Crystallization and preliminary X-ray diffraction analysis of [beta]-cinnamomin, an elicitin secreted by the phytopathogenic fungus Phytophthora cinnamomin

# 2000 International Union of Crystallography Printed in Denmark ± all rights reserved Cinnamomin (CIN) belongs to a family of 10 kDa proteins designated as elicitins. Some of these proteins induce a hypersensitive response in diverse plant species, leading to resistance against fungal and bacterial plant pathogens. CIN was crystallized by the vapourdiffusion method using either ammonium sulfate or polyethyleneglycol (PEG) as precipitants in solutions buffered at around pH 7. These crystals are isomorphous and belong to the triclinic space group, with unit-cell parameters a = 31.69, b = 36.99, c = 44.09 AÊ , = 76.86, = 84.41, = 80.26 . A frozen crystal diffracted X-rays beyond 1.45 AÊ resolution on a synchrotron-radiation source. Received 12 October 1999 Accepted 6 January 2000


Introduction
Cinnamomin is a polypeptide secreted by the oomycetous fungus P. cinnamomi Rands, a highly aggressive ®lamentous soilborne fungus and one of the world's most destructive plant pathogens. Recently, the fungus has been shown to be involved in widespread deaths of cork oaks (Quercus suber L.) in Portugal and Spain (Brasier, 1992(Brasier, , 1993Brasier et al., 1993). Cinnamomin belongs to a group of proteins called elicitins that are produced by members of the genus Phytophthora and by Pythium vexans (Kamoun et al., 1994;Ricci et al., 1989;Huet et al., 1995). These polypeptides are holoproteins containing 98 residues.
Elicitins can cause a hypersensitive response, including leaf necrosis and cell death, and induce systemic acquired resistance in some plant species (Kamoun et al., 1994;Ricci et al., 1992). According to their toxic activity, elicitins were classi®ed as acidic -elicitins and basic -elicitins, the former being less necrotic than the latter (Kamoun et al., 1993). Moreover, -elicitins have a hydrophilic residue at position 13, usually a lysine, whereas -elicitins have a hydrophobic valine at this position.
The solution structures of -capsicein from P. capsici (Bouaziz et al., 1994) was determined using multidimensional heteronuclear nuclear magnetic resonance spectroscopy. Furthermore, the three-dimensional structure of -cryptogein from P. cryptogea was solved independently by X-ray crystallography (Boissy et al., 1996) and by NMR spectroscopy (Fefeu et al., 1997). The overall structure reveals a novel fold, consisting of six -helices and a two-stranded -sheet facing an loop (Boissy et al., 1996). Recently, the crystal structure of an elicitin±ergosterol complex was determined (Boissy et al., 1999), which is consistent with the hypothesis that elicitins may be sterol carrier proteins, as proposed by Mikes et al. (1998). More structural information for these compounds showing elicitor activity is needed for a better understanding of how elicitors bind to plant receptors.
In order to study the role of CIN in the pathogenesis of P. cinnamomi and Q. suber, a synthetic gene was constructed and expressed in a biologically active form (Duclos et al., 1998). In this work, we report the crystallization and preliminary X-ray diffraction studies of the recombinant -cinnamomin.

Protein expression and purification
A synthetic gene encoding -cinnamomin was designed using the previously determined amino-acid sequence . The gene was inserted into the methylotrophic yeast Pichia pastoris secretion vector pIC9K in frame and downstream to the -mating factor pre-pro secretion signal sequence of Saccharomyces cerevisiae. The Pi. pastoris culture supernatant was concentrated in an AMICON stirred cell concentrator using a YM10 membrane (cutoff 10 kDa) at a pressure of 3 Â 10 5 Pa; the ionic strength of this solution was decreased to 20 mM sodium acetate pH 4.0. The resulting sample was loaded onto a Macro-Prep S column equilibrated with 20 mM sodium acetate pH 4.0. A NaCl gradient was applied from 0 to 100 mM NaCl in the working buffer and the CIN fraction eluted from 25 to 50 mM NaCl. The eluate was concentrated and applied onto a Superdex 75 FPLC column equilibrated with 20 mM potassium phosphate buffered at pH 7.0 and supplemented with 150 mM NaCl.

