Catalytic Convergence of Manganese and Iron Lipoxygenases by Replacement of a Single Amino Acid*

Background: 13R-MnLOX catalyzes suprafacial hydrogen abstraction and oxygenation in contrast to sLOX-1. Results: Mutation of one residue of 13R-MnLOX altered hydrogen abstraction and oxygenation to antarafacial. Conclusion: Replacement with the corresponding residue of soybean LOX-1 yielded catalytic convergence. Significance: The suprafacial oxygenation mechanism can be attributed to a single amino acid substitution. Lipoxygenases (LOXs) contain a hydrophobic substrate channel with the conserved Gly/Ala determinant of regio- and stereospecificity and a conserved Leu residue near the catalytic non-heme iron. Our goal was to study the importance of this region (Gly332, Leu336, and Phe337) of a lipoxygenase with catalytic manganese (13R-MnLOX). Recombinant 13R-MnLOX oxidizes 18:2n-6 and 18:3n-3 to 13R-, 11(S or R)-, and 9S-hydroperoxy metabolites (∼80–85, 15–20, and 2–3%, respectively) by suprafacial hydrogen abstraction and oxygenation. Replacement of Phe337 with Ile changed the stereochemistry of the 13-hydroperoxy metabolites of 18:2n-6 and 18:3n-3 (from ∼100% R to 69–74% S) with little effect on regiospecificity. The abstraction of the pro-S hydrogen of 18:2n-6 was retained, suggesting antarafacial hydrogen abstraction and oxygenation. Replacement of Leu336 with smaller hydrophobic residues (Val, Ala, and Gly) shifted the oxygenation from C-13 toward C-9 with formation of 9S- and 9R-hydroperoxy metabolites of 18:2n-6 and 18:3n-3. Replacement of Gly332 and Leu336 with larger hydrophobic residues (G332A and L336F) selectively augmented dehydration of 13R-hydroperoxyoctadeca-9Z,11E,15Z-trienoic acid and increased the oxidation at C-13 of 18:1n-6. We conclude that hydrophobic replacements of Leu336 can modify the hydroperoxide configurations at C-9 with little effect on the R configuration at C-13 of the 18:2n-6 and 18:3n-3 metabolites. Replacement of Phe337 with Ile changed the stereospecific oxidation of 18:2n-6 and 18:3n-3 with formation of 13S-hydroperoxides by hydrogen abstraction and oxygenation in analogy with soybean LOX-1.

Lipoxygenases (LOXs) 2 are fatty acid dioxygenases with a non-heme catalytic metal, usually iron (1)(2)(3). They oxidize polyunsaturated fatty acids with 1Z,4Z-pentadiene units to hydroperoxides, which are precursors to biological mediators in mammals, plants, and fungi (2,4,5). In humans, LOX products are formed from oxidation of arachidonic acid, and these eicosanoids interact with G-protein-coupled receptors; modulate asthma and allergic inflammation, atherosclerosis, and skin water permeability; and contribute to oxidative stress and cancer development (4, 6 -8). The versatile role of LOX metabolites has led to the development of enzyme inhibitors and receptor antagonists (4). Plant LOX pathways oxygenate C 18 fatty acids to volatile jasmonates, aldehydes, and a series of other metabolites (9). These oxylipins participate in the chemical warfare between plants and microorganisms or herbivores and act as signals in plant development (2,5,9,10).
All LOXs belong to the same gene family, which is characterized by sequence homology of their catalytic domains and by conserved iron ligands (1,3,(11)(12)(13). Crystal structures are available for several soybean LOXs (sLOXs), coral 8R-LOX, arachidonate 15-and 5-LOX, and one prokaryotic LOX (14 -19). The binding of fatty acids to the active sites has so far with one recent exception (18) been deduced by modeling, site-directed mutagenesis, and oxidation of fatty acids with different chain lengths and numbers of double bonds (16, 20 -25).
