Purification, crystallization and preliminary X-ray diffraction analysis of pathogen-inducible oxygenase (PIOX) from <em>Oryza sativa</em>

1. Introduction

Several enzyme systems are activated as a result of pathogen attack on plants, including phospholipase, lipoxygenase, allene oxide synthase and divinyl ether synthase. These enzymes produce oxylipins, bioactive lipid mediators from 18–22 carbon polyunsaturated fatty-acid (PUFA) substrates that play key roles in a plant’s defense reaction against pathogens (Hammond-Kosack &amp; Jones, 1996 ▶; Creelman &amp; Mullet, 1997 ▶). In 1998, Sanz and colleagues identified a 75 kDa oxygenase, termed PIOX (pathogen-induced oxygenase), that was induced upon the infection of tobacco leaves with the bacterium Pseudomonas syringae (Sanz et al., 1998 ▶). Subsequent analysis of the biological function revealed that PIOX catalyzed a non-lipoxygenase type of PUFA oxidation. Further database searches have identified PIOX in other plants, including other tobacco species (Nicotiana attenuata; Hermsmeier et al., 2001 ▶), mustard seed (Arabidopsis thaliana; Sanz et al., 1998 ▶; Liu et al., 2004 ▶), rice (Oryza sativa; Koeduka et al., 2002 ▶), pepper (Capsicum annuum; Kim et al., 2002 ▶) and tomato (Lycopersicon esculentum; Tirajoh et al., 2005 ▶). PIOX belongs to a larger family of heme-containing fatty-acid α-­dioxygenases that includes the mammalian cyclooxygenases (COX-­1 and COX-2; Hamberg et al., 2002 ▶), linoleate diol synthase (LDS) from the fungus Gaeumannomyces graminis (Brodowsky et al., 1992 ▶; Su &amp; Oliw, 1996 ▶) and a protein of unknown function encoded by the open reading frame OrfX from P. alcaligenes (Gerritse et al., 1998 ▶). While it was previously thought that plants do not possess a homolog equivalent to the COX enzymes found in mammals, analysis of the sequence of PIOX and other members of the fatty-acid α-­dioxygenase family indicate potential structural and catalytic similarities (Hamberg et al., 2002 ▶).

PIOX converts linoleic acid, α-linolenic acid, arachidonic acid and other PUFAs to their corresponding 2R-hydroperoxides via stereoselective dioxygenation (Sanz et al., 1998 ▶; Hamberg et al., 1999 ▶; Koeduka et al., 2002 ▶). The resulting 2R-hydroperoxides undergo spontaneous decarboxylation to shorter aldehydes and fatty acids (Hamberg et al., 1999 ▶; Koeduka et al., 2002 ▶). The stereoselective dioxygenation catalyzed by PIOX differs from that observed for COX and other plant lipoxygenases in that an unsaturated region of the PUFA substrate is not targeted for oxygen addition (Hamberg et al., 1999 ▶; Liu et al., 2004 ▶). Many residues crucial for activity in mammalian COX are conserved in PIOX. These residues include the distal and proximal histidine residues involved in heme binding and a catalytic tyrosine which is responsible for the initiation of catalysis (Sanz et al., 1998 ▶; Koeduka et al., 2002 ▶; Garavito et al., 2002 ▶; Liu et al., 2004 ▶). The complete conservation of these residues within the fatty-acid α-dioxygenase family suggests that the dioxygenations carried out by PIOX are initiated by the conserved tyrosine and that all of the members bind a heme group (Hamberg et al., 2002 ▶).