crystallization papers 2.2. Crystallization
Initial crystallization conditions were screened at 278 and 293 K with the Hampton Research Crystal Screen kit I. The sitting-drop vapour-diffusion method was employed. The droplets contained equal amounts (2 ml) of protein (10 mg ml À1 in 10 mM Tris±HCl pH 7.5) and reservoir solutions equilibrated against 0.7 ml of reservoir solution in the well. Crystalline material was obtained under different experimental conditions. Small thin platelike crystals which were prone to forming aggregates were grown at both temperatures from 30% PEG 4K, 0.1 M Tris±HCl pH 8.5, 0.2 M lithium sulfate or ammonium acetate. Similar crystal forms were obtained in the cold room from 20% PEG 4K, 10% 2-propanol, 0.1 M HEPES pH 7.5. Many microcrystals also appeared at room temperature in 1.4 M sodium citrate buffered with 0.1 M HEPES pH 7.5.
A narrower screening was performed around these conditions, which included variations in the protein concentration, the nature and concentration of precipitants, pH and type of buffer, addition of different salts or organic solvents, drop proportion (protein to mother liquor ratio) and temperature. It was observed that the addition of certain alcohols and ammonium sulfate improved the quality of the crystals. Thicker plates were grown using ammonium sulfate as precipitant.
Finally, two crystallization conditions were optimized at 293 K, yielding crystals suitable for X-ray diffraction analysis. The protein concentration used was 15 mg ml À1 . In the ®rst conditions, the initial reservoir solution contained 2.0 M ammonium sulfate, 5% MPD, 0.1 M HEPES pH 7.0 or 7.5. After 3 d, the precipitant concentration was increased to 2.2 M. The nucleation rate was thus decreased, leading to only a few single colourless crystals within 4±5 d (Fig. 1). The second crystallization conditions consisted of 30% PEG 4K, 10% butanol, 0.2 M ammonium sulfate, 0.1 M Tris±HCl pH 7.5; it took about two weeks for crystals to grow. Crystals grown under both conditions showed similar morphologies, with maximum crystal dimensions of around 0.3 Â 0.2 Â 0.06 mm.

Diffraction data collection and processing
A crystal grown in the ®rst conditions described above was transferred using a rayon-®bre loop into mother liquor complemented with 10% glycerol (crystal A in Table 1) and was placed directly into a nitrogen stream at 115 K (Oxford Cryosystems). X-ray diffraction experiments were carried out using a MAR Research imageplate detector (30 cm) mounted on an Enraf±Nonius rotating-anode generator operating at 4.5 kW with Cu K radiation. This crystal diffracted to 2.1 A Ê resolution.
Subsequently, another crystal grown under the same conditions was taken to the ESRF in Grenoble (crystal B in Table 1). Immediately prior to the X-ray measurements, the crystal was brie¯y immersed in a cryo-protectant solution containing 30% PEG 400, 5% MPD buffered in 0.1 M HEPES pH 7.0. The crystal was then mounted in a loop and¯ash-frozen at 110 K in a cold nitrogen stream. Diffraction data were collected at the ID14-EH4 beamline using a ADSC Quantum 4 CCD detector, at a wavelength of 0.932 A Ê . Since the crystal diffracted beyond 1.45 A Ê , data collection at the synchrotron followed a two-step protocol. First, X-ray data from the frozen crystal was measured using a crystal-todetector distance of 120 mm with an oscillation range of 0.5 per ®lm (1 s exposure time) for a total of 360 images. To collect the lower resolution data, oscillation pictures were taken for 0.5 s with a frame width of 1.5 and a crystal-to-detector distance of 180 mm (120 images were measured in total). Data were processed and scaled with the HKL suite of programs (Otwinowski & Minor, 1995).

