An ensemble of lipoxygenase structures reveals novel conformations of the Fe coordination sphere

Abstract

The regio‐ and stereo‐specific oxygenation of polyunsaturated fatty acids is catalyzed by lipoxygenases (LOX); both Fe and Mn forms of the enzyme have been described. Structural elements of the Fe and Mn coordination spheres and the helical catalytic domain in which the metal center resides are highly conserved. However, animal, plant, and microbial LOX each have distinct features. We report five crystal structures of a LOX from the fungal plant pathogen Fusarium graminearum. This LOX displays a novel amino terminal extension that provides a wrapping domain for dimerization. Moreover, this extension appears to interfere with the iron coordination sphere, as the typical LOX configuration is not observed at the catalytic metal when the enzyme is dimeric. Instead novel tetra‐, penta‐, and hexa‐coordinate Fe²⁺ ligations are apparent. In contrast, a monomeric structure indicates that with repositioning of the amino terminal segment, the enzyme can assume a productive conformation with the canonical Fe²⁺ coordination sphere.

Short abstract

PDB Code(s): 6NS2, 6NS3, 6NS4, 6NS5 and 6NS6;

Abbreviations

LOX

lipoxygenase

PUFA

polyunsaturated fatty acid

Introduction

Lipoxygenases (LOX) are generally iron enzymes that catalyze the dioxygenation of polyunsaturated fatty acids. Crystal structures of plant,1, 2 animal,3, 4 bacterial,5, 6 and fungal LOX7 have all been described, with that of the fungal LOX representing the first description of a Mn‐LOX. Plant (~900 amino acids) and animal (~700 amino acids) LOX are composed of two domains: an amino‐terminal β‐sheet domain (~125 amino acids) followed by a much larger α‐helical domain. Three invariant His side chains, a fourth His or an Asn, and the free carboxy terminus position the catalytic iron in the active site, located in the larger helical domain. For the animal LOX this site is markedly U‐shaped and accommodates the substrate arachidonic acid, which contains four cis double bonds that form three pentadiene moieties. In contrast, in the plant enzymes the active site is more elongated or boomerang‐shaped. The primary substrates for the plant enzymes are linoleic and α‐linolenic acids, with one and two pentadienes, respectively. The central hydrophobic core of the active site, where a fatty acid pentadiene must be positioned for catalysis, is highly conserved. The metal cofactor mediates H abstraction from the central carbon of the susceptible pentadiene moiety and the resulting free radical reacts with molecular oxygen to yield a hydroperoxy product.

The products of LOX in both plant and animal enzymes are potent lipid mediators that allow the organisms to respond to stress or threats. For example, jasmonic acid, a plant and fungal hormone that functions in growth and the stress response, is synthesized from linolenic acid in a LOX pathway.8, 9 In mammals, LOXs produce pro‐inflammatory lipid mediators such as leukotrienes.10 The functions of LOXs in microbes are less well understood, but roles in host–pathogen interactions have been suggested.9, 11, 12, 13 While the central framework of the catalytic domain structure is conserved throughout the enzyme family (17), the bacterial and fungal enzymes lack the amino‐terminal β‐domain. They may instead have additional amino terminal helices associated with the helical catalytic domain,5 a distorted β‐barrel6 or only encompass the catalytic domain.7, 14

The fungus Fusarium graminearum (Fg), which infects wheat and barley, is an economically important plant pathogen.15 The Fg genome encodes for a single LOX protein, a 13S‐LOX16 designated here as FgLOX. A 13S‐LOX with linoleic and α‐linolenic acids as substrates has previously been characterized from Fusarium oxysporum (~70% sequence identity),17 and a related gene is conserved in other Fusarium family members.

We report an ensemble of structures of FgLOX, which represent monomeric and dimeric conformations, derived from a total of five crystal forms. In the dimeric enzyme, an extended amino‐terminus serves as a wrapping domain; the wrapping domain is not discernable in the monomeric forms. In addition to the difference in oligomerization, these structures reveal distinct conformations at the coordination sphere.

