Manganese Lipoxygenase

A linoleic acid (13R)-lipoxygenase was purified to homogeneity from the culture medium of Gäumannomyces graminis, the take-all fungus, by hydrophobic interaction, cation exchange, lectin affinity, and size-exclusion chromatography. The purified dioxygenase lacked light absorption between 300 and 700 nm. Gel filtration indicated an apparent molecular mass of ∼135 kDa in 6 murea and ∼160 kDa in buffer. SDS-polyacrylamide gel electrophoresis (PAGE) showed that the enzyme was heterogeneous in size and consisted of diffuse protein bands of 100–140 kDa. Treatment with glycosidases for N- and O-linked oligosaccharides yielded a distinct protein of ∼73 kDa on SDS-PAGE. Atomic emission spectroscopy indicated 0.5–1.0 manganese atom/enzyme molecule. The isoelectric point was ∼9.7, and the enzyme was active between pH 5 and 11 with optimum activity at pH 7.0. For molecular oxygen,K m was 30 μm andV max 10 μmol mg−1min−1; for linoleic acid,K m was 4.4 μmol, V max 8.2 μmol mg−1min−1, and the turnover number 1100 min−1. The enzyme oxidized linolenic acid twice as fast as linoleic acid. The main products were identified by mass spectrometry as 13-hydroperoxy-(9Z,11E,15Z)-octadecatrienoic and 13-hydroperoxy-(9Z,11E)-octadecadienoic acids, respectively. After reduction of the hydroperoxide, steric analysis of methyl 13-hydroxyoctadecadienoate by chiral high performance liquid chromatography yielded one enantiomer (>95%), which co-eluted with the R-stereoisomer of methyl (13R,13S)-hydroxyoctadecadienoate. Arachidonic and dihomogammalinolenic acids were not substrates, while oxygen consumption, UV analysis, and mass spectrometric analysis indicated that γ-linolenic acid was oxygenated both at C-11 and C-13. The enzyme was active at 60 °C and after treatment with 6 murea. It was strongly inhibited by 10–50 μmconcentrations of eicosatetraynoic acid and a lipoxygenase inhibitor (N-(3-phenoxycinnamyl)acetohydroxamic acid), but many other lipoxygenase inhibitors (100 μm) were without effect. We conclude that, after deglycosylation, the enzyme has the same size on SDS-PAGE as mammalian and marine lipoxygenases, but it differs from all previously described lipoxygenases in three ways. It is secreted, it forms (13R)-hydroperoxy-(9Z,11E)-octadecadienoic acid, and it contains manganese.

The catalytic center of lipoxygenases contains iron, which according to analysis of crystal structures and sequence information is bound to three conserved histidine residues, a Cterminal isoleucine, water, and a variable sixth ligand (9 -11). The sixth ligand of plant enzymes is an asparagine residue, while mammalian enzymes have a histidine, asparagine, or serine at this position (12). The arachidonate (8R)-lipoxygenase of P. homomalla has conserved these histidine and asparagine residues but differs from S-lipoxygenases in its C-terminal amino acid, which is a threonine residue (3). The difference between S-and R-lipoxygenases might therefore be due to substrate binding and control of oxygenation rather than fundamental differences in metal ligands at the active site. A recent comparison of the crystal structures of soybean lipoxygenase-1 and a reticulocyte 15-lipoxygenase shows that the three-dimensional structure has been remarkably well conserved between these S-lipoxygenases (13).
Mammalian lipoxygenases metabolize arachidonic acid as their most important physiological substrate, while plant lipoxygenases oxygenate linoleic and linolenic acids (1)(2)14). These fatty acid hydroperoxy metabolites can be further converted to a number of products with a diverse range of biological actions, e.g. to leukotrienes in animals and the growth regulatory hormone jasmonic acid in plants (4,5). Many investigators believe that plant lipoxygenases may contribute to the chemical defense of plants against pathogens (14,15). For example, lipoxygenases might be involved in the resistance of rice to the rice blast fungus, Magnaporthe grisea, oats to Puccinia coronata avena, and tomato to Cladosporium fulvum (16 -18). The possibility that fungal pathogens can secret a lipoxygenase has not been explored.
