Gene Deletion of 7,8-Linoleate Diol Synthase of the Rice Blast Fungus

Linoleate diol synthases (LDS) are heme enzymes, which oxygenate 18:2n-6 sequentially to (8R)-hydroperoxylinoleic acid ((8R)-HPODE) and to (5S,8R)-dihydroxy-, (7S,8S)-dihydroxy-, or (8R,11S)-dihydroxylinoleic acids (DiHODE). The genome of the rice blast fungus, Magnaporthe oryzae, contains two genes with homology to LDS. M. oryzae oxidized 18:2n-6 to (8R)-HPODE and to (7S,8S)-DiHODE, (6S,8R)-DiHODE, and (8R,11S)-HODE. Small amounts of 10-hydroxy-(8E,12Z)-octadecadienoic acid and traces of 5,8-DiHODE were also detected by liquid chromatography-mass spectrometry. The contribution of the 7,8-LDS gene to M. oryzae pathogenicity was evaluated by replacement of the catalytic domain with hygromycin and green fluorescent protein variant (SGFP) cassettes. This genetically modified strain Δ7,8-LDS infected rice leaves and roots and formed appressoria and conidia as the native fungus. The Δ7,8-LDS mutant had lost the capacity to biosynthesize all the metabolites except small amounts of 8-hydroxylinoleic acid. Studies with stereospecifically deuterated linoleic acids showed that (8R)-HPODE was formed by abstraction of the pro-S hydrogen at C-8 and antarafacial oxygenation, whereas (7S,8S)-DiHODE and (8R,11S)-DiHODE were formed from (8R)-HPODE by suprafacial hydrogen abstraction and oxygenation at C-7 and C-11, respectively. A mac1 suppressor mutant (Δmac1 sum1–99) of M. oryzae, which shows cAMP-independent protein kinase A activity, oxygenated 18:2n-6 to increased amounts of (10R)-HPODE and (5S,8R)-DiHODE. Expression of the 7,8-LDS gene but not of the second homologue was detected in the suppressor mutant. This suggests that PKA-mediated signaling pathway regulates the dioxygenase and hydroperoxide isomerase activities of M. oryzae.

Studies on the biological function of fungal oxylipins started 20 years ago in A. nidulans. (8R)-HODE and (5S,8R)-DiHODE were found to induce premature sexual sporulation (20). Gene deletion of 5,8-LDS and (10R)-DOX in A. nidulans and A. fumigatus was later found to influence sporulation, development, mycotoxin production, and pathogenicity (16,21,22). In contrast, little is known about the biological function of 7,8-LDS of G. graminis. The hyphae of G. graminis penetrate and destroy wheat roots, and the take-all disease causes significant economical losses (23). The genome of G. graminis is not yet available. Transformation and genetic manipulation of G. graminis are difficult and preclude gene deletion studies of 7,8-LDS. We therefore turned our attention to the genetically tractable M. oryzae, another fungal grass pathogen of the Magnaporthaceae family (24), to better understand the biological roles of the 7,8-LDS. Amino acid identities between deduced oxygenases of the two genes of M. oryzae with homology to 7,8-LDS of G. graminis and fatty acid oxygenases of aspergilli are summarized in Table 1. M. oryzae has previously been found to oxygenate 18:2n-6 to 7,8-DiHODE and 8-HODE, and the transcript of the tentative 7,8-LDS gene was sequenced (14). 6 M. oryzae is a model organism for studies of fungal diseases of blasts (25), and it also infects rice roots (26). The spores of M. oryzae attach to the hydrophobic surface of leaves and contain all that is needed to develop an injection apparatus for cuticle penetration, designated an appressorium with penetrating hyphae. In this process, the conidium undergoes apoptosis with transfer of lipid bodies, trehalose, and glycerol to the appressorium. The subsequent mobilization and oxidation of fatty acids of lipid bodies are controlled by PKA, MAPK, triacylglycerol lipases, and ␤-oxidation (27)(28)(29)(30)(31)(32). Glycerol accumulates and generates the unprecedented internal turgor pressure of the appressorium (27). These processes are deranged in mutants of the cAMP-protein kinase cascade (29). Whether LDS enzymes are involved in lipid mobilization and pathogenicity and can be regulated by kinases are unknown.