LOXs abstract the bisallylic hydrogen from C-3 of the 1Z,4Zpentadiene of unsaturated fatty acids followed by insertion of molecular oxygen at C-1, C-5, or occasionally C-3 with biosynthesis of hydroperoxy fatty acids (26,27); the lipoxidation of 18:2n-6 is outlined in Fig. 1A. Recently, a U-shaped substrate channel of LOX was proposed (16). In this model based on the coral 8R-LOX crystal structure, the carboxyl group of fatty acids binds to one of the two channel entrances with the -end embedded in the protein (16). The two opposite orientations will allow hydrogen abstraction from C-3 of the 1Z,4Z-pentadiene with antarafacial oxygen insertion at either C-1 (e.g. 8R-LOX) or at C-5 (e.g. sLOX-1) (16). Three of the four conserved metal ligands are His residues, and the fourth is the carboxyl of the C-terminal amino acid, usually Ile. Two of these His residues are found in a specific sequence hexamer (His-(Leu/ Trp)-Leu-(Asn/Arg)-(Thr/Gly)-His). The iron also binds water and interacts with a distant Asn or His residue, forming an octahedral ligand configuration (3). The metal-water complex constitutes the catalytic base (Fe 3ϩ OH Ϫ ) for hydrogen abstraction and redox cycling.
13R-MnLOX of Gaeumannomyces graminis and 9S-MnLOX 3 of Magnaporthe salvinii, fungal root and stem pathogens of wheat and rice, respectively, differ from FeLOX in three aspects (28). First, the metal-coordinating spheres are not identical. The manganese ligands are likely three His residues, the carboxylate of the C-terminal amino acid residue, and an Asn residue in analogy to iron ligands of FeLOX (11,29), but the hexamer motif of FeLOX is truncated to a pentamer, His-Val-Leu-Phe-His, in MnLOX (11). Second, MnLOX oxidizes hydroperoxides to peroxyl radicals at a rate almost 2 orders of magnitude higher than that of sLOX-1 (30). The mechanism is unknown but could be related to the metal centers. The redox potentials of Mn 2ϩ/3ϩ and Fe 2ϩ/3ϩ differ by a factor of 2, which is reflected in the catalysis of organic Mn-and FeLOX mimics (30,31). Third, 13R-and 9S-MnLOX 3 catalyze suprafacial hydrogen abstraction and oxygenation, whereas sLOX-1 and presumably all FeLOXs catalyze antarafacial hydrogen abstraction and oxygenation (26) (see Fig. 1A). How can this catalytic difference be explained?
9S-and 13R-MnLOX and sLOX-1 likely bind 18:2n-6 in the same "head-tail" orientation at pH 9, and all three enzymes abstract the pro-S hydrogen at C-11 and form a pentadienyl radical (26,32). 3 The same faces of the 9Z and 12Z double bonds of this radical are shielded from oxygen insertion by 9Sand 13R-MnLOX, respectively, whereas the other side of the 12Z double bond is shielded from oxygen in sLOX-1 (see Fig.  1A). Linoleic acid is thus likely positioned so that hydrophobic residues protect opposite faces of the 12Z double bond from oxygen insertion in 13R-MnLOX and sLOX-1. We lack detailed structural information on the active sites of 9S-and 13R-MnLOX, but it seems likely that many structural features of FeLOX are conserved in the vicinity of the catalytic manganese. It is therefore conceivable that minor structural changes could contribute to the unique oxidation mechanism of MnLOX. In this region, a conserved Leu of FeLOX is believed to position the 1Z,4Z-pentadiene close to the catalytic iron (16,33). Replacements of these Leu residues of sLOX-1 and coral 8R-LOX and the adjacent Ile residue of coral 8R-LOX have been investigated and found to reduce or even abolish catalysis (16,33). This Leu residue is conserved in 13R-and 9S-MnLOX. 3 The adjacent Phe residue is characteristic of 9S-and 13R-MnLOX as all other LOXs have Ile or Val residues at this position (see Fig. 1B).
We therefore hypothesized that Leu 336 , Phe 337 , and the pentamer motif could be important for substrate positioning and the catalytic properties of 13R-MnLOX. We also considered the nearby Gly 332 of the Gly/Ala determinant of regio-and stereospecific (R/S) oxidation at C-1 and C-5 of the 1Z,4Z-pentadiene (34). A few LOXs, including 9S-MnLOX, deviate from this rule of R/S stereospecificity (35,36). The Gly/Ala position also influences oxygen access and the hydroperoxide isomerase activities of eLOX-3 and 13R-MnLOX (37,38).
