J., Lange, B.B., Merbitz-Zahradnik, T., Zwicker, K., Hill, P.,Meunier,B.,Palsdottir,H.,Hunte,C.,Meshnick,S.& Trumpower, B.L. (2003) Molecular basis for atovaquone binding to cytochrome bc 1 complex. J. Biol. Chem. 278, 31312–31318. 9. Howell, N. & Gilbert, K. (1988) Mutational analysis of the mouse mitochondrial cytochrome b gene. J. Mol. Biol. 203, 607–618. 10. Chin, K.M., Chavaillaz, D., Kaesbohrer, M., Staub, T. & Felsenstein, F.G. (2001) Characterizing resistance risk of Erysiphe graminis f.sp. tritici to strobilurins. Crop Protection 20, 87–96. 11. Sierotzki, H., Parisi, S., Steinfeld, U., Tenzer, I., Poirey, S. & Gisi, U. (2000) Mode of resistance to respiration inhibitors at the cytochrome bc 1 enzyme comples of Mycosphaerella fijiensis field isolates. Pest Manag. Sci. 56, 833–841. 12. Hunte, C., Koepke, J., Lange, C., Rossmanith, T. & Michel, H. (2000) Structure at 2.3 angstrom resolution of the cytochrome bc 1 complex from the yeast Saccharomyces cerevisiae co-crystallized with an antibody Fv fragment. Structure Fold. Des. 8, 669–684. 13. Humphrey, W., Dalke, A. & Schulten, K. (1996) VMD – visual molecular dynamics. J. Mol. Graph. 14, 33–38. Ó FEBS 2004 Modeling the Q o site of crop pathogens (Eur. J. Biochem. 271) 2271 . azoxystrobin and pyra- clostrobin, which bind at the proximal lobe of the Q o site (Fig. 3 and Table 1). The control strains, AB1, AB4, AB7 and AB9 were. Specific residues in the Q o site of yeast cytochrome b were modified to obtain four new forms mimicking the Q o binding site of Erysiphe graminis, Venturia inaequalis,