The first goal of this study was to characterize the oxygenation of 18:2n-6 by M. oryzae with respect to all products, oxygenation mechanisms, and stereochemistry. Our second goal was to use gene targeting to identify and study the biological importance of 7,8-LDS during rice infection and sporulation. The third goal was to determine whether LDS activity is influenced by the cAMP-mediated signaling pathway. Finally, we also assayed oxygenation of unsaturated fatty acids by mycelia to identify the major products formed by cytochrome P450.
The following genetically modified strains of M. oryzae were studied: deletion of adenylyl cyclase MAC1 gene (⌬mac1) (35); deletion of the adenylyl cyclase MAC1 gene with a suppressor mutation in the cAMP binding domain of the regulatory subunit of PKA (⌬mac1 sum1-99 (29)); deletion of the catalytic subunit of protein kinase A (⌬cpka) (36); deletion of PMK1 (PMK1, pathogenicity MAPK 1) (31); and deletion of the MAPK substrate and transcription factor MST12 (32). The mycelium was ground up in liquid nitrogen and homogenized, and the supernatant was assayed for enzyme activity with 18:2n-6 as substrate (see below). The experiment was performed in triplicate.
Enzyme Assays-Mycelia (0.5-20 g wet weight) of M. oryzae were incubated in 0.1 M sodium borate buffer (pH 8.0) with 0.5-1 mg of fatty acids ml Ϫ1 (4 -5 h, 21°C), as described previously (14). Mycelia were also ground up in liquid nitrogen, and the fine powder was stored at Ϫ80°C. The nitrogen powder was homogenized (glass-Teflon, 10 passes; 4°C) in 10 volumes (w/v) of 0.1 mM KHPO 4 buffer (pH 7.3), 2 mM EDTA, 0.04% Tween 20, centrifuged at 13,000 ϫ g (10 min, 4°C), and used immediately for enzyme assay. Microsomal fractions and high speed supernatants were prepared by centrifugation (100,000 ϫ g, 60 min; 4°C) of low speed supernatant and assayed for enzyme activity. An aliquot (0.5-1 ml) was incubated with 100 M of fatty acids for 30 -40 min on ice. In some experiments, 50 pmol of (13R)-[ 2 H 4 ]HODE was added as an internal standard. The products were extracted with ethyl acetate or on SepPak/C 18 (19). Triphenylphosphine or NaBH 4 was used to reduce hydroperoxides to alcohols to simplify analysis. The products from incubations with mycelia were purified by TLC (Kieselgel 60, Merck) or by preparative RP-HPLC.
A. tumefaciens-mediated Transformation-Competent A. tumefaciens containing pCAMB⌬7,8-LDS was used to transform M. oryzae Guy11 conidia (39). Selection medium agar consisted of complete medium supplemented with 250 g/ml hygromycin B and with 200 M cefotaxime to eliminate A. tumefaciens. Transformants were grown individually on selection medium in 24-well plates before analysis.
Analysis of ⌬7,8-LDS Transformants-Transformants were analyzed by PCR, Southern blot, and for expression of LDS activity. PCR was used as a first screening to distinguish between homologous and nonhomologous recombination. Primers F-MG-43646 (5Ј-GGTTGTTTGTAGTACTGCAG-CAGC) and R-SGFP-149 (5Ј-GCAGATGAACTTCAGGTG-CAGCTT) amplified a fragment of 1736 bp in transformants resulting from homologous recombination. This primer pair would generate amplicons of different size, or no amplicon at all, in ectopic transformants. Primers F-⌬lds (5Ј-AACG-GCAACGGTATACATCAGAAC-3) and R-⌬lds (5Ј-TATT-GCCAAATGTTTGAACGATCGG-3Ј) generated amplicons of 1105 bp in ectopic transformants, because the R-⌬lds primer binds to a sequence in the T-DNA of pCAMB⌬7,8-LDS outside of the gene replacement construct. Confirmation of the integration pattern was determined by Southern hybridization (see supplemental material). Finally, the deletion mutants of 7,8-LDS were assayed for enzyme activity, as above.