13R-MnLOX is secreted by the take-all fungus and can be expressed in Pichia pastoris as a secreted, glycosylated protein of 602 amino acids (ϳ91 kDa; Refs. 11 and 28). Recombinant 13R-MnLOX is transformed by spontaneous hydrolysis during storage to a smaller enzyme of uniform size (ϳ67.4 kDa) with virtually unchanged catalytic activity (11). This mini 13R-MnLOX is formed by loss of a sequence of glycosylated amino acid residues at the N-terminal end of the C2 domain, but the cleavage position is unknown. A crystal screening of mini 13R-MnLOX suggested that this protein might be amendable for crystallization. This led us to investigate whether it could be possible to determine the cleavage position and to express recombinant mini 13R-MnLOX as mini 13R-LOX retained the catalytic parameters of the native enzyme (11) and to use it for further studies. A recent report demonstrates that the C2 domain of mammalian LOX has little influence on catalytic activity and regiospecificity (39).
The main objective of the present investigation was to study replacements near the catalytic metal of 13R-MnLOX based on crystal structural information on the substrate channels of sLOX-1, coral 8R-LOX, human LOX, and sequence homology as outlined in Fig. 1B. Our main goal was to determine whether the His-Val-Leu-Phe-His motif and hydrophobic replacements of the Leu 336 and Phe 337 residues could influence the catalytic properties of 13R-MnLOX.
Expression Constructs of Mini 13R-MnLOX-The 13R-MnLOX precursor consists of 618 amino acids (supplemental Fig. S1). 13R-MnLOX (602 amino acids) was previously expressed and secreted by P. pastoris using pPICZ␣_ MnLOX_602 (11). This plasmid was used to obtain shorter expression constructs. The expression plasmid was modified by PCR technology to include an N-terminal His tag. In short, the codons for His 6 -Gln-Gln-Leu were introduced by using the EcoRI restriction site in the cloning region of pPICZ␣_MnLOX (aattcCATCATCATCATCATCATctgcagc) along with a restriction site for PstI (underlined). Next we created a unique restriction site for AflII (ctcaag 3 cttaag) without changing the coded sequence (L176K; numbered from the 13R-MnLOX precursor) by site-directed mutagenesis. This vector was restricted with PstI and AflII and then ligated with PCR fragments of different length to shorten the N-terminal end of the open reading frame. This yielded three constructs with 580, 570, or 555 residues to the C-terminal end (supplemental Fig. S1). pPICZ␣_Mini-MnLOX_580 thus coded for the sequence Glu-Phe-His 6 -Leu-Gln-Thr 39 -Val-Leu-Pro. . . . ..Val 618 . The DNA sequence differed in two positions (K52N and Y158C) from the sequence reported in GenBank TM (accession number AAK81882).
Site-directed Mutagenesis of 13R-MnLOX-pPICZ␣A_ MnLO_580 was modified by replacements of Leu 336 and Phe 337 by site-directed mutagenesis using Pfu and Phire polymerases and the oligonucleotides listed in supplemental Table S1. For L336A and L336G, the pPICZ␣A_MnLO_580_L336V was used as the template in the PCR. The PCR products were restricted with DpnI, analyzed by agarose gel electrophoresis, and used to transform E. coli (NEB5␣). Plasmid DNA was purified by anion exchange chromatography (Nucleobond AX) and screened by restriction analysis and sequencing at Uppsala Genome Center.
Expression of Recombinant Enzymes-P. pastoris was transformed as described (11). Transformants were selected on yeast peptone-dextrose-agar plates with Zeocin (100 g/ml) at 28°C (11). PCR analysis of genomic DNA was used to confirm recombination of the plasmid into the genomic DNA (42). Resistant colonies were first grown to generate biomass and then grown in buffered minimal medium as described (11). Protein biosynthesis was induced by 0.5-1% methanol added daily in the medium for 4 -5 days (11,43). The Pichia cells were harvested by centrifugation (5,000 ϫ g) and could be reused for additional expression. Secreted recombinant 13R-MnLOX was purified from the supernatant after centrifugation of the yeast suspension.