Results

A total of five crystal forms of FgLOX were obtained (Table 1). As four of the crystal forms had two monomers in the asymmetric unit, a total of nine crystallographically independent structures were determined representing both monomeric and dimeric enzymes.

Table 1.

Data Collection, Phasing, and Refinement Statistics

	P212121	I222	C2	C2_2	P21
PDB Code	6NS2	6NS3	6NS4	6NS5	6NS6
Wavelength (Å)	1.7413	1.38079	1.38079	0.97918	0.97918
Resolution (Å)	2.79	2.84	2.40	2.79	3.30
Temperature (K)	100	100	100	100	100
Space group	P212121	I222	C2	C2	P21
Cell dimensions
a (Å)	90.489	93.931	116.96	123.37	73.01
b (Å)	95.060	95.537	121.12	113.67	94.77
c (Å)	189.490	186.249	102.12	99.44	105.08
β (°)	–	–	95.13	90.25	106.16
Molecules per asymmetric unit	2	1	2	2	2
No. of unique reflections	40 654	19 682	55 511	33 864	20 760
Rpim a , b (%)	7.2 (39.5)	6.2 (30.0)	8.3 (35.9)	7.6 (42.8)	5.7 (42.5)
Completeness (%)	98.5 (97.9)	97.5 (78.7)	99.7 (97.8)	100 (100)	99.5 (99.8)
Redundancies	6.2 (6.2)	7.0 (5.2)	3.7 (3.3)	3.8 (3.6)	3.0 (2.9)
I/σ(I)	8.2 (2.2)	12.2 (2.2)	9.6 (1.7)	8.6 (1.9)	9.7 (1.9)
CC(1/2), last shell	0.828	0.911	0.710	0.581	0.763
Refinement statistics
Resolution range (Å)	47.58–2.79	47.81–2.84	40–2.40	64.08–2.79	40.00–3.30
No. of reflections	38 588	19 085	54 372	32 762	20 082
σ cutoff	None	None	None	None	None
R /R free c (%)	20.55/24.94	23.41/27.24	19.03/22.09	17.48/23.72	19.38/24.12
Number of refined atoms
Protein	10 103	5 040	10 791	10 596	10 534
Hetero atoms	2	1	30	2	2
Water	53	8	300	10	2
Average B‐factors (Å2)
Protein	62.9	88.5	34.5	47.7	120.0
Water	35.4	49.5	24.4	26.4	43.0
Fe2+	44.5	65.6	28.4	31.8	91.6
RMS deviations
Bonds (Å)	0.004	0.002	0.005	0.012	0.003
Angles (°)	1.314	1.206	1.270	1.528	1.182
Ramachandran plot (%)
Favored	95.2	92.5	97.0	95.7	91.9
Disallowed	0.2	0.2	0.2	0.2	0.3

^aValues in parentheses are for the highest‐resolution shell.

^b R _pim is a redundancy‐independent measure of the quality of intensity measurements. R _pim = ∑ _hkl (1/[n − 1])^1/2 ∑ _i |I _hkl,i − <I _hkl >|/∑ _hkl ∑ _i I _hkl,i, where I _hkl,i is the scaled intensity of the measurement of reflection h, k, l, <I _hkl > is the average intensity for that reflection, and n is the redundancy.

^c R = ∑|| F _o | − | F _c ||/∑|F _o|, where F _o and F _c are the observed and calculated structure factors amplitudes. R _free is calculated using 5.0, 3.0, 2.1, 3.2, 3.2% of reflections omitted from the refinement for the P2₁2₁2₁, I222, C2, C2_2, and P2₁ structures, respectively.

Overall structure of the FgLOX dimer

The individual FgLOX monomers (Fig. 1) are composed of 29 α‐helices. All but two of these segments are gathered to form a flattened, triangular bundle. The helices diverge from the apex of the triangle, and 6 β‐strands wrap one side of the base of this bundle. Centered in the triangular, wedge‐like shape is the catalytic iron positioned by His‐407, ‐412 and ‐596. These side chains are provided by two long helices (386–425, 577–608) that span the height of the triangle. The side of the wedge‐like assembly that is strapped and fortified by β‐strands appears to provide the bulk of the support for the catalytic center and protects it from access by solvent. The Fe²⁺ sits closer to the opposite face, and in the dimeric structure appears solvent accessible.

