crystallization papers

Acta Crystallographica Section D Biological Purification, crystallization and preliminary X-ray Crystallography diffraction analysis of the Trichoderma reesei ISSN 0907-4449 hydrophobin HFBI

Sanna Askolin,a*³ Johan P. Hydrophobins are fungal that are capable of altering the Received 14 June 2003 Turkenburg,b Maija Tenkanen,a§ hydrophobicity of surfaces by self-assembly at hydrophilic±hydro- Accepted 9 August 2004 Sinikka Uotila,c Keith S. Wilson,b phobic interfaces. Here, the growth of hydrophobin crystals suitable a c for X-ray crystallography is reported. The hydrophobin HFBI from Merja PenttilaÈ and Kalevi Visuri Trichoderma reesei was crystallized by vapour diffusion in hanging drops in 30% PEG 4000, 0.1 M sodium citrate pH 4.3 buffer containing 0.2 M ammonium acetate and CYMAL-5 detergent aVTT Biotechnology, FIN-02044 VTT, Finland, bUniversity of York, Chemistry Department, (initial concentration of 2.4 mM). HFBI crystals are hexagonal and Structural Biology Laboratory,York YO10 5YW, belong to space group P61 (or P65), with unit-cell parameters England, and cMacrocrystal Oy, Ruukintie 20 F, a = b = 45.9, c = 307.2 AÊ . The HFBI used in the crystallization FIN-02320 Espoo, Finland experiments was puri®ed from fungal cell walls.

³ Present address: Finnzymes Oy, 1. Introduction spectroscopy (WoÈsten & de Vocht, 2000). Self- Keilaranta 16 A, FIN-02150 Espoo, Finland. assembly on a hydrophobic solid induced § Present address: University of Helsinki, Hydrophobins are small (approximately 100 -helical structure, while in contrast the Department of Applied Chemistry and amino acids) moderately hydrophobic proteins percentage of -sheet in class I hydrophobins Microbiology, PO Box 27, 00014 University of secreted by ®lamentous fungi. They self- Helsinki, Finland. assemble into amphiphilic membranes when increased upon self-assembly at a water±air confronted with hydrophobic±hydrophilic interface. interfaces. This hydrophobin membrane The ®lamentous T. reesei produces Correspondence e-mail: confers hydrophobicity on fungal structures three class II hydrophobins: HFBI, HFBII and [email protected] È È and mediates the attachment of the fungal HFBIII (Nakari-Setala et al., 1996, 1997; È hyphae to hydrophobic surfaces and the Penttila et al., 2000). In contrast to the - dispersal of into the air (WoÈsten, 2001). wall-speci®c HFBII, HFBI (7.5 kDa) has been Moreover, hydrophobins are some of the most isolated from submerged fungal cell walls of surface-active molecules known (WoÈsten & de vegetative mycelia in addition to liquid culture Vocht, 2000). They lower the surface tension of medium (Nakari-SetaÈlaÈ et al., 1996, 1997). The liquids, enabling the escape of fungi from amino-acid similarity between HFBI and È È aqueous solution to the air (WoÈsten et al., HFBII is 69% (Nakari-Setala et al., 1996). 1999). Hydrophobins have several potential Atomic force microscopy studies showed that applications, including modi®cation of both the hydrophobins formed highly ordered biophysical surface properties such as wett- Langmuir±Blodgett monolayer ®lms (Paan- ability, immobilization and emulsi®- anen et al., 2003). HFBI and HFBII formed cation (Wessels, 1997). Based on their ®brillar aggregates by shaking (Torkkeli et al., hydropathy patterns and the solubility of the 2002). Contrary to the ®brillar aggregates of hydrophobin assemblages, hydrophobins are HFBI, those of HFBII were stable and their divided into classes I and II (Wessels, 1994). low-resolution structure was able to be studied Hydrophobins contain eight resi- (Torkkeli et al., 2002). Here, we report for the dues at conserved positions in the sequence, ®rst time the crystallization and preliminary which appear to form disul®de bridges. The crystallographic studies of HFBI. locations of four disul®de bridges have been determined for HFBII from Trichoderma 2. Experimental reesei (HakanpaÈaÈ et al., 2004). HFBII is the only hydrophobin for which the atomic reso- 2.1. Protein production and purification lution structure is available. The hydrophobin To produce HFBI, the T. reesei HFBI over- has a single-domain structure containing one producing strain VTT D-98692 (Askolin et al., -helix and four antiparallel -strands 2001) was cultivated in feed-batch fermenta- (HakanpaÈaÈ et al., 2004). An NMR study of the tions on glucose-containing medium essentially hydrophobin EAS indi- as described previously (Askolin et al., 2001). cated that the monomers were mostly HFBI was puri®ed using two puri®cation unstructured apart form a small -sheet region methods. The ®rst included mycelial extraction (Mackay et al., 2001). Changes in the secondary with 1% SDS at pH 9, SDS removal, hydro- # 2004 International Union of Crystallography structure of hydrophobins during self-assembly phobic interaction and ion-exchange chroma- Printed in Denmark ± all rights reserved have been observed by circular-dichroism tography (Askolin et al., 2001), buffer

