Structural Biology and Crystallization Communications Preliminary Crystallographic Analysis of L-2-keto-3-deoxyarabonate Dehydratase, an Enzyme Involved in an Alternative Bacterial Pathway of L-arabinose Metabolism

l-2-Keto-3-deoxyarabonate (l-KDA) dehydratase is a novel member of the dihydrodipicolinate synthase (DHDPS)/N-acetylneuraminate lyase (NAL) protein family and catalyzes the hydration of l-KDA to-ketoglutaric semialdehyde. l-KDA dehydratase was overexpressed, purified and crystallized at 291 K using the hanging-drop vapour-diffusion method. The crystal diffracts to 2.0 A ˚ resolution using synchrotron radiation and belongs to the trigonal space group P3 1 21 or its enantiomorph P3 2 21, with unit-cell parameters a = b = 78.91, c = 207.71 A ˚ .


Introduction
Azospirillum brasilense, a nitrogen-fixing bacterium, metabolizes l-arabinose through a hypothetical pathway via nonphosphorylated intermediates, which differs from the well known bacterial and fungal pathways. We are in the process of characterizing this alternative l-arabinose pathway enzymatically and genetically and have recently identified a set of five metabolic genes (Watanabe, Kodaki et al., 2006a,b;Watanabe, Shimada et al., 2006;Watanabe et al., 2007). l-2-Keto-3-deoxyarabonate (l-KDA) dehydratase (EC 4.2.1.43; encoded by the AraD gene; GenBank accession No. BAE94270) catalyzes the fourth reaction step in this pathway, converting l-KDA to -ketoglutaric semialdehyde (Fig. 1). This enzyme consists of 309 amino-acid residues and belongs to the dihydrodipicolinate synthase (DHDPS)/N-acetylneuraminate lyase (NAL) protein family, which includes the archetypal DHDPS and NAL and also d-5-keto-4-deoxyglucarate dehydratase, d-2-keto-deoxygluconate aldolase, trans-o-hydroxybenzylidenepyruvate hydrolase-aldolase and trans-2 0 -carboxybenzalpyruvate hydratase-aldolase (Watanabe, Shimada et al., 2006). Although these enzymes form single clusters in the phylogenetic tree, l-KDA dehydratase shows a poor relationship to all of the subclasses. Many crystallographic analyses of DHDPS/NAL proteins, including DHDPS (Blagova et al., 2006;Blickling et al., 1997;Mirwaldt et al., 1995;Pearce et al., 2006;Tam et al., 2004), NAL (Barbosa et al., 2000;Izard et al., 1994) and d-2-keto-deoxygluconate aldolase (Theodossis et al., 2004), have revealed a common (/) 8barrel fold and a common reaction step in their reactions, namely the formation of a Schiff-base intermediate between a strictly conserved lysine residue and the C 2 carbon of a common -keto acid moiety of the substrate. In the case of l-KDA dehydratase, the structurally equivalent lysine residue has been assigned in the amino-acid sequence (Lys171), but the enzyme reaction does not include cleavage of the C-C or C C bond of the substrate, in contrast to the reactions catalyzed by other DHDPS/NAL enzymes; therefore, three-dimensional structure analysis should provide a novel insight # 2007 International Union of Crystallography All rights reserved Figure 1 The enzyme reaction catalyzed by l-KDA dehydratase.
into the catalytic mechanism of this unique enzyme. Here, we describe the crystallization and preliminary X-ray crystallographic analysis of l-KDA dehydratase.