Figure 1.

Dimeric FgLOX. (A) Cartoon rendering of a monomer of the FgLOX dimer (N to C, blue to red). The star indicates the first visible amino acid observed in the monomeric structures. The Fe (rust sphere) is positioned by His‐407, ‐412, ‐596. (B) Cartoon rendering with monomers green and cyan. The two amino terminal helices (blue) from one monomer flank the two C‐terminal helices (red) of the partner protein. The surface rendering has been sliced so that the large central depression, a consequence of little contact between the monomer bundles in the absence of the wrapping domain (blue helices in A) is visible. The Fe²⁺ is accessible via the central depression.

An ~100 amino acid amino terminal extension in FgLOX breaks away from the helical bundle. In the dimeric structure, it plays a fundamental role as a wrapping domain, burying a total of ~2200 Ų of surface area with its embrace of the monomer mate. This area is roughly 50% of the total buried surface area (~4500 Ų) of the monomer at the dimer interface. Much of the dimer interface is helical, and the first two short helical segments of the structure flank the two C‐terminal helices of the monomer mate such that each helical bundle in the dimer contains segments from two polypeptide chains. Because the dimer interface relies extensively on amino terminal segments which otherwise have no contact with the helical bundle of the same polypeptide, there is a large gap between the two helical bundles of the dimer. The consequence of this organization is a large depression which allows access to the active sites that lie opposed at the center of the dimer; that is, dimerization does not block access to the catalytic sites. Yet dimerization appears to significantly impact the iron coordination sphere, as in this form the main‐chain carboxyl of the C‐terminus is not visible in the electron density of the Fe²⁺ coordination sphere.

Monomeric FgLOX

A monomeric form of FgLOX was observed in two crystal packings. Analysis of the crystal packing contacts of these forms with the Protein, Interfaces, Structures, and Assemblies server (PDBePISA18) yields maximum dimer interfaces of 649 Ų or 763 Ų, areas not likely sufficient to promote dimer formation in solution. In this structure, 100 amino acids from the N‐terminus are not visible in the electron density map, and the first visible amino acid (Gly‐100) is positioned quite remotely from the same segment in the dimeric structure, some 30 Å away from where it is located in that model. In the monomeric crystal structures, the C‐terminal main chain carboxyl and Asn‐600 complete the iron coordination sphere.

The iron coordination sphere

Three invariant His, an Asn or His, and the main chain carboxyl of a highly conserved C‐terminal Ile position the active site Fe²⁺ in all LOX structures to date. A catalytic water molecule completes the octahedral sphere. This precise arrangement of metal ligands is observed in the monomeric I222 crystal form of FgLOX, depicted in Figure 2. There are, however, striking differences in the metal coordination spheres among the five crystal structures of FgLox described here. While both monomeric structures (I222 and P2₁2₁2₁) conform to the usual arrangement with His‐407, His‐412, His‐596, Asn‐600, Ile‐745, and a water molecule comprising the distorted octahedral sphere, density for a water molecule is not visible in one of the proteins in the P2₁2₁2₁ asymmetric unit. Moreover, there is a significant change in the position of Asn‐600, with the Fe–O distance varying by 1 Å in the I222 versus P2₁2₁2₁ structures (Table 2). The dimeric structures all lack electron density for the last seven amino acids, thus the terminal Ile‐745 is no longer coordinated to Fe via the main chain carboxyl. Moreover, each of the three crystal forms, obtained from similar crystallization conditions (22%–25% PEG 3350, 0.3 M NH₄Acetate), displays unique coordination geometry. In one the Fe²⁺ is hexacoordinated with His‐407, His‐412, and His‐596 and three water molecules ligating the catalytic iron; the side chain of Asn‐600 is no longer a participant in metal ligation. Asn‐600 also does not coordinate to the iron in a second dimeric structure. However, this coordination sphere is only tetracoordinated with a water molecule as the fourth ligand. The third and final dimeric structure shows penta‐coordination with Asn‐600 and a water molecule as the additional ligands in addition to the invariant histidines. This array of observed coordination geometries suggests that alternative metal coordination might be exploited to “tune” enzyme activity. Conformational flexibility might come into play with specific substrates, for example, the ability of FgLOX and its F. oxysporum homolog to exhibit “hydroperoxidase”19, 20 activity with α‐linolenic acid as substrate—the initial 13‐hydroperoxide is rearranged to an 11‐hydroxy‐12,13‐epoxy derivative, a reaction that does not occur with the corresponding 13‐hydroperoxide of linoleic acid.17

Figure 2.