Enzyme Purification-Solid (NH 4 ) 2 SO 4 was added to the supernatant to a concentration of ϳ1 M, and the pH was adjusted to 7.0 -7.2 with 10 M KOH. After centrifugation and filtration, the material was loaded on a column for hydrophobic interaction chromatography (33 ϫ 40 mm, phenyl-Sepharose) in 1 M (NH 4 ) 2 SO 4 , 0.025 M potassium phosphate buffer (pH 7.1) and washed with the same buffer, and absorbed proteins were eluted with 25 mM potassium phosphate buffer (pH 7.1) (28). The column was washed with 70% ethanol and 6 M urea before the next analysis. The peak fractions were combined, concentrated to 1-2 ml by diafiltration (Ultracel 30K, Millipore), and stored at 4°C in 0.5-2 ml of 100 mM Tris-HCl (pH 8), 150 mM NaCl, 0.04% Tween 20, 1 mM NaN 3 . The proteins were analyzed by SDS-PAGE, and 13R-MnLOX and its mutants constituted at least 25-30% of the proteins as judged from the SDS-PAGE analysis (see Fig. 2A). Further partial purification was achieved by gel filtration (Superdex 200 HR 10/30) in 0.05 M potassium phosphate buffer (pH 7.4), 0.15 M NaCl, 1 mM NaN 3 , 0.04% Tween 20 with UV detection (280 nm).
Lipoxygenase Assay and UV Spectroscopy-Light absorbance was measured with a dual beam spectrophotometer (Shimadzu UV-2101PC). The cis-trans conjugated hydro(pero)xy fatty acids were assumed to have an extinction coefficient of 25,000 cm Ϫ1 M Ϫ1 . LOX activity was monitored by UV spectroscopy (237 nm) in 0.1 M NaBO 3 (pH 9.0) with 50 -100 M 18:3n-3 and 18:2n-6 as substrates. Apparent K m values were estimated with 18:3n-3 (2.5-55 M) in triplicates from the linear rate of biosynthesis of cis-trans conjugated products. We used GraFit (Erithacus Software) for non-linear estimates of K m values, based on seven to eight data points, which were fitted to the Michaelis-Menten equation. Specific activities were determined with 100 M 18:3n-3.
13R-MnLOX forms 11S-HPODE and 11R-HPOTrE as intermediates during the linear phase of oxidation and the 13R-hydroperoxides as main end products (26). The products formed by the mutants were therefore analyzed at two time points: when the UV absorbance at 235 or 237 nm had ceased to increase (apparent end point; UV absorption, 1.5-2 absorbance units) and at the middle of the linear increase in UV absorbance (midpoint; UV absorption, 0.5-1 absorbance unit).
LC-MS/MS Analysis-Reversed phase HPLC was performed with a Surveyor MS pump (ThermoFisher) and an octadecyl silica column (5 m; 2 ϫ 150 mm; Phenomenex; equipped with 2 ϫ 4-mm C 18 guard cartridge), which was eluted at 0.3 ml/min with 750:250:0.1 or 800:200:0.1 methanol/water/acetic acid. The effluent was subject to electrospray ionization in a linear ion trap mass spectrometer (LTQ, ThermoFisher) as described (40). The heated transfer capillary was set at 315°C, the isolation width was set at 1.5 or 5 (for unstable hydroperoxides; Refs. 44 and 45), and the collision energy was set at 35%. A trihydroxy fatty acid (prostaglandin F 1␣ ; 10 ng/min) was infused for tuning. Hydroperoxides were occasionally reduced to alcohols by treatment with triphenylphosphine (1-10 g) before LC-MS/MS analysis. Normal phase HPLC was performed as described (30).
Miscellaneous Methods-SDS-PAGE was performed as described (11). Protein bands were excised and treated with trypsin as described (47). Peptides were analyzed by MALDI-TOF (Bruker Ultraflex TOF/TOF), and the Mascot program was used for analysis. N-terminal sequencing of mini 13R-MnLOX was performed at the Protein Analysis Center, Karolinska Institute, Stockholm, Sweden.

Sequencing and Expression of Mini 13R-MnLOX-Secreted
13R-MnLOX of 602 amino acids was transformed during storage to a smaller and less glycosylated protein of ϳ67 kDa that retained the enzyme activity and the catalytic parameters of native 13R-MnLOX (11).
SDS-PAGE purification, trypsin digestion, and MALDI-TOF analysis of peptides showed that this mini form of 13R-MnLOX contained at least 562 C-terminal amino acids (supplemental Fig. S1). Sequential N-terminal sequencing suggested that mini 13R-MnLOX was mainly formed by cleavage between Thr 37 and Thr 38 as judged from five amino acids in correct sequence order from this point (data not shown). This corresponded to a protein of 580 amino acids with a calculated molecular mass of 63.8 kDa.