Results and discussion
The availability of a reasonable amount of protein with a high degree of purity is an important requirement for crystallization purposes. The methylotrophic yeast Pi. pastoris was chosen for the expression of -cinnamomin as it is able to express and secrete high levels of heterologous proteins (Sreekrishna et al., 1997). The transformants seem to secrete the apparently fully processed protein.
The recombinant protein was ®rst puri®ed by using cation-exchange chromatography. As the ionic strength of the media prevented the correct binding of CIN to the Macroprep S resin, the Pi. pastoris culture supernatant was not directly loaded onto this ionexchange matrix but was ®rst concentrated and dialysed through membranes with a cutoff value of 10 kDa. This ultra®ltration step led to the removal of large quantities of UV-absorbing substances without loss of the CIN. Based on the calculated isoelectric point of the protein (pI 7.55), the elution of the CIN fraction from the Macroprep S column was performed using a salt gradient in sodium acetate buffered at pH 4.0. SDS±PAGE and immunoblotting analysis showed that the fractionated protein was still contaminated by molecules of higher molecular weight and that a further puri®-  crystallization papers cation step was required. Thus, a ®nal passage through a molecular-exclusion column (Superdex75 FPLC) was necessary. The purity of the protein was assessed by SDS±PAGE stained with silver nitrate, which showed the presence of a single band corresponding to a molecular weight 9 10 kDa, the same as the natural protein.
The yield of this puri®cation protocol was about 48 mg of puri®ed recombinant CIN per litre of culture. After the initial crystallization screening, crystalline material was obtained within a week. However, the optimization of these experimental conditions was a dif®cult process, owing to the tendency of the protein to form clusters of thin-shaped plates and to the non-reproducibility of the crystallization experimental results. Finally, high-quality single crystals were produced after 3±4 months of trial-and-error crystallization experiments.
Cryo-stabilization of the crystals was ®rst achieved by the addition of glycerol or MPD (to a ®nal concentration of 10±15%) to the crystallization buffers. Since these cryoprotectant solutions sometimes formed icerings, different conditions were chosen for the synchrotron data collection, namely a solution containing 30% PEG 400, 5% MPD buffered at pH 7.0.
Complete data sets were collected from single frozen crystals, which diffracted X-rays to about 2.1 A Ê resolution with a conventional X-ray source (crystal A in Table 1) and beyond 1.45 A Ê with synchrotron radiation (crystal B in Table 1). CIN crystals belong to space group P1, with unitcell parameters a = 31.69, b = 36.99, c = 44.09 A Ê , = 76.86, = 84.41, = 80.26 for the crystal measured at the ESRF (Table 1). Details of the data-collection and processing statistics from both crystals are presented in Table 1. Assuming that a dimer is present in the asymmetric unit, a V m of 2.41 A Ê 3 Da À1 is obtained, which corresponds to a solvent content of around 49%. CIN crystals were also formed using PEG as precipitant, but since these crystals are isomorphous to the ones grown from ammonium sulfate and diffract no better, no complete data set was measured for these crystals.
The CIN structure determination by molecular replacement using -cryptogein (PDB entry 1beo) as a starting model is now in progress.
We acknowledge the opportunity to collect X-ray data at the ESRF synchrotron, Grenoble, France. We are grateful to Carlos Fraza Ä o and Pedro Matias for help with data collection. MLR and MA thank Fundac Ëa Ä o para a Cie Ã ncia e Tecnologia (FCT) for their PRAXISXXI research grants (BIC/17185/98 and BPD/17265/98). We thank FCT (PRAXIS XXI 3/3.2/FLOR/2112/95) and NATO (SfS PO-CORKOAKS II) for ®nancial support.