We now report that filtrate of the culture medium of G. graminis contains a linoleate dioxygenase, which appears to be secreted by the fungus in large quantities. The secretion of a dioxygenase by an important pathogen has not been reported previously. The goal of the present report was to purify and to characterize this enzyme. The dioxygenase was found to have several unprecedented features. First, it showed R stereospecificity in a typical lipoxygenase reaction with polyunsaturated C 18 fatty acids as substrates. The main metabolites of linoleic and linolenic acids were cis-trans-conjugated (13R)-hydroperoxy fatty acid derivatives (13-HPODE and 13-HPOTrE(n-3), respectively). Second, the dioxygenase was found to contain manganese. Third, the enzyme catalyzed hydrogen abstraction, oxygen insertion, and biosynthesis of an intermediate hydroperoxide in a different way than all previously described lipoxygenases with either S or R stereospecificity. Based on these observations, the new enzyme is designated manganese lipoxygenase (Mn-LO). The unique oxygenation mechanism is described in detail in the accompanying paper (26 (20,27). Equipment and reagents for SDS-PAGE and isoelectric focusing were from Bio-Rad. Buffers were prepared with water from Milli-Q plus 185 (Waters). High molecular mass protein standards for SDS-PAGE were from Amersham Pharmacia Biotech; standards for gel filtration were from Bio-Rad. Endoglycosidase F/N-glycosidase F (50 milliunits/l) and O-glycosidase (bovine serum albumin-free, 1 milliunit/l) were from Boehringer Mannheim (Mannheim, Germany). Chelex 100, CHAPS, GSH, esculetin, and methyl-␣-D-glucopyranoside were from Sigma. Tween 20 was from U. S. Biochemical Corp. BW 4AC (N-(3-phenoxycinnamyl)acetohydroxamic acid) was from Wellcome Research Laboratories (Beckenham, UK), ICI 230487 from Zeneca Pharmaceuticals (Macclesfield, UK), zileuton from Abbott Laboratories (Chicago, IL), and ETYA from Hoffman-La Roche (Basel, Switzerland). Racemic 13-HODE was prepared by autoxidation, and (13S)-HPODE, (13S)-HPOTrE(n-6), and (13S)-HPOTrE(n-3) were prepared enzymatically (lipoxidase type IV, Sigma) as described (27).
Enzyme Assays-Enzyme activity was monitored by UV analysis, oxygen consumption or TLC analys. For UV analysis, the enzyme was added to microcuvette containing 0.24 mM linoleic acid (in 0.15 ml of buffer A: 10 mM triethanolamine-HCl (pH 7.0), 1 mM EDTA, 2 mM NaN 3 , 0.04% Tween 20) at room temperature. The absorbance at 235 nm was followed, and the rate was estimated from a linear part of the curve. Other fatty acids were incubated in the same way. Kinetic constants were determined at 25°C in triplicate from five concentrations of 18:2n-6 and 18:3n-3. Soybean lipoxygenase (Sigma) was used as a reference enzyme. Oxygen consumption was monitored at 25°C in buffer A with 0.13 mM substrates added in ethanol. Kinetic constants were estimated from triplicate analysis of five different concentrations of oxygen. For analysis by TLC, enzyme was mixed with 85 l of buffer A and incubated, in a total volume of 100 l, with 1 nmol of [ 14 C]18:2n-6 (10 M). After termination with ethanol and extractive isolation on SepPak/C 18 , the products were separated by TLC (toluene/ dioxane/formic acid/acetic acid, 81/14/1/1) and the plates were scanned for radioactivity (Berthold Dü nnschicht scanner model II, Kebo, Stockholm). The effect of lipoxygenase inhibitors on the enzyme was studied with TLC or with UV analysis.