Acta Cryst. (2004). D60, 1903±1905 DOI: 10.1107/S0907444904019754 1903 crystallization papers exchange to 20 mM Tris±HCl buffer pH 8.0 HFBI sample and an equal volume of crys- Table 1 by gel ®ltration (Bio-Gel P-6DG, Bio-Rad) tallization reagent (total volume of 40 ml) Diffraction data statistics. and concentration by ultra®ltration (YM1 were seeded with previously grown HFBI Values in parentheses correspond to re¯ections in the membrane, Amicon). Prior to crystal- crystals to induce nucleation. outer resolution shell. lization, the pH was adjusted to the pH of Space group P65 or P61 the crystallization reagent, approximately 2.3. Data collection and processing Unit-cell parameters (AÊ ) a = b = 45.9, c = 307.2 pH 4.5, by adding an equal volume of Resolution range (AÊ ) 20±2.5 (2.59±2.5) A cryoprotectant solution was prepared Observed re¯ections 63019 (6255) 100 mM sodium acetate/acetic acid buffer by replacing 25% of the water in the crys- Unique re¯ections 12649 (1251) pH 4.5. tallization solution with glycerol. Crystals Completeness (%) 100 (100) For the second puri®cation method of Rmerge (%) 8.0 (20) were soaked brie¯y in this solution and I/(I) 21.7 (9.3) HFBI, the mycelium (1.73 kg wet weight, mounted in a rayon-®bre loop and placed h i 13.8% dry weight) was separated by ®ltra- directly into a stream of N2 gas at 120 K. tion through a 0.25 mm cloth from a working A crystal of suitable quality was chosen critical factors in the crystallization was pH. fermentation volume of 25 l (New Bruns- and used to collect native X-ray diffraction HFBI crystallized within the pH range 4.0± wick Scienti®c IF40) after 4 d of cultivation. data in the home laboratory at 120 K using a 4.5 at 288 K, somewhat different from the The mycelium was extracted with 100 mM MAR345 imaging-plate detector on a calculated isoelectric point of 5.7. HFBI sodium acetate buffer pH 5.0 containing 1% Rigaku RUH3R rotating-anode X-ray crystals of varying length (0.1±0.6 mm) were SDS (10 ml of buffer per 1 g dry weight) at generator with a Cu target operating at obtained using 24±30% PEG 4000 and the room temperature. This lower pH was used 50 kV and 100 mA equipped with multilayer other crystallization reagents mentioned to prevent the deamidation of two N-term- focusing X-ray optics (Osmic). The same above at pH 4.0. inal asparagines (Asn2 and Asn4; Askolin et crystal was subsequently used to collect data The crystals belong to space group P65 or al., 2001). Puri®cation was continued with on beamline ID14-2 at the ESRF, Grenoble P61, with unit-cell parameters a = b = 45.9, Ê SDS removal from the mycelial extract by using a MAR Research CCD detector at a c = 307.2 A. Attempts to reduce the data in precipitation with 0.4 sample volumes of 2 M wavelength of 0.934 AÊ . Both data sets were space groups belonging to Laue group 622 KCl, centrifugation and ammonium sulfate processed and reduced using the DENZO result in an Rmerge of more than 30% and precipitation in approximately 0.6 M and SCALEPACK programs (Otwinowski rejection of almost 15% of the observations, ammonium sulfate at 277 K overnight. The & Minor, 1997). indicating that the crystals contain pseudo- precipitate was removed by centrifugation 622 point symmetry. Calculation of a self- and dissolved in 60% ethanol (400 ml). Part rotation function using the program of the solution (50 ml) was buffered to 3. Results and discussion MOLREP (Vagin & Teplyakov, 1997) does 25 mM Tris±HCl pH 8.5 by gel ®ltration The ®rst HFBI crystals were obtained with indeed show non-crystallographic twofold (Bio-Gel P-6DG, Bio-Rad). Proteins eluted 30% PEG 4000, 0.1 M sodium acetate pH axes perpendicular to the sixfold axis, but from the chromatography column were 4.6 containing 0.2 M ammonium acetate. reveals no other signi®cant features (data detected by absorbance at 215 nm. The pH The crystals were very thin hexagonal plates not shown). With four molecules in the Ê 3 1 of the puri®ed HFBI was adjusted to pH 6 that grew from spherulites. HFBI crystals asymmetric unit, VM is 3.1 A Da , which with 100 mM sodium acetate buffer pH 3.8 larger in all three dimensions were obtained corresponds to a solvent content of 60% immediately after the chromatography run. by adding the non-ionic detergent (Matthews, 1968). Alternatively, six mole- Prior to crystallization, the pH was lowered CYMAL-5 (cyclohexyl-pentyl- -d-malto- cules in the asymmetric unit would result in Ê 3 1 further to the pH of the crystallization side). HFBI crystals suitable for X-ray a VM of 2.1 A Da , which corresponds to a reagent solution. HFBI content and purity analysis were obtained after approximately solvent content of 40%. The crystals were analysed by HPLC essentially as two months with 30% PEG 4000, 0.1 M diffracted strongly in the home laboratory, described by Askolin et al. (2001). Based on sodium acetate buffer pH 4.3 containing which would be in keeping with 4±6 mole- the HPLC analysis, the puri®ed HFBI was 0.2 M ammonium acetate and 1 the critical cules per asymmetric unit, although the homogeneous and intact, indicating that micelle concentration of CYMAL-5 prior to latter value is more probable. For the data extraction at lower pH prevented N-term- equilibration at 277 K (or 288 K) (Fig. 1). collected at the ESRF, the resolution was inal modi®cations. limited to 2.5 AÊ owing to the long c axis The drop consisted of 5 ml of HFBI sample 1 (307.1 AÊ ) (Table 1). (2.6 g l , pH adjusted to 4.5), 1 mlof24mM Determination of the three-dimensional 2.2. Crystallization CYMAL-5 and 4 ml of the optimized crys- tallization reagent solution. One of the most structure will provide details of surface Crystallization conditions were screened conformation and the amino-acid residues and optimized by vapour diffusion in participating in surface recognition in inter- hanging drops (initial volume of 4.5±10 ml) facial self-assembly. It will allow directed using puri®ed HFBI. Initial screening was amino-acid substitutions to broaden the carried out using a standard crystallization usefulness of the protein in potential screening kit (HR2-110, Hampton medical and technical applications or even Research) at 288 K. The effect of additional in nanotechnology. detergents on crystallization was studied with a detergent screening kit (HF2-410, Hampton Research). The authors would like to thank Michael HFBI puri®ed by the new puri®cation Figure 1 Bailey for fermenter cultivations, Riitta Crystals of HFBI. The crystal dimensions are method (see 2.1) was crystallized using approximately 190 106 mm. The bar represents Isoniemi for excellent technical assistance in x  microseeding at 277 K. Batches containing 100 mm. protein puri®cation, the ESRF, Grenoble for