While there are global differences between PIOX and COX enzymes with respect to their potential overall structural architecture, substrate specificities and signaling pathways, the mechanisms by which these enzymes stereoselectively oxygenate PUFAs are strikingly similar. Therefore, elucidation of the structure of PIOX and the subsequent mechanism controlling catalysis will enhance the existing knowledge of how PUFAs are stereoselectively oxygenated and how this oxygenation is controlled. This information will also provide useful insights into other systems besides COX enzymes which discriminate between the saturation and chain length of a PUFA substrate, such as lipoxygenases (Brash, 1999 ▶; Kuhn, 2000 ▶; Schneider &amp; Brash, 2002 ▶; Coffa &amp; Brash, 2004 ▶), cytosolic phospholipase A₂ (Clark et al., 1995 ▶) and lysophospholipid acyl transferases (Reitz et al., 1969 ▶). We report here the purification, solubilization, crystallization and data collection of rice PIOX.

2. Methods

2.1. Protein purification

The expression vector containing the cDNA encoding for PIOX from rice (O. sativa) was obtained from Dr Kenji Matsui of the Department of Biological Chemistry, Faculty of Agriculture, Yamaguchi University, Japan. PIOX was expressed and purified according to the procedure reported previously (Koeduka et al., 2002 ▶) with significant modifications to the purification protocol which allowed the isolation of larger quantities of homogeneous protein required for subsequent crystallization trials. Briefly, decyl maltoside (C₁₀M) was added to the cell lysate from a 6 l growth to a final concentration of 0.7%(w/v) and stirred for 60 min at 277 K to solubilize the protein. The detergent-solubilized mixture was then ultracentrifuged for 45 min at 277 K, followed by passing the clarified supernatant through a 0.22 µm filter. The filtered supernatant was applied onto a Talon immobilized metal-affinity column (10 ml bed volume) equilibrated with buffer A [50 mM sodium phosphate pH 7.0, 300 mM NaCl, 0.7%(w/v) C₁₀M]. PIOX was eluted from the column in buffer A containing 100 mM EDTA, which strips the metal (and subsequent bound protein) from the column support matrix. Fractions containing O₂-uptake activity (measured as described in Koeduka et al., 2002 ▶) were pooled and dialyzed overnight at 277 K against 50 mM Bis-Tris pH 7.0, 300 mM NaCl and 0.7%(w/v) C₁₀M to remove cobalt and EDTA and to initiate detergent exchange. At this point, either 0.1%(w/v) C₁₀M or 0.5%(w/v) β-octylglucoside (β-OG) can be used as the crystallization detergent. Both have been used successfully to generate diffraction-quality crystals. The dialyzed protein was then concentrated to ∼2 ml using a Millipore Ultrafree-15 spin concentrator with a 50 kDa nominal molecular-weight cutoff and loaded onto a HiPrep Sephacryl S-300 HR 16/60 gel-filtration column (16 mm diameter and 120 ml bed volume) equilibrated with 50 mM Bis-Tris pH 7.0, 150 mM NaCl and either 0.1%(w/v) C₁₀M or 0.5%(w/v) β-OG to finalize detergent exchange and remove aggregates.

2.2. Crystallization

Peak fractions were pooled based on O₂-uptake activity and sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then concentrated to 5–10 mg ml ^{for crystallization trials. Initial crystallization screening was carried out using the microbatch-under-oil method in conjunction with the High Throughput Crystallization Screening Robot located within the Hauptman–Woodward Institute (Luft} et al., 2003 ▶; http://www.hwi.buffalo.edu/). To supplement the search for initial crystallization leads, commercial screens from Hampton Research (Crystal Screens 1 and 2, MembFac, PEG/Ion, PEG/LiCl and MPD) were set up using the sitting-drop vapor-diffusion method. For optimization, the sitting-drop vapor-diffusion method was employed with experimental drop volumes of 5 µl and reservoir solution volumes of 1 ml. Drops were prepared with a 1:1(v:v) ratio of protein:precipitant solution. All crystallizations were carried out and stored at 294 K.