Enzyme Purification-The fungus was cultivated in 2 liters of culture medium (containing (per liter): 20 g of D-glucose and 5 g each of yeast extract (Merck), peptone from soy meal (Merck), K 2 HPO 4 , and NaCl (pH ϳ 7.7)) in a 5-liter Erlenmayer flask under normal atmosphere with gentle shaking (125 rpm) for 7 days at 22°C. The broth was then filtered through a nylon net to remove the mycelia (typically 50 g of mycelia (wet weight)/liter). The following steps were performed at 4°C. The medium was adjusted to 0.55 M ammonium sulfate, 2 mM NaN 3 (pH 7.2) (NaOH), carefully mixed, and then filtered (0.45 m, type HA, Millipore). The filtrate was loaded on a Phenyl-Sepharose CL-4B column (2.6 ϫ 20 cm) in 0.05 M KPB (pH 7.2) with 0.6 M ammonium sulfate at 5 ml/min. The column was then washed with two to three column volumes of the same buffer. The column was then eluted with 0.01 M KPB (pH 6.8), 0.5 mM GSH, 2 mM NaN 3 , 0.04% Tween 20, which yielded enzyme fraction I. Additional enzyme (fraction II) could be eluted with 5 mM KPB (pH 6.8), 6 M urea, 5 mM EDTA, 0.5 mM GSH. Fractions I and II were initially purified separately, but they were found to contain the same enzyme as judged from the isoelectric points, chromatography on Superdex 200 HR 10/30, and SDS-PAGE before and after deglycosylation. Fractions I and II were therefore combined. After dialysis, the material was further purified by cation exchange chromatography (CM Sepharose CL-4B, 1.8 ϫ 3 cm) in 0.01 mM KPB (pH 6.8) with 2 mM NaN 3 , 0.5 mM GSH, and 0.04% Tween 20. The column was then eluted in one step with 0.2 M NaCl in the same buffer. The material was then purified either by affinity chromatography on ConA-Sepharose (2. Material from the final gel filtration was referred to as purified enzyme, and it was used throughout this and the accompanying paper (26).
Isoelectric Focusing-The purified enzyme was not retained on Mono P HR 5/5 in 0.025 M diethylamine-HCl (pH 9.5). Isoelectric focusing was performed in a preparative cell (Rotofor, Bio-Rad) with 1% Bio-Lyte 3/10 (Bio-Rad). Isoelectric focusing was performed at 12 watts (4°C, 3 h) with purified enzyme from fraction I (20 g) and fraction II (12 g). The fractions from the Rotofor were assayed for enzyme activity and protein content and analyzed by SDS-PAGE.
LC-MS Analysis-The pump for HPLC was from Thermo Separation Products (SpectraSYSTEM P2000) equipped with a degasser (Uniflow Dagasys DG-1310, Tokyo, Japan). The column contained octadecasilane silica (5 m, 250 ϫ 2 mm or 150 ϫ 2 mm; Kromasil 5 C 18 100A, Phenomenex) and a guard column (Opti-Gard, 1 mm C 18 ; Optimize Technologies). The RP-HPLC column was eluted with methanol/water/ acetic acid, 80/20/0.01, at 0.2 ml/min. The effluent first passed a UV detector (Kratos Spectroflow 757) equipped with an integrator (Merck Hitatchi D-2500) and then into an ion trap mass spectrometer (LCQ, Finnigan MAT, San Jose, CA), where it was subject to atmospheric pressure chemical ionization. The vaporizer temperature was 450°C, the capillary temperature was 150°C, and the collision energy was set to 25-35%. Negative ions were monitored, and prostaglandin F 1␣ was used for tuning. In a few experiments, electrospray ionization was used.
GC-MS Analysis-Hydroperoxides were reduced with SnCl 2 or with NaBH 4 . Methyl esters were prepared with diazomethane. Hydrogenation was performed with Pd/C in ethanol for 2 min with a gentle flow of hydrogen. Trimethylsilyl ethers were prepared by treatment with bis(trimethylsilyl)trifluoroacetamide and pyridine and methyl esters with diazomethane (27). GC-MS analysis with electron impact ionization was performed as described (27). C-values were determined by the retention times of fatty acid methyl esters (18:0, 20:0, 22:0, and 24:0).
Spectroscopy-Light absorption spectroscopy was performed with a dual beam spectrophotometer (Shimadzu UV-2101PC), which was equipped with a thermostated cuvette holder. Microcuvettes (0.15 ml) were used routinely. The cis-trans-conjugated hydro(pero)xy fatty acids were assumed to have a molar extinction coefficient of 23,000 (29). Protein concentration was determined as described by Bradford (30) using bovine albumin as a standard.