Observed Fe²⁺ coordination spheres. (A) Monomeric FgLOX (space group I222) displays the distorted octahedral coordination typical of lipoxygenases. (B, C, D) coordination spheres observed in dimeric FgLOX structures: hexa‐, tetra‐, and penta‐coordinate (Table 2, C2, C2_2, and P21, respectively).

Table 2.

Fe²⁺‐Ligand Distances Observed in the Different Crystal Forms of FgLOX

		P212121		C2		C2_2		P21
	I222	Aa	B	A	B	A	B	A	B
Fe…NE2(H407), Å	2.11	2.31	2.31	2.31	2.30	1.89	1.86	2.01	1.95
Fe…NE2(H412), Å	2.10	2.30	2.32	2.27	2.27	1.92	1.92	1.90	1.87
Fe…NE2(H596), Å	2.11	2.30	2.31	2.30	2.30	1.88	1.92	1.95	1.94
Fe…OD1(N600), Å	2.22	3.30	3.23	–	–	–	–	2.86	2.87
Fe…OXT(I745), Å	2.25	2.43	2.44	–	–	–	–	–	–
Fe…H2O, Å	2.16	2.17	–	2.00	2.01	2.02	2.06	1.96	2.23
				2.01	2.01
				2.48	2.02

^aValues for two independent molecules A and B are listed.

The protein context of the different Fe²⁺ coordination spheres

The absence of the C‐terminus as an iron ligand in dimeric FgLOX is likely a consequence of localized protein disorder, as no electron density is visible for the last seven amino acids of the enzyme, or for a segment of a proximal peptide region (556–668). The obvious question here is what promotes the disorder and consequently a complete coordination of the Fe²⁺ in the dimeric form. In addition to the breaks in density suggesting mobility of the C‐terminal segment and this proximal peptide region, there is a remodeling of a helical segment (678–688). Two turns of a helix in the dimer are unwound in the monomeric enzyme so that the polypeptide forms a β‐hairpin (Fig. 3). The hairpin sits where a short α‐helix of a dimer mate flanks the penultimate helix of a protomer. Interestingly, this same β‐hairpin structure is observed in the structure of Pseudomonas aeruginosa LOX,5 which is monomeric. However, the wrapping domain of the FgLOX dimer appears to prevent hairpin formation. Essentially, the dimeric structure is incompatible with the formation of a productive Fe²⁺‐coordination sphere of the monomer mate.

Figure 3.

Secondary structure differences in monomeric and dimeric FgLOX. (A) Ribbon drawing of amino acids 638–745 as observed in the monomeric structure (purple) and the corresponding region of the dimeric structures (green). The active site metal is an orange sphere. Amino acids 678–688 are helical in the dimer, and form a β‐hairpin in the monomer. The ribbons are superimposed on transparent cartoon rendering of the dimer (yellow and lavender). Note the hairpin is inconsistent with the placement of a helix of the monomer mate (pale yellow). (B) Surface rendering of one protomer of the dimer (white) with the remodeled region of the monomeric form (purple). In the dimeric protein, the Fe (indicated by the arrow) is accessible to solvent and the active site cavity deceptively large due to the disordering of the C‐terminal amino acids.