PCR Analysis of Gene Transcripts in M. oryza, the ⌬mac1 Mutant, and the Suppressor Mutant (⌬mac1 sum1-99)-Total RNA was prepared from nitrogen powder of mycelia (grown for 7-14 days) using LiCl extraction (17) and treated with DNase I. Quality was assessed on the Experion Automated Electrophoresis System (Bio-Rad). First-strand cDNA was synthesized using Superscript III. Real time PCR was performed in an iCycler (Bio-Rad) with SYBR Green Supermix. 500 nM forward and reverse primers and ϳ200 ng of cDNA were used in 50 l. Primers used for the detection of MGG_13239 were 5Ј-AGC-CTTCAACACGTTGATGAAG (forward) and 5Ј-GGAG-GAACGTCGAGTCCTTG (reverse) and for MGG_10859 were 5Ј-ACCGCGTCTTTGTATCCTTTG (forward) and 5Ј-CATCTCGGTGATGGCAATCTG (reverse). Both primers were designed to bridge an intron to distinguish amplification of cDNA from any contaminating genomic DNA. Primers for detection of actin as housekeeping gene (40) were 5Ј-CCTG-GCACCGTCGTCGATGAAGG and 5Ј-GCGAGGCGAG-AATGGAACCACCG. Cycling conditions were as follows: 95°C, 3 min followed by 50 cycles (95°C, 30 s; 58°C, 30 s; 72°C, 30 s). Melting point curve analysis and agarose gel electrophoresis verified amplification of one single product of the expected size. Cycle threshold (C T ) values were obtained from the iCycler software (Bio-Rad).
Infection Assays-Rice (Oryzae sativa spp. indica var. CO39) seedlings were grown at 85% relative humidity, 25°C, and 16-h light/8-h dark photoperiod. Two-week-old plants with second and third expanded leaves were used for leaf infection assays. Plants were spray-inoculated with 2 ml of a suspension of 10 5 conidia/ml with 0.25% gelatin per pot. Symptoms were scored after 5 days.
Root infection assays were carried out using thick and moist vermiculite. Wet vermiculite was prepared by immersing in distilled water (2 h), and excess water was then removed by a sieve. We filled a 50-ml Falcon tube with 30 cm of wet vermiculite, followed by a mycelial plug with the same diameter as the Falcon tube, a further layer of 5 cm of wet vermiculite, and five rice seeds of the same strain covered with the other 5-cm layer of wet vermiculite. The tube was sealed with parafilm to prevent loss of humidity. Lesions were scored and compared with wild type strain Guy11 after 15 days.
Appressorium Formation-Aliquots (150 l) of conidia suspensions (10 4 spores/ml) were placed on plastic microscope coverslips (BDH, Germany) and incubated under humid conditions at 25°C in 16-h light/8-h dark photoperiod. At least 200 conidia per strain were examined microscopically for germination and formation of appressoria in each experiment.
Radial Growth and Asexual Sporulation-Mycelial plugs of M. oryzae Guy11 and ⌬7,8-LDS were placed in triplicates on agar plates with complete and minimal medium. The fungal growth area, and the number of conidia were determined after 10 days (22°C, fluorescent light as above). The conidia were harvested as above, and the conidia were counted in a hemocytometer and corrected for the fungal growth area.
GC-MS analysis, hydrogenation, and synthesis of methyl ester and TMS ether derivatives were performed as described previously (41,42). Carbon values were estimated from the retention times of saturated fatty acid methyl esters (41). Steric analysis was performed after chemical reduction of HPODE to alcohols. 8-HODE consisted of over 95% of the 8R stereoisomer (supplemental material). We confirmed that (8R)-HPODE was a precursor of 7,8-and 6,8-DiHODE. The absolute configuration of (7,8S)-DiHODE was determined by GC-MS analysis after hydrogenation (Fig. 4A); the designation 8S in (7S,8S)-DiHODE is due to the hydroxyl at C-7 and the Cahn-Ingold-Prelog nomenclature rules. The retention time of the biological product was the same as the threo stereoisomer of 7,8-octadecanoic acid (methyl ester TMS ether derivative, as described in the supplemental material), and only a few percent of the erythro isomer was detected. The configuration at C-6 of 6, (8R)-DiHODE was also determined by GC-MS analysis after hydrogenation. The hydrogenated product had the same retention time as the anti stereoisomer of 6,8-DiHODE (TMS ether methyl ester derivative, supplemental material), as shown in Fig. 4B. This suggested that (6S,8R)-DiHODE was formed. We conclude that both dihydroxymetabolites were formed with high stereo selectivity.  Table 2. The deuterium labels were retained in both (8R)-HODE and in (7S,8S)-DiHODE, respectively. This suggested antarafacial hydrogen abstraction and oxygen insertion at C-8 and suprafacial hydrogen abstraction and oxygenation at C-7. (11S)-[ 2 H]18:2n-6 (Ͼ95%) was transformed to (8R,11S)-DiHODE with loss of the deuterium label, suggesting suprafacial oxygenation at C-11 (Table 2).