For atomic emission spectroscopy, plastic tubes, pipette tips, and tubings were washed overnight with 15% HNO 3 and rinsed with metal-free water. Purified enzyme (1 mg) in 0.5 ml of buffer (0.05 M KPB, 3 mM NaN 3 ) was slowly passed through a column (16 ϫ 10 mm, Chelex 100, 50 -100 dry mesh) before analysis. The instrument, an inductively coupled plasma with atomic emission spectrometry (Spectroflame P), was used for quantitative and qualitative determination of the transition and other elements in the sample. The samples were diluted 1:5 with pure water and introduced at 1 ml/min, and the plasma power was 1200 watts. Iron was also determined with graphite furnace atomic absorption spectrometry.
Other Analyses-A Clark-type oxygen electrode (probe 5331 with a high sensitivity membrane and oxygen monitor 5300, YSI Inc., Yellow Springs, OH) was used to measure oxygen consumption during enzyme catalysis in a thermostated oxygen cell (cell volume 1.5 ml, Gilson; 25°C). The cell was first filled with air-saturated buffer of 25°C and the substrate (in ethanol) was then added. After stabilization (1-2 min), the reaction was started by injection of a few microliters of the enzyme. Nitrogen was used to obtain different concentrations of oxygen in the cell. Enzyme activity was estimated from the initial rate of oxygen uptake, assuming an initial concentration of 0.24 mM oxygen. SDS-PAGE was performed with 7.5% resolving and 3.0% stacking gels (with MiniProtean II, Bio-Rad) using the method of Laemmli (31). Runs were carried out at 100 V for 10 min, then at 200 V for 35 min. Proteins were detected by silver staining as described (20).

Enzyme Purification
The results of the enzyme purification are shown in Table I. Two protocols were used, which differed only in cation exchange chromatography (Mono S HR 5/5) or affinity chromatography (ConA-Sepharose). The latter method was more con-venient. The enzyme was purified 188 -286-fold to an apparent specific activity of 8.1-8.4 mol Ϫ1 min Ϫ1 mg Ϫ1 with linoleic acid as a substrate ( Table I).
The enzyme was purified in four main steps. It was first "captured" by hydrophobic interaction chromatography (Fig.  1A). It could only be partly eluted by a decrease in salt concentration, which yielded fraction I. A second fraction with enzyme activity eluted with 6 M urea (fraction II). Fractions I and II were initially purified separately by both protocols, but found to give the same protein after deglycosylation (see below). The two fractions were therefore combined and purified by open column cation exchange chromatography (CM Sepharose CL-6B). The enzyme eluted with 0.2 M NaCl and was then purified by affinity chromatography (ConA-Sepharose) or by cation exchange HPLC (Mono S HR 5/5). The enzyme eluted mainly in a broad peak from this HPLC cation exchange column (Fig. 1B).
The enzyme was finally purified by gel filtration. Gel filtration of native enzyme in buffer with 0.15 M NaCl and 0.5 mM CHAPS indicated a molecular mass of ϳ160 kDa (Fig. 1C). Eleven fractions (0.2 ml) were collected over this peak and subjected to SDS-PAGE analysis. Each fraction contained a diffuse protein band, which appeared to decrease in mean size from about 140 to 100 kDa ( Fig. 2A). Gel filtration was also performed in 6 M urea in 0.05 M KPB. This yielded a mass of 135 kDa (data not shown), which is within the 100 -140-kDa range obtained by SDS-PAGE. Each fraction of the purified enzyme from gel filtration showed a single, broad band on SDS-PAGE with silver staining. The material from gel filtration was clearly heterogeneous in size, but appeared free from other contaminating proteins.
These results suggested that the enzyme could be glycosylated. This was confirmed by enzymatic deglycosylations with endoglycosidases for both N-and O-linked oligosaccharides. SDS-PAGE now showed a distinct protein of ϳ73 kDa (Fig. 2B). As discussed above, the hydrophobic interaction chromatography yielded two enzyme fractions (Fig. 1A). After purification and deglycosylations, fractions I and II yielded the same protein band of ϳ73 kDa on SDS-PAGE (Fig. 2C).

Prosthetic Group
The purified enzyme was colorless and lacked significant light absorption between 300 and 700 nm (Fig. 3), which suggested that the enzyme lacked heme. However, concentrated solutions (1 mg/ml) had a pale yellow color.