The active site

In addition to the canonical Fe²⁺ coordination sphere, the core active site of monomeric FgLOX displays other LOX hallmarks, which include an invariant Leu (Leu‐454) positioned to straddle the fatty acid pentadiene once lined up in the active site for hydrogen abstraction.21 The location of the reactive group of the substrate, as inferred from the one LOX structure in complex with substrate,22 is flanked by the catalytic iron and a pocket for the O₂ co‐substrate. As predicted,23 Ala‐450 is in this pocket to help direct the O₂ to the appropriate position on the pentadiene. What is not as obvious is how substrate gains entry into this site, as FgLOX lacks the gatekeeper helix (α2) described for both animal and plant structures21 and there is no open portal in the monomeric structure (e.g. Ref. 4, 24). However, the helix which provides the clamping Leu (Leu‐454), the “arched helix” typical of this enzyme family, has elevated temperature factors, suggesting a possible path of entry might be to slide by this helix. Such a means of entry has been proposed by Bradshaw and Gaffney from EPR experiments with Soybean LOX‐1.25 Alternatively, one might look to the dimeric configuration to suggest an entry portal. The dimeric enzyme structure reveals a very large cavity with direct access to the catalytic iron (Fig. 1). However, one must keep in mind that the terminal heptapeptide in this structure is disordered, and this cavity, which must accommodate the missing piece of peptide for a fully functional conformation, appears deceptively large (Fig. 3).

In contrast to ready identification of the active site core defined by the catalytic iron, O₂ pocket and “clamping” Leu, the amino acids which lie outside this core but are essential to confer substrate and product specificity are not obvious. In the monomeric structure, there is not a volume large enough to accommodate the substrate. Moreover, one cannot extrapolate defined determinants for pocket depth in the animal enzymes to the fungal enzyme (e.g., the Sloan determinant26) as the fungal counterpart amino acids deduced from a structure‐based sequence lineup are distal to the active site cavity.

Discussion

Brodhun et al.17 characterized a LOX from F. oxysporum (EXK38530) which is 70% identical in sequence to the FgLOX described here (ESU07624). That enzyme, the sequence of which also has an amino terminal extension that might serve as a wrapping domain, was observed to be dimeric in solution. Like the animal and plant LOXs, it is an Fe²⁺ enzyme that catalyzes the antarafacial oxygenation of substrate. Both the deduced dimeric structure and the presence of a conserved O₂ pocket for antarafacial oxygenation are revealed in the crystal structure of this homologous enzyme FgLOX.

As one can see from Supporting Information Figure S1, a similar wrapping domain is likely to be present in LOX's from other fungi that share ~70% sequence identity or more with FgLOX. However, it may also be found in enzymes with much less sequence identity. For example, while LOX from the fungi Botrytis cinera and Pleurotus ostreatus (~40% sequence identity with Fg LOX) lack this domain, the LOX from Zymoseptoria tritici, which is also only 40% identical in sequence to FgLOX, likely shares it (Supporting Information Fig. S1). It is interesting to point out that the enzymes which carry the amino terminal extension also conserve the sequence of the C‐terminal segment (FgLOX 678–688, Fig. 3) that remodels from helix to hairpin with the transition from dimer to monomer.

Two fungal LOX structures have previously been reported; both are Mn²⁺ rather than Fe²⁺ LOX. These LOX (5FX8 and 5FN0, ~56% pairwise identity; ~20% identity with FgLOX) also lack the amino terminal β‐barrel described in animal and plant enzymes. Two structures from cyanobacteria have also been reported, (5MED and 5EK8, pairwise identity ~18%, ~20% with FgLOX), one of which contains an amino terminal β‐barrel domain roughly the same size as that of the plant and animal enzymes (Fig. 4), but with distinct strand topology and positioning relative to the catalytic domain. A P. aeruginosa LOX (~23% identical in sequence to FgLOX) also lacks the β‐barrel, but like the Fg LOX has additional amino terminal helices. However, the placement of the amino terminal helices and the presence of the helical wrapping domain is distinct in FgLOX.

Figure 4.

Diversity in LOX structures. Cartoon representations of lipoxygenases from diverse organisms with the amino termini colored in blue, the conserved helical bundle in red.

Conclusion

The suite of LOX structures described herein reveals strikingly different conformations of the Fe²⁺ coordination sphere, and the differences in coordination geometry appear to be provoked to the presence of a wrapping dimerization domain. It is tempting to speculate that this oligomerization‐dependent metal coordination sphere is an important regulatory mechanism. Moreover, these views provide a structural context for understanding how Fe²⁺ ligation may change upon substrate binding27 or shift between penta‐ and hexa‐coordination,28 as previously described for Soybean LOX1.