We expressed 13R-MnLOX with 580, 570, and 555 amino acids and confirmed protein expression by SDS-PAGE. The 580-amino acid construct retained catalytic activity, whereas the two shorter constructs were inactive. We confirmed the structure by MALDI-TOF analysis of tryptic peptide fragments.
The recombinant mini 13R-MnLOX of 580 residues oxidized 18:2n-6 with formation of 11-and 13-hydroperoxy metabolites as formed by native 13R-MnLOX. Unexpectedly, recombinant mini 13R-MnLOX was glycosylated and in some experiments to the same extent as recombinant 13R-MnLOX_602; SDS-PAGE thus revealed a broad protein band due to heterogenous glycosylation ( Fig. 2A). The recombinant mini form with 580 amino acids was used in this report unless otherwise stated as its catalytic activities appeared to be identical to those of 13R-MnLOX (see Ref. 11), and this recombinant minienzyme is referred to as 13R-MnLOX for simplicity.
The effects of replacements at the Leu 336 and Phe 337 positions of 13R-MnLOX on catalysis are summarized in Tables 1  and 2. Except for L336V, the catalytic efficiency of the mutants was markedly reduced.
Replacement of Leu 336 of 13R-MnLOX with Small Hydrophobic Residues-The kinetic lag phase appeared to be increased with some mutants (e.g. L336A⅐13R-MnLOX), but K m remained below 10 M 18:3n-3 for all the Leu 336 mutants albeit significantly larger than the K m for 13R-MnLOX (Table 2 and supplemental Fig. S3).
Replacement of Leu 336 and Gly 332 with Larger Hydrophobic Residues-Replacement of Leu 336 with Phe had relatively small effects on the regiospecific oxidation of 18:2n-6 and 18:3n-3 as 13R-hydroperoxy metabolites remained the main products (86 -90%; Table 1). This is also in analogy with G332A, which only slightly increased the oxygenation at C-9 (38). K m for 18:3n-3 remained in the low M range for L336F (Table 2). For comparison, K m for G332A⅐13R-MnLOX was determined to be 2.5 Ϯ 0.8 M.
L336F⅐13R-MnLOX oxidized 18:2n-6 to products with UV absorbance at 235 nm as shown in comparison with L336V⅐13R-MnLOX in Fig. 4A. After a prolonged kinetic lag phase, the absorption reached a maximum and then began to decline. This phenomenon was augmented with 18:3n-3 as a substrate (Fig. 4B). As the absorbance at 237 nm declined, there was an increase in UV absorbance at 280 nm (Fig. 4B, inset) apparently due to dehydration of 13R-HPOTrE. These events were unchanged in buffer saturated with oxygen (Fig. 4B); the lag phase appeared to be shortened in oxygen-saturated buffer, but this was not further investigated.
G332A⅐13R-MnLOX has previously been found to enhance the dehydration of 13R-HPOTrE (38). A notable difference a Relative data from analysis of end products from incubation with 13R-MnLOX and its mutants with 100 M substrate in 0.1 M NaBO 3 (pH 9.0). The relative amounts of products were estimated by selective ion chromatograms and integration under the area of the peaks of the ion chromatograms. The incubation was followed until the increase in UV absorbance ceased. L336F⅐13R-MnLOX was assayed during the linear phase of 18:3n-3 oxidation due to its hydroperoxide isomerase activity. Specific activities of 13R-MnLOX (apparent k cat ϳ720 min Ϫ1 ; 100%) and the mutants were determined after partial enzyme purification by hydrophobic interaction chromatography and gel filtration, and these numbers provide an index of catalytic turnover. L336 designates recombinant 13R-MnLOX. b These enzymes oxidized [11S-2 H]18:2n-6 with a kinetic isotope effect (k H /k D ) of ϳ32 (L336V) and ϳ23 (L336A; see Ref. 33) and with apparent loss of the deuterium label of 13-HODE as judged from LC-MS analysis. The k H /k D of L336A was not further investigated due to low specific activity. c Incubations with 18:3n-3 generated significant amounts of 11R-HPOTrE as an end product for L336V (ϳ4%), L336G (ϳ9%), and L336F (ϳ5%), but the table shows the percentage of oxidation at C-9 and C-13.  from L336F was that the rate of dehydration of 13R-HPOTrE by G332A was reduced in buffer saturated with oxygen (Fig. 4C), whereas this reaction of L336F was largely unaffected (Fig. 4B).