Atomic emission spectroscopy indicated that the enzyme sample contained, in quantities above 25 ng/ml, only Na, K, S, P, and Mn (see Fig. 3, inset). Mg, Ni, Fe (Fig. 3, inset) and Cu could not be detected. For K, Na, and P, the concentration was the same as in the buffer. The Mn concentration was 2 M, the S concentration 0.15 mM, and the protein concentration 0.3 mg/ml. This corresponds to ϳ2.2 M apoprotein at a mass of pH Optimum and Isoelectric Point pH Optimum-Enzyme activity was first tested in 1 pH unit intervals from pH 2 to 12 and then at 0.5 pH unit intervals around the broad pH optimum, which was found to be centered at pH 7.0. Changes of pH had little effect on the enzyme activity in the range pH 6 -8 with linoleic acid as a substrate. Incubations were routinely performed at pH 7.0, but in some experiments pH 7.4 was used. Significant enzyme activity was detectable at both ends of the pH scale, at pH 5 about 64% and at pH 11 69% of the activity at pH 7.
Isoelectric Point-The enzyme was not retained on a chromatofocusing column (Mono P HR 5/5) at pH 9.5. Isoelectric focusing (Rotofor, Bio-Rad) of the purified enzyme yielded the highest specific activity in the fraction corresponding to pI ϳ9.7. This activity also corresponded to the peak of proteins as confirmed by SDS-PAGE of this and nearby fractions. About 50% of the enzyme appeared to be inactivated after the isoelectric focusing.

Catalytic Properties
To obtain consistent and reproducible results with high concentration of substrates, it was essential to have 0.04% Tween 20 present in the assay buffer.

UV Analysis
Biosynthesis from 18:2n-6 and 18:3n-3 started after a short time lag and proceeded at a linear rate for at least 25 min, as judged from the increase in UV absorption at 235 nm (data not shown). The enzyme metabolized 18:3n-3 twice as fast as 18: 2n-6, while 18:3n-6 was metabolized to UV absorbing metabolites only slowly (Fig. 4A). Neither 20:4n-6 nor 20:3n-6 were metabolized to a significant degree. 22:6n-3 appeared to be metabolized slowly. The methyl ester derivatives of 18:2n-6 and 18:3n-3 were not converted to UV absorbing products (data not shown). Addition of Fe 2ϩ , Fe 3ϩ , or Mn 2ϩ did not appear to stimulate enzyme activity.

Oxygen Consumption
The oxygen consumption of the enzyme with 18:2n-6 and 18:3n-3 as substrates appeared to be in the same order as the biosynthesis of metabolites on UV analysis. Oxygen consumption in the presence of 18:2n-6, 18:3n-6, and 20:4n-6 is shown in Fig. 4B. 18:2n-6 caused a large increase in oxygen consumption, but there was no apparent oxygen consumption with 20:4n-6 (or 20:3n-6; data not shown). 18:3n-6 caused a signifi-cant increase in oxygen consumption, which appeared to be larger than expected from the formation of metabolites with UV absorbance at 235 nm. This was probably due to biosynthesis of the 11-hydroperoxy metabolite of 18:3n-6 as discussed below.

Linoleic Acid
18:2n-6 was incubated with enzyme in a cuvette until the substrate appeared to be consumed. The UV spectrum showed max at 235 nm. After extractive isolation, the material was analyzed by LC-MS with atmospheric pressure chemical ionization (direct injection). This yielded a strong signal at m/z 311 ((M Ϫ H) Ϫ ), which suggested a hydroperoxy metabolite. MS-MS of m/z 311 supported this notion. After reduction with SnCl 2 , only one major metabolite was formed, as judged from RP-HPLC or SP-HPLC with UV analysis (235 nm). LC-MS and GC-MS of this compound yielded mass spectra, which were were similar to those reported for 13-HODE and its methyl ester trimethylsilyt ether derivative, respectively (27,32). GC-MS analysis of the hydrogenated compound yielded consistent results, but the hydrogenated sample also contained a few percent of hydrogenated 9-HODE as judged from weak signals at m/z 259 and m/z 229. We conclude that the main product was 13-HPODE under conditions with substrate depletion.
Metabolites formed from 18:2n-6 by the enzyme at pH 7.0 and pH 9.0 were analyzed and compared in order to detect possible formation of 9-HODE. SP-HPLC showed no evidence for a significant biosynthesis of 9-HODE in comparison with 13-HODE at either pH. The Mn-LO thus differs from soybean lipoxygenase (33). However, biosynthesis of 2-3% of 9-HODE can be detected (26).