Material and Methods

Protein expression and purification

BL21(DE3) cells (Novagen) were transformed with a pET‐28b plasmid harboring the FgLOX gene with a six‐histidine tag at the amino terminus. Cultures of 500 mL were grown in Terrific Broth plus 50 μg/mL kanamycin in 2.8 L flasks for 21 hours at 37°C, 250 rpm. The cultures were then cold shocked at 4° for 30 minutes and returned to the shaker/incubator at 21°C, 250 rpm for 21 hours. Cells were harvested by centrifugation and pellets frozen at −80°. Cell pellets were resuspended in Bugbuster (EMD Millipore, Burlington, MA) according to the manufactures instructions, with the addition of pepstatin, leupeptin, and DNase I. Cells were lysed by sonication and the lysates clarified by centrifugation (30 minutes, 46,000 g). The supernatant was applied to a Ni‐bound chelating agarose column (GE Healthcare, Chicago, IL) and the column was washed extensively with binding buffer (20 mM Tris pH 8, 500 mM NaCl). The protein was eluted with an imidazole gradient (0–200 mM in binding buffer). Further purification was performed on a Superdex‐200 column (GE Healthcare) in 20 mM Tris pH 8, 150 mM NaCl. Protein purity was monitor by SDS gel.

Protein crystallization

FgLOX crystallizes in a variety of crystal forms in similar conditions with PEG3350 as the precipitating agent. Five different crystal forms have been obtained.

P2₁2₁2₁

Plate‐like crystals of the enzyme were obtained using the hanging drop vapor‐diffusion method by mixing equal volumes of protein (4 mg/mL in 20 mM Tris pH 8.0, 150 mM NaCl) and the reservoir solution (10% PEG 3350, 0.05M proline, 0.1M imidazole acetate pH = 7.0) at 22°C. The crystals grew in approximately 1 week and belonged to the orthorhombic space group P2₁2₁2₁.

I222

Plate‐like crystals were obtained using the hanging drop vapor‐diffusion method by mixing equal volumes of protein (6 mg/mL in 20 mM Tris pH 8.0, 150 mM NaCl) and the reservoir solution (11% PEG 3350, 0.35M CaCl₂, 0.1M Tris pH 8.0) at 22°C. The crystals grew in couple of days and belonged to the orthorhombic space group I222.

C2

A cluster of plates was obtained using the hanging drop vapor‐diffusion method by mixing equal volumes of protein (12 mg/mL in 20 mM Tris pH 8.0, 150 mM NaCl) and the reservoir solution (25% PEG 3350, 0.3M ammonium acetate, 0.1M Hepes pH 7.46) at 22°C. The crystals belonged to the monoclinic space group C2.

Second C2 crystal form (C2_2)

Three‐dimensional crystals were obtained using the hanging drop vapor‐diffusion method by mixing equal volumes of protein (12 mg/mL in 20 mM Tris pH 8.0, 150 mM NaCl) and the reservoir solution (22% PEG 3350, 0.3M ammonium acetate, 0.1M Tris pH 8.2) at 22°C. The crystals grew in approximately 3 weeks, were highly irreproducible and belonged to the monoclinic space group C2.

P2₁

Small rod‐like crystals grew using the hanging drop vapor‐diffusion method by mixing equal volumes of protein (6 mg/mL in 20 mM Tris pH 8.0, 150 mM NaCl) and the reservoir solution (22% PEG 3350, 0.3M ammonium acetate, 0.1M Tris pH 8.46) at 22°C. The crystals appear in approximately a week and belonged to the monoclinic space group P2₁.