TABLE 2 Apparent K m values for oxidation of ␣-linolenic acid to cis-trans conjugated hydroperoxides with UV absorbance at 237 nm by 13R-MnLOX and its Leu 336 and Phe 337 mutants
To reduce the influence of the n-9 and n-3 double bonds on oxygenation, we also examined 18:1n-6 as a substrate. 13R-MnLOX oxidizes 18:1n-6 at C-11 and C-13 in a ratio of ϳ4:1. As shown in Fig. 5A, G332A changed this ratio to ϳ4:3. L336F also augmented the relative oxygenation at C-13 to the same extent (Fig. 5B); the inset shows for comparison that L336V only slightly affected the oxidation of 18:1n-6. We conclude that the main effect of L336F was an augmented transformation of 13R-HPOTrE to 13-KOTrE and increased oxidation at C-13 of 18:1n-6 in analogy with G332A.

DISCUSSION
We report two main observations. First, F337I⅐13R-MnLOX changed the absolute configuration of the main 13-hydroperoxy metabolites of 18:2n-6 and 18:3n-3 from ϳ100% R to ϳ70% S but did not alter abstraction of the pro-S hydrogen at C-11 of 18:2n-6. The hydrogen abstraction and oxygenation was thus altered from exclusively suprafacial to predominantly antarafacial, which mimics the antarafacial oxygenation mechanism of sLOX-1 and other FeLOXs. Second, replacement of Leu 336 with smaller or larger hydrophobic residues changed the regioand stereospecific oxygenation of C 18 fatty acids in a substratedependent way, the disposition of molecular oxygen to the catalytic center, and the hydroperoxide isomerase activity. An overview of these effects is shown in Fig. 6. All mutants except L336V showed markedly reduced specific activities (Table 1).
Site-directed mutagenesis of the active site of LOX can change the oxygen insertion in three principal ways, whereas the hydrogen abstraction remains invariant. First, replacements with small or large hydrophobic residues can shift the position of hydrogen abstraction at C-3 and oxygenation at C-5 of one 1Z,4Z-pentadiene of the substrate to these positions of the overlapping, vicinal 1Z,4Z-pentadiene as first described for 15-LOX by Sloane et al. (20). This finding was recently extended to a triad model for substrate positioning (48). By a similar mechanism, swapping Gly and Ala at one specific position of many LOXs changed the stereochemistry of the main product by oxygen insertion at the opposite ends of the 1Z,4Zpentadiene (34). The substrate alignment in the active site is changed, but antarafacial hydrogen abstraction and oxygen insertion are retained (34). Second, the substrate can bind to the active site in different orientations. At acidic pH, the uncharged carboxyl group of 18:2n-6 can enter the active site of sLOX-1 for oxidation in reverse head-tail orientation (49). In analogy,  replacement of a hydrophobic amino acid at the bottom of the substrate channel of cucumber 13-LOX unmasked a positively charged residue and shifted oxygenation from C-13 to C-9 (50). This mechanism was recently highlighted by conversion of human 5-LOX to 15-LOX by replacement of Ser 663 with an Asp residue, a mimic of phosphorylation (18). The antarafacial relationship between hydrogen abstraction and oxygenation is not altered by the substrate in the opposite head-tail orientation. Third, radical leakage may be increased by structural perturbations of the active site in analogy with the radical leakage of sLOX-3. The latter releases the 1Z,4Z-pentadiene radical so that an almost racemic mixture of hydroperoxides is generated.
Catalytic Convergence-F337I⅐13R-MnLOX describes an unprecedented modification of the oxidation process. 13R-MnLOX and its 9S-MnLOX homologue of M. salvinii are characterized by suprafacial hydrogen abstraction and oxygen insertion (26). 3 The F337I mutant retained this hydrogen abstraction but changed oxygen insertion from suprafacial to mainly antarafacial. Binding of 18:2n-6 and 18:3n-3 in the reverse orientation would be expected to give 9R-hydroperoxy metabolites as the main products, but this was not the case. The F337I mutant likely altered the oxygenation at C-13 from R to S by allowing oxygen access to the other face of the 12Z double bond. A comparison of the active sites of 13R-MnLOX and sLOX-1 based on modeling is shown in Fig. 7. It seems likely that Phe 337 of 13R-MnLOX shields the face of the 12Z double bond, which is subject to 13S oxygenation by sLOX-1, whereas Ile 547 of sLOX-1 shields this double bond for oxygen insertion at C-13 with R configuration. Schematic views of these steric effects of Phe 337 and Ile 547 are presented in Fig. 7, C and D.