Protein expression and purification
The l-KDA dehydratase gene from A. brasilense was cloned into pQE-80L (Qiagen), a vector that adds an N-terminal His 6 tag to expressed proteins. The gene with 13 additional residues (MRGSHHHHHHGST) at its N-terminus was transformed into Escherichia coli DH5 cells (Watanabe, Shimada et al., 2006). This recombinant E. coli was grown in Super Broth medium (pH 7.0; 12 g peptone, 24 g yeast extract, 5 ml glycerol, 3.81 g K 2 H 2 PO 4 and 12.5 g of K 2 HPO 4 per litre) containing 50 mg l À1 ampicillin at 310 K for 3 h. After addition of 1 mM isopropyl -d-thiogalactopyranoside, the culture was grown for a further 6 h to induce the expression of His 6tagged protein. Cells were harvested, resuspended in buffer A (50 mM sodium phosphate pH 7.5 containing 300 mM NaCl, 10 mM imidazole and 1 mM sodium pyruvate) and lysed by sonication. After centrifugation, the supernatant was loaded onto an Ni-NTA Superflow column (Qiagen) equilibrated with buffer A linked to an Ä KTA Purifier system (Amersham Biosciences). The column was washed with buffer B [buffer A containing 10%(v/v) glycerol and 50 mM instead of 10 mM imidazole] and the proteins were eluted with buffer C (buffer B containing 250 mM instead of 50 mM imidazole). The solution was concentrated by ultrafiltration with Centriplus YM-30 (Millipore) and loaded onto a HiLoad 16/60 Superdex 200 column (1.6 Â 60 cm, Amersham Biosciences) equilibrated with buffer D (50 mM Tris-HCl pH 7.5 containing 1 mM sodium pyruvate). The main single-peak fractions were collected and concentrated to 80 mg ml À1 . These purification procedures gave an enzyme preparation that showed a single band on SDS-PAGE which corresponded to polypeptide with a subunit molecular weight of $35 kDa. The native molecular weight was estimated to be $85 kDa by gel filtration, suggesting a dimeric enzyme structure. The specific activity of the recombinant enzyme for l-KDA was 33.3 units per milligram of protein, which was similar to that of the native enzyme (19.8 units per milligram of protein; Watanabe, Shimada et al., 2006). This enzyme sample was dialyzed against buffer D containing 50%(v/v) glycerol and stored at 238 K until use.

Crystallization
The stock solution of l-KDA dehydratase was dialyzed at 277 K against distilled water for 2 h without significant inactivation and adjusted to an appropriate concentration with distilled water by the method of Lowry et al. (Lowry et al., 1951). When Tris-HCl buffer was used as a dialyzing solvent instead of distilled water, no further improvement was observed. All crystallization experiments were carried out by the hanging-drop vapour-diffusion method in 24-well Linbro tissue-culture plates (ICN Inc.) at 291 K. Each drop was formed by mixing equal volumes (5 ml) of protein solution and reservoir solution. The initial trial was carried out using Crystal Screens I and II (Hampton Research), Wizard I and II (Emerald BioSystems) and JB Screen Classic (Jena Bioscience). The best crystal of l-KDA dehydratase was obtained within 2 d using a reservoir solution consisting of 50 mM Tris-HCl pH 7.5, 0.4 M NH 4 H 2 PO 4 and 5%(v/v) 2-propanol and a protein concentration of 4-10 mg ml À1 (Fig. 2a).

X-ray analysis
A crystal of the enzyme picked up from a droplet was transferred to reservoir solution containing 30%(v/v) 2-methyl-2,4-pentanediol as a cryoprotectant. The crystal was mounted on a nylon loop (Hampton Research, Laguna Niguel, CA, USA) and placed directly in a cold nitrogen-gas stream at 100 K. X-ray diffraction images were collected at 100 K from crystals in the nitrogen-gas stream using a Quantum 4R CCD detector and synchrotron radiation of wavelength 1.0 Å at station BL-38B1 of SPring-8 (Hyogo, Japan; Fig. 2b). The distance between the crystal and detector was set at 22.5 cm and 0.8 oscillation images were recorded with 10 s exposure. Diffraction data were obtained from the crystal in the resolution range 48.8-2.0 Å and were processed using the HKL-2000 program package (Otwinowski & Minor, 1997). The space group of the crystal was determined to be P3 1 21 or its enantiomorph P3 2 21 (trigonal), with unit-cell parameters a = b = 78.91, c = 207.71 Å . Table 1   possibility of collecting higher resolution data; however, the crystalto-detector distance was limited in order to separate the spots during these measurements. The V M value (crystal volume per unit protein molecular weight; Matthews, 1968) was calculated to be 2.7 Å 3 Da À1 assuming the presence of two molecules of the enzyme in the asymmetric unit; the solvent content was 53.5%. The V M value and solvent content lie within the ranges usually found for protein crystals.
To date, several crystal structures of DHDPS/NAL family enzymes have been reported, as described in x1. Of these, DHDPS from Thermotoga maritima (PDB code 1o5k) shows the highest sequence similarity to l-KDA dehydratase (29% identity). We attempted molecular replacement using CNS v.1.1 (Brü nger et al., 1998) with various parts of the T. maritima DHDPS structure as the search probe, but were unsuccessful. This is probably because sequence homology between l-KDA dehydratase and other DHDPS/NAL proteins is limited within the N-terminal half of the protein. As an alternative, the preparation of heavy-atom derivatives, including iodine, platinum and mercury derivatives, for use in the multiplewavelength anomalous dispersion method is currently in progress.