Kinetic Constants
Apparent and turnover numbers for Mn-LO are summarized in Table II. The data are based on UV analysis at 235 nm for metabolites of 18:2n-6 and 18:3n-3. 18:2n-6 is metabolized to ϳ20% 11-HPODE without UV absorption at 235 nm. The V max for 18:2n-6 (and 18:3n-3) should therefore be considered as a lower estimate only. K m and V max for oxygen were determined with linoleic acid as a substrate (Table II). The apparent turnover number for 18:2n-6 was ϳ1100 min Ϫ1 , and it was 2400 min Ϫ1 for 18:3n-3.
Heating above 63°C for 5 min decreased enzyme activity (Fig. 7B). The enzyme activity decreased sharply at higher temperatures and was totally lost at 75°C. This was determined at 25°C by UV analysis of metabolites after initial cooling on ice. The data suggested a thermal inactivation energy of Ϫ186 kJ/mol (Arrhenius plot; r ϭ 0.93). For comparison, Fig. 7B also shows the effect of temperature on the activity of soybean lipoxygenase, and the data suggest a lower thermal inactivation energy (Ϫ76 kJ/mol; r ϭ 0.99).
The purified enzyme could be stored for several months at 4°C in buffer with 3 mM NaN 3 and 0.2 mM GSH without apparent loss of enzyme activity. Enzyme activity as measured directly after freezing (Ϫ80°C) and thawing was reduced by about 50%, but the enzyme activity then appeared to increase with time during storage at 4°C. Treatment of the purified enzyme with 6 M urea had little effect on enzyme activity (as measured after removal of the urea). The peak marked with an arrow contained material with the same mass spectrum and C-value as (13S)-HOTrE(n-6) formed by soybean lipoxygenase. The metabolite of the largest peak was identified as 11-HOTrE(n-6) (see "␥-Linolenic Acid" under "Results"). Before analysis, the metabolites were reduced with NaBH 4 and extracted with ethyl acetate. Trimethylsilyl ether methyl ester derivatives and electron impact ionization were used.

Effects of Enzyme Inhibitors
The formation of radiolabeled products from 10 M 14 Clabeled 18:2n-6 was inhibited by over 50% by 10 M ETYA (and by 10 M arachidonic acid). Three lipoxygenase inhibitors (22)

Different Strains of Take-all
Five different isolates of G.graminis (var. tritici, var. avenae, var. graminis) were investigated for secretion of Mn-LO to the culture medium. All isolates secreted the enzyme to the medium. DISCUSSION We report three novel findings. First, the take-all fungus secretes a stable dioxygenase (Mn-LO), which can be recovered from cell-free filtrates of the culture medium in large amounts (Table I). This is close to an ideal naturally occurring enzyme expression system. Second, we show that this enzyme forms (13R)-HPOTrE(n-3) and (13R)-HPODE as major products from 18:3n-3 and 18:2n-6, respectively, whereas 18:3n-6 is a relatively poor substrate. It is converted to 11-HPOTrE(n-6) and 13-HPOTrE(n-6) as two main metabolites. Third, the enzyme contains manganese. Manganese has not been reported previously in any lipoxygenase, but manganese has been found in other non-heme dioxygenases (36,37).
Filamentous fungi can secrete many different enzymes, and the majority are believed to be glycosylated. This process appears to be associated with the growth of the hyphal tip (38). Secreted enzyme can digest cell walls and cell membranes of plants. It is well known that fungi infecting plants can induce lipoxygenases of plant origin, presumably as a defense reaction. We now report for the first time that a lipoxygenase-like dioxygenase can be secreted by an important pathogen, the take-all fungus. The only other secreted lipoxygenase known to us is arachidonate 15-lipoxygenase, presumably of type 2 (39), which is secreted into human seminal fluid in subcellular organelles ("prostasomes") from the prostate gland (40,41).
The secreted Mn-LO metabolized polyunsaturated C 18 fatty acids, but it did not oxygenate arachidonic and dihomogammalinolenic acids. This finding was not unexpected. G. graminis belongs to the family of Ascomycetes, which have low or undetectable levels of C 20 fatty acids (42). G. graminis infects grasses and these are also relatively poor in C 20 fatty acids.