Oxylipins Formed by Mutants Implicated in the cAMP-dependent and Mitogen-activated Protein Kinase Cascades-There are two important signaling pathways required for M. oryzae appressorium development and leaf penetration as follows: the PMK1 pathway (31), and the cAMP-activated PKA pathway (36). Nitrogen powder preparations of M. oryzae mutants lacking the adenylyl cyclase (⌬mac1), a catalytic sub-   The deuterium content was determined in recovered 18:2n-6 from the experiments to correct for dilution of ͓ 2 H͔18:2n-6 by endogenous 18:2n-6 in the preparations (cf. Fig. 3A).
Real Time PCR Analysis-Real time PCR was performed to investigate whether the increased formation of (5S,8R)-DiHODE and (10R)-HPODE in ⌬mac1 sum1-99 strain could be due to relative up-regulation of the second gene with homology to 7,8-LDS, MGG_10859. We therefore compared the expression of the transcripts of MGG_10859, MGG_13239, and actin in three strains (Guy11 and two mutants, ⌬mac1 and  ⌬mac1 sum1-99).
In Guy11 and ⌬mac1 mycelia, transcripts of MGG_13239 (7,8-LDS) were more (⌬C T ϳ 1.9) and less abundant (⌬C T ϳ 1.9) compared with transcripts of actin. This suggested that 7,8-LDS could be down-regulated in ⌬mac1 (⌬C T ϳ 3.8). The transcripts of MGG_10859 were less abundant than actin in Guy11 (⌬C T ϳ 10.6) and ⌬mac1 (⌬C T ϳ 5.7) and thus up-regulated in the latter (⌬C T ϳ 4.9). We confirmed that the nitrogen powder preparations used for RNA extraction oxidized linoleic  In the ⌬mac1 sum1-99 mutant, transcripts of MGG_13239 (7, were detected by real time PCR analysis, but we could not detect transcripts of MGG_10859 (n ϭ 7; 50 cycles). We confirmed that the PCR efficiencies for both primer pairs of MGG_10859 and MGG_13239 were comparable with genomic DNA as template. We also confirmed that the nitrogen powder preparation of the ⌬mac1 sum1-99 mutant used for RNA extraction formed 5,8-DiHODE and 10-HODE as major metabolites.
Hydroxylation and Epoxidation of Unsaturated Fatty Acids-Mycelia of M. oryzae catalyzed 2and/or 3-hydroxylation of all investigated unsaturated fatty acids and oxidation to vicinal diols also occurred. -Hydroxymetabolites could not be detected. The diols were likely formed by epoxidation of the terminal double bonds of the fatty acids followed by enzymatic hydrolysis. Over 50% of the fatty acids were typically consumed in 4 -5 h, and these hydroxy-and dihydroxymetabolites were obtained in significant yields, as judged from TLC analysis. Experimental details of GC-MS analysis and a summary of the major metabolites of unsaturated C 18 -C 22 fatty acids are given in the supplemental material.

DISCUSSION
We have studied oxygenation of 18:2n-6 by M. oryzae Guy11 and report four major findings. First, gene deletion allowed us to identify the 7,8-LDS gene (MGG_13239). Second, we could also determine the metabolites of 18:2n-6 associated with the 7,8-LDS activity, their structures, and mechanism of formation with the aid of stereospecifically deuterated 18:2n-6 and by chiral analysis. Third, gene deletion showed that 7,8-LDS activity was not critical for sporulation or rice infection. (8R,11S)-DiHODE was formed by antarafacial dioxygenation at C-8 and suprafacial hydroxylation at C-7 and C-11. This mechanism of biosynthesis appears to be a characteristic feature of the family of LDS enzymes (7,16,42). We conclude that the metabolites discussed above can be formed by 7,8-LDS, as summarized in Fig. 8.
