Crystallization and preliminary X-ray investigation of lipoxygenase 1 from soybeans.

Soybean lipoxygenase 1 has been crystallized by the vapor diffusion method in 8-10% polyethylene glycol (average Mr 3400), 0.2 M sodium acetate buffer, pH 5.2-5.6, at a protein concentration of 6-12 mg/ml. Microseeding was employed to obtain growth of large single crystals. The crystals, which diffract to at least 2.2-A spacings, are monoclinic and of space group P2(1). Cell constants are: a = 95.4, b = 94.2, and c = 50.4 A and beta = 91.4 degrees. The calculated value of Vm (2.41 A3/Da) is consistent with the probable presence of one molecule of lipoxygenase/crystallographic asymmetric unit (Z = 2).

Lipoxygenases constitute a class of non-heme, non-sulfur iron oxygenases that act upon lipids containing a cis,ci.spentadiene moiety. Dioxygen is incorporated into the primary reaction products, hydroperoxides. These enzymes are widely distributed among plants, animals, and microorganisms. The presence of lipoxygenases in mammalian tissue has attracted considerable interest since their natural substrate, arachidonic acid, is the precursor of a number of highly potent physiological effecters (1).
Soybeans contain a number of distinct isozymes. The three major isozymes that have been purified from seeds and have been studied are L-l,' L-2, and L-3. The primary sequences of L-l (2) and L-2 (3) have been reported from this laboratory, whereas that of L-3 has been reported by Yenofsky et al. (4). L-l, L-2, and L-3 have M, values of 94,038,97,035 and 96,541, respectively. All contain a single iron atom/molecule. L-2/L-1 and L-2/L-3 are 81 and 74% identical in amino acid sequence, respectively. L-l is activated by its own hydroperoxide product (5). Activation is accompanied by a shift in the optical spectrum to longer wavelength (6-9). L-2 (8, 10) and L-3 (8) behave in similar fashion. Activation results in the conversion of high-spin Fe'+ to high-spin Fe3+ (11,12). This change is reflected in the appearance of an EPR signal at g = 6 in L-l (6, 7, 13), L-2 (8, lo), and L-3 (8).
The soybean isozymes share a homologous histidine-rich region, which we first suggested as a possible iron-binding site (3). Extended x-ray absorption fine structure (14,15)  residues as ligands for the iron in L-l. However, it has been reported (17) that 2 of the histidine residues in the histidinerich region in the 5-lipoxygenase are not required for activity. The sequences of human reticulocyte B-lipoxygenase (18,19), human reticulocyte 15-lipoxygenase (20), and rabbit reticulocyte 15-lipoxygenase (21) have recently been determined. The histidine-rich region noted above occurs in the animal enzymes with reasonably good conservation. Another feature common to the soybean and mammalian lipoxygenases is the presence of a moderate but clearly conserved stretch of 12-13 amino acids which occurs between the histidine-rich region and the carboxyl terminus. Although the animal enzymes are about 150 amino acids shorter than L-l, all of them have greater than 60% similarity to L-l in the carboxyl-terminal half of their polypeptide chains. The mammalian enzymes catalyze the same primary reaction as the soybean lipoxygenases, namely the hydroperoxidation of a cis,cis-1,4-pentadiene moiety in lipids; and they all contain one atom of iron/ molecule of enzyme. It is therefore likely that they have a common tertiary structure, at least in part, and that L-l may serve as a model for at least a portion of the structure of other lipoxygenases, including the highly important human enzymes (1).
In this paper, we describe the crystallization and the prelimiaary characterization of crystals of L-l, which diffract to 2.2-A spacings and appear to be suitable for structure determination by x-ray crystallography.

Enzyme
Preparation-L-l was prepared from defatted soybean seeds as previously described (22), except that DE52 ion-exchange cellulose (Whatman) was used instead of DEAE-Sephadex.

RESULTS
AND DISCUSSION Crystals were initially grown by the hanging-drop method using 12-16% (w/v) solutions of PEG 3400 buffered with 0.1 M sodium acetate, pH 4.6. The first crystals, which appeared in 6-12 days, were poorly shaped needles or clusters of needles and occasional hollow rods. One of the better crystals (-0.5 mm long) was ground in a small glass Potter-Elvehjem homogenizer in 0.5 ml of precipitant solution (24). The stock of microcrystals so obtained was diluted 4-fold. An aliquot from this dilution was again diluted 4-fold and so on until nine dilutions were obtained. The dilutions, which were made in precipitant solution, were used for microseeding (2 ~1 added to each drop of 10 ~1) in experiments over the pH range of 4.4-4.9 for 0.05-0.2 M sodium acetate buffer in lo-14% PEG.
Protein stock solution concentration was 12 mg/ml. The best crystals were obtained with precipitant solutions of 0.2 M buffer, pH 4.9, in 14% PEG. One of these was crushed with a horsehair in a depression plate well containing 20 ~1 of crystallizing solution (10 hl of protein stock solution + 10 ~1 of precipitant solution). Ten ~1 of the crystal suspension was transferred to a second well containing 10 ~1 of crystallizing solution and mixed. In this manner, serial dilutions were continued for a total of 11 dilutions. In all cases, small uniform thick platelets were obtained. Suspensions of these crystals were used directly for seeding, without grinding, under the conditions noted above. Good single crystals, generally appearing as rods, were obtained in at least 75% of the sittingdrop experiments.
The isoelectric point of L-l is 5.68 (25). All concentrations refer to stock solutions. To save time, when evaluating the suitability of the seed dilutions, a preliminary test was carried out under the optimum conditions shown, except that 14% PEG was employed. At this concentration, enough growth was seen after 2 or 3 h to permit selection of a dilution that would generate a small number of crystals. Based on the results of the preliminary tests, l-3 ~1 of the selected seed dilution was added to 20 ~1 of crystallization droplets. Usually 2-3 days under the optimum conditions noted above was sufficient for producing crystals that were suitable for diffraction studies. A typical crystal is shown in Fig. 1. The crystals often appeared six-sided in end view, but were not regular hexagons. They were mechanically robust and were not unu!ually sensitive to x-radiation. A knowledge of the tertiary structure of L-l will complement studies directed to site-specific mutations of the lipoxygenase isozymes currently underway in our laboratory with an L-l gene-bearing plasmid that has been successfully expressed in Escherichia coli.
" Since submission of this manuscript, a report (27)