2.3. Data collection and analysis

Diffraction data were collected on a Rigaku R-AXIS IV image-plate detector mounted on an RU-H3RHB 18 kW rotating-anode X-­ray generator (Cu target) equipped with Osmic Max-Flux confocal multi-layer mirrors and an X-Stream cryogenic system. Crystals were removed from sitting drops and directly flash-frozen in the nitrogen stream (100 K). Individual 1° oscillation frames were collected over a range of 360°. Data were initially processed and scaled using DENZO and SCALEPACK (Otwinowski &amp; Minor, 1997 ▶) and subsequently with MOSFLM (Leslie, 1992 ▶) combined with SCALA from the CCP4 suite of programs (Collaborative Computational Project, Number 4, 1994 ▶). Self-rotation function and Matthews coefficient calculations were performed using MOLREP (Vagin &amp; Teplyakov, 1997 ▶) and Cell Content Analysis in CCP4. Data-collection statistics are given in Table 1 ▶.

3. Results and discussion

Purified PIOX was found to be homogeneous as analyzed by SDS–PAGE and size-exclusion chromatography (data not shown). Over 50 hits were identified in the initial crystallization screening experiments carried out using the HWI crystallization robot. Analysis of the hits revealed common features: (i) low-molecular-weight polyethylene glycol (PEG 400 or PEG 1K) or 2-methyl-4-pentanediol (MPD) as the precipitant, (ii) a pH range of 8.5–10.5 and (iii) 50–150 mM of a monovalent salt (NaCl, KCl, LiCl). The conditions were further refined through the construction of standard grid screens using the above identified variables. The largest single crystals could be reproducibly grown in sitting drops by combining 2.5 µl protein solution (5–10 mg ml^{) with 2.5 µl of 45, 50 or 55%(}w/v) MPD supplemented with 100 mM Bicine pH 9.0, 100 mM KCl and either C₁₀M or β-OG [at final concentrations of 0.05 and 0.33%(w/v), respectively] as the crystallization detergent. These 5 µl drops were subsequently equilibrated over reservoir solutions consisting of 50–65%(w/v) MPD, 100 mM Bicine pH 9.0 and 50–150 mM KCl. PIOX crystals routinely grow to maximum dimensions of 0.2 × 0.2 × 0.05 mm over the course of four weeks (Fig. 1 ▶). The optimized crystallization conditions proved ideal for cryopreservation owing to the high concentration of MPD used as the precipitating agent. Hence, crystals could be easily removed from the sitting drops and directly flash-frozen in a stream of cold nitrogen gas.

Figure 1

Crystals of rice PIOX grown using C₁₀M as the crystallization detergent.

The cryocooled PIOX crystals diffracted weakly to 3.0 Å, but were sufficiently stable in the X-ray beam to allow a complete data set to be collected. The indexing and scaling routines of both DENZO/SCALEPACK and MOSFLM/SCALA identified the space group as P1, with unit-cell parameters a = 74.70, b = 90.42, c = 119.44 Å, α = 69.67, β = 73.64, γ = 88.40°. Unsuccessful attempts at processing the data in other crystal systems and space groups further validated this solution. Self-rotation functions indicate that there are likely to be four PIOX molecules in the asymmetric unit related by 222 noncrystallographic symmetry (Fig. 2 ▶), which is in agreement with the calculated values of 2.5 Å^{Da and 51% for the Matthews coefficient and solvent content, respectively. Initial attempts to solve the structure by molecular replacement using a monomer of ovine COX-­1 (PDB code} 1diy; Malkowski et al., 2000 ▶) as the search model proved unsuccessful. Both COX isoforms share ∼12% sequence identity with PIOX and to date are the most homologous search models available. Structure solution utilizing multiple isomorphous replacement and multiwavelength anomalous dispersion methods are currently being pursued.

Figure 2

Stereographic projection of the self-rotation function in spherical polar angles plotting θ/ϕ at χ = 180°. The three orthogonal peaks ∼90° apart indicate the presence of noncrystallographic 222 symmetry and four PIOX monomers in the asymmetric unit.