However, the stereospecific R-lipoxygenation at C-13 of linoleic acid was unexpected. This is the first report on an R-lipoxygenase, which introduces oxygen at the 6 carbon of linoleic acid.
Mammalian lipoxygenases will form stereochemically pure hydroperoxides with S configuration. This occurs by hydrogen abstraction from the bis-allylic carbon of the pentadiene and antarafacial insertion of oxygen. In contrast to mammals, fresh and salt water species of invertebrate contain R-lipoxygenases   (3,43,44). These enzymes oxygenate arachidonate in the 5R, 8R, 11R, or 12R configurations. The arachidonate (8R)-lipoxygenases have been studied in detail from two corals, Pseudoplexaura porosa and P. homomalla (46 -48), and from nervous tissue of Aplysia californica (43). Brash and co-workers (3) cloned and sequenced the (8R)-lipoxygenase from P. homomalla and found that it belonged to the lipoxygenase gene family. As far as is known, these marine R-lipoxygenases also abstract hydrogen and insert oxygen in an antarafacial way (3,45).
Marine R-lipoxygenases are generally believed to contain iron, but this may not have been confirmed by direct analysis. Lipoxygenases were originally named based on their catalytic activity (49). They were later found to catalyze stereospecific formation of a cis-trans-conjugated hydroperoxy fatty acids (50). The dioxygenase we have purified thus catalyzes a typical lipoxygenase reaction, although it can also oxygenate C-11. An interesting question is whether the Mn-LO is a member of the lipoxygenase gene family. The Mn-LO had an apparent mean size of ϳ135 kDa in 6 M urea, but after enzymatic deglycosylation the apparent mass was reduced to ϳ73 kDa. The latter figure is smaller than the mass of many plant lipoxygenases (ϳ95 kDa), but close to the mass of both mammalian S-lipoxygenases (ϳ75 kDa) and the (8R)-lipoxygenase of P. homomalla (predicted mass ϳ76 kDa from cDNA sequence). The similarity in size suggests that the Mn-LO could be related to mammalian and marine lipoxygenases. On the other hand, the Mn-LO was inhibited by ETYA and BW 4AC but not by 0.1 mM concentrations of three other lipoxygenase inhibitors. The Mn-LO also has a different oxygenation mechanism than lipoxygenases (26). These observations may argue against a structural similarity between Mn-LO and Fe-lipoxygenases. However, there are structurally related non-heme dioxygenases, which depend either on iron or on manganese for catalytic activity.
Bacterial dioxygenases can cleave catechols either between the hydroxyls or at one side of the diol. Extradiol-cleaving catechol dioxygenases are usually iron-dependent, but the 3,4dihydroxyphenylactetate 2,3-dioxygenases of Bacillus brevis and Arthrobacter globiformis CM-2 are manganese-dependent for catalytic activity (37,51). There is 78% identity in 337 of 365 amino acids between the manganese-dependent 2,3-dioxygenase of Arthrobacter and the corresponding iron-dependent enzyme of Brevibacterium fuscum (52). The similarity to other iron-dependent extradiol cleaving dioxygenases is less striking, but the three known manganese-dependent 2,3-dioxygenases are considered to belong to the major extradiol dioxygenase family (36,53). Site-directed mutagenesis of the Arthrobacter 2,3-dioxygenase suggested that iron-and manganese-dependent enzymes employ the same conserved metal binding amino acid ligands, two histidines and a glutamic acid residue (51). Superoxide dismutase also uses identical amino acid ligands, three histidines and an aspartic acid residue, for the Fe-, Mn-, and Fe/Mn-dependent enzymes (54 -56). It will clearly be of interest to sequence the Mn-LO, to determine the manganese binding residues and to determine whether this protein is related to the lipoxygenase gene family.
The biological function of Mn-LO of G. graminis is unknown. The fact that the enzyme is secreted, highly glycosylated, and very stable suggests that it might be involved in infection of grass roots. Wheat is particularly sensitive to G. graminis. The secreted enzyme could form 13-HPODE and 13-HPOTrE(n-3), which might conceivably lead to oxidative damage of wheat roots. The take-all fungus causes reduced wheat crops worldwide. New methods to combat G. graminis would be useful, and the biological importance of the secreted Mn-LO may merit further investigation.