X‐ray data collection

Prior to data collection, suitable crystals were dipped for 30 seconds in a mother liquor solution with the addition of 10%–15% glycerol as a cryoprotectant. Diffraction data were collected at 100 K at the Advance Photon Source 24‐ID‐C beamline (P2₁2₁2₁) equipped with a Pilatus 6M detector and 24‐ID‐E beamline (C2_2 and P2₁) equipped with a Dectris Eiger 16M pixel detector. Data were also collected at the Center for Advanced Microstructures and Devices (Baton Rouge, Louisiana) with a Mar CCD detector (I222 and C2). The images were processed using the XDS program suite29 (P2₁2₁2₁, C2_2, and P2₁) or HKL200030 (I222 and C2) and scaled using the Scala or Scalepack program. Data collection and data processing statistics are given in Table 1.

Crystal structure determination

The molecular replacement procedure was applied to obtain a solution for the P2₁2₁2₁ crystal form using the program MOLREP.31 A monomer of the catalytic domain of 12‐lipoxygenase (PDB accession code 3RDE) was used as a search model. Two monomers were located in the asymmetric unit. The positioned MR model was refined using the maximum likelihood refinement in REFMAC31 with NCS restraints and one TLS group per each monomer. COOT32 was used for model building throughout the refinement. A monomer of the refined FgLox P2₁2₁2₁ crystal form was used to locate the molecular replacement solutions in all other crystal forms. Details of the refinement of every particular crystal structure follow.

P2₁2₁2₁ crystal form (6NS2)

The final model consists of protein residues 100–210, 223–745 for both molecules, 2 Fe⁺² ions and 53 water molecules.

I222 crystal form (6NS3)

The final model consists of protein residues 101–210, 223–745, one Fe⁺² ion and eight water molecules.

C2 crystal form (6NS4)

The final model consists of protein residues 8–134, 160–165, 171–211, 225–664, 669–739 (Molecule A), 8–134, 171–210, 225–664, 669–686, 694–739 (Molecule B), 2 Fe⁺² ions, 4 acetate anions, 2 glycerol molecules, and 300 water molecules. Alternate conformations have been built for protein residue Arg‐78, Glu‐102, Arg‐370, Glu‐415, and Glu‐579 (Chain A).

Second C2 crystal form (C2_2, 6NS5)

Detection and analysis of crystal twinning was performed in the CCP4 suite.31 The crystal was pseudomerohedrally twinned (the β angle is close to 90° (90.25°)) with the twin operator h,‐k,‐l and a twin fraction of 0.19. Refinement was done with REFMAC with automatic twin refinement. The final model consists of protein residues 9–134, 171–211, 225–664, 669–686, 694–736 (Molecule A), 8–134, 160–165, 171–210, 225–664, 671–684, 695–736 (Molecule B), 2 Fe⁺² ions and 20 water molecules. Alternate conformations have been built for protein residue Glu‐188 (Chain A).

P2₁ crystal form (6NS6)

The final model consists of protein residues 9–134, 171–211, 225–664, 670–685, 694–736 (Molecule A), 9–134, 160–165, 171–209, 226–664, 670–686, 694–736 (Molecule B), two Fe⁺² ions and two water molecules.

Conflict of Interests

The authors have no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Supporting information

Supplement Figure S1: Sequence alignment of Fg‐LOX with a selection of fungal lipoxygenases

A: The amino acid iron ligands are highlighted in yellow. The Ala/Gly switch, typically diagnostic of S‐LOX as Ala, R‐LOX as Gly is highlighted in red (at amino acid 450 in Fg‐LOX), thus predicting that all these sequences are S‐LOX, as established experimentally. The invariant Leucine clamp is in magenta (position 454 in Fg‐LOX), and the position corresponding to the “Sloane determinant” (that adjusts available space at the deep end of the fatty acid binding channel in LOX enzymes) is in purple. The alignments indicate that this determinant is not applicable to these fungal LOX enzymes. The boxed sequence (green) around amino acids 678–684 in Fg‐LOX marks the region that remodels in transition from dimer to monomer, the change in secondary structure apparently provoked by the absence of the wrapping domain. This is missing in the other fungal LOX, an argument for the lack of a similar dimerization/catalytic center link in those enzymes.

B: Phylogenetic tree and percent sequence identities for the same fungal LOX enzymes. The total number of amino acids in each LOX sequence is listed after the name (e.g. 745aa), and GenBank accession numbers.