It will be of interest to determine whether replacement of Ile or Val at this position could influence the steric oxidation of other LOXs in analogy with 13R-MnLOX. 8R-LOX contains an Ile residue in this position that has been replaced by Trp and Ala residues. I433A⅐8R-LOX retained catalytic activity and slightly changed the stereospecific oxidation ratios in the direction of suprafacial hydrogen abstraction and oxygen insertion (8R/8S, 95:5; 12R/12S, 16:84), whereas I433W⅐8R-LOX was inactive (16).
Unshielding of the 9Z Double Bond for Oxygen Insertion-We evaluated the effects of substitutions of Leu 336 in 13R-MnLOX with 18:3n-3 and 18:2n-6 as substrates. This Leu residue is conserved in all LOXs and assumed to position the substrate near the metal for oxygenation (16,33). 18:2n-6 and 18:3n-3 were oxidized by all Leu 336 mutants at C-13 as the major products FIGURE 6. Overview of the catalytic effects of hydrophobic amino acid replacements near the metal center of 13R-MnLOX. L336F and G332A mainly increased the hydroperoxide isomerase activity. Replacement of Leu 336 partly changed the oxidation of 18:3n-3 and 18:2n-6 from C-13 with retained R stereospecificity (Ͼ95%) toward C-9 with formation of 9S-and 9R-hydroperoxides. F337I changed the configuration of the 13-hydroperoxy metabolites of 18:3n-3 and 18:2n-6 from ϳ100% R to ϳ70% S. (Ն74%) and consistently with Ͼ98% R configuration with only one exception, L336A (Table 1). Mutations at the Leu 336 position seemed to influence the oxygenation at C-9. The absolute configuration at C-9 was mainly 9R with 18:3n-3 as a substrate. These changes from 9S to 9R configuration were less pronounced with 18:2n-6. Substituting Leu with Val thus had smaller effects on 18:2n-6 (70% 9S) than on 18:3n-3 (17% 9S). These results support that Leu 336 restrains the substrate for stereospecific oxygen insertion in the 9S position, which can be perturbed by replacements.
The replacements of 13R-MnLOX at the Leu 336 position had little influence on K m for 18:3n-3 in comparison with the markedly reduced lipoxygenation (Tables 1 and 2). For comparison of homologous replacements, L546A⅐sLOX-1 lost over 98% of total catalytic activity and increased the relative oxygenation at C-9 of 18:2n-6 to 13%. The latter was attributed to reverse substrate orientation in the active site (33). The L432A and L432F mutants of 8R-LOX also lost most of the enzyme activity (Ն95%), but they increased the oxygenation of 20:4n-6 at C-12 from 2 to 10% and to 34%, respectively, with retention of the R configuration at C-8; L432V and L432I did not affect the regiospecificity (16). We conclude that hydrophobic replacements of Leu 336 changed the relative oxygenation at C-13 and C-9 of 18:2n-6 and 18:3n-3 in analogy with L546A⅐sLOX and in analogy with oxidation of 20:4n-6 by mutants of coral 8R-LOX.
Oxygen Access-As previously reported, G332A⅐13R-MnLOX augmented the hydroperoxide isomerase activity and the formation of epoxy alcohols from 13R-HPOTrE (38). We found that L336F⅐13R-MnLOX also possessed prominent hydroperoxide isomerase activity. In contrast to G332A⅐13R-MnLOX, the hydroperoxide isomerase activity of L336F⅐13R-MnLOX could not be reduced in oxygen-saturated buffer. The decrease of the catalytic space with G332A or L336F replacement likely restricts the disposition of molecular oxygen in the active site. Interestingly, Zheng and Brash (51) found that A451G⅐eLOX-3 allowed molecular oxygen access to the enzyme center, which reduced the hydroperoxide isomerase activity and revealed LOX activity.
Summary-We report that F337I⅐13R-MnLOX changed the stereospecificity and formed the 13S-hydroperoxides of 18:2n-6 and 18:3n-3 as main metabolites. This occurred by antarafacial hydrogen abstraction and oxygenation in analogy with sLOX-1. The Leu 336 residue was important for substratedependent regio-and stereospecific oxygenation at C-9, oxygen access to the catalytic center, and consequently the hydroperoxide isomerase activity (51).