⌬7,8-LDS infected rice leaves and roots to the same extent as native M. oryzae. We found no apparent defects in conidia and appressoria formation or in radial growth. This might due to alternative compensatory pathways, which seem to be common in rice blast (45). Recent genome analysis suggests that LDS homologues occur in many filamentous fungi (46), and it may be linked to housekeeping functions. A protein with homology to LDS is abundant in teliospores of Ustilago maydis and is thought to participate in the mobilization of stored lipids, although gene deletion of this protein did not produce a distinct phenotype (47). In contrast, gene deletion of 5,8-LDS and (10R)-DOX affected biological processes of A. fumigatus and A. nidulans (15,16).
Degradation of lipid bodies, activation of lipases, and generation of glycerol are important steps in formation of the turgor pressure of appressoria of M. oryzae (27). Protein kinases are important regulators of this process (29 -32). Glycogen formation and lipid degradation occur rapidly in the ⌬mac1 sum1-99 mutant (27). Interestingly, mycelia of ⌬mac1 sum1-99 formed much larger amounts of (10R)-HPODE and (5S,8R)-DiHODE in comparison with the ⌬mac1 mutant and M. oryzae Guy11. Could this be due to up-regulation of the second gene product (MGG_10859) with homology to 7,8-LDS?
Transcripts of the two LDS homologues were detected by PCR analysis in Guy11 and the ⌬mac1 mutant, and MGG_10859 mRNA appeared to be up-regulated in the ⌬mac1 mutant. In contrast, MGG_10859 mRNA was not detectable in the mutant with cAMP-independent PKA activity (⌬mac1 sum1-99). MGG_10859 could therefore not be linked to biosynthesis of (10R)-HPODE nor (5S,8R)-DiHODE. As shown in Table 1, MGG_10859 did not show specific sequence homology to any of the listed enzymes, and the homology did not suggest a specific catalytic activity. Further studies will be needed to determine its function.
The genome of M. oryzae contains over 120 CYP genes, grouped with only a few members per family (5). It is conceivable that 5,8-DiHODE could be formed by any of these P450s of the suppressor mutant. We found, however, that 5,8-LDS (and 7,8-LDS) activities were present in the high speed supernatant and negligible in the microsomal fraction, which seems to exclude microsomal enzymes.
The P450 domain of 7,8-LDS contains several sequences with homology to the PKA motif, Arg-Arg-Xaa-(Ser/Thr). Phosphorylation of P450 at this motif by PKA has so far only been found to reduce P450 activity without influence on product specificity (48). Whether the oxylipin biosynthesis of the ⌬mac1 sum1-99 mutant can be attributed to post-translational modification of 7,8-LDS merits further investigation.
Mycelia of M. oryzae also oxidized medium and long chain polyunsaturated fatty acids by hydroxylation and epoxidation, as summarized in the supplemental material. CYP102 of Bacillus megaterium is the prototype microbiological fatty acid 2and 3-hydroxylase and fatty acid epoxygenase (49). Blast search with CYP102 revealed three homologous hypothetical enzymes (numbers 3-5) of the CYP505A family in M. oryzae. Further studies will be needed to link CYP505 of M. oryzae to fatty acid oxygenation in the same way as the terminal oxygenases of Fusarium oxysporum were linked to CYP404 (50). Recent studies of Talbot and co-workers (28) showed that peroxisomal ␤-oxidation is essential for M. oryzae during appressorium-mediated rice infection, and the lipid bodies of spores may provide the fuel for the ␤-oxidation.
In summary, we report the first studies on the biological function of 7,8-LDS in M. oryzae, an important fungal grass pathogen and model organism. We identified the 7,8-LDS gene (MGG_13239) and determined the structure and mechanism of formation of metabolites. Gene deletion showed that 7,8-LDS was not critical for sporulation nor for rice infection. PKA, which has profound effects on these processes, augmented formation of two previously unrecognized oxylipins of M. oryzae, (5S,8R)-DIHODE and (10R)-HPODE. We conclude that M. oryzae can oxidize to 18:2n-6 to a more complex set of oxylipins than previously anticipated.