1904 Askolin et al.  HFBI Acta Cryst. (2004). D60, 1903±1905 crystallization papers provision of excellent data-collection facil- Mackay, J. P., Matthews, J. M., Wine®eld, R. D., PenttilaÈ, M., Nakari-SetaÈlaÈ, T., FagerstroÈm, R., ities and the beamline staff for assistance in Mackay, L. G., Haverkamp, R. G. & Templeton, Selber, K., Kula, M.-R., Linder, M. & Tjerneld, Structure the use thereof. The ®nancial support of M. D. (2001). , 9, 83±91. F. (2000). Int. Pat. Appl. PCT/FI00/00249. Matthews, B. W. (1968). J. Mol. Biol. 33, 491± Torkkeli, M., Serimaa, R., Ikkala, O. & Linder, M. Neste Oy's Foundation (personal grant to 497. (2002). Biophys. J. 83, 2240±2247. SA) and National Technology Agency of Nakari-SetaÈlaÈ, T., Aro, N., Ilmen, M., Kalkkinen, Vagin, A. & Teplyakov, A. (1997). J. Appl. Cryst. Finland (Tekes) is gratefully acknowledged. N. & PenttilaÈ, M. (1997). Eur. J. Biochem. 248, 30, 1022±1025. 415±423. Wessels, J. G. H. (1994). Ann. Rev. Phytopathol. Nakari-SetaÈlaÈ, T., Aro, N., Kalkkinen, N., Alatalo, 32, 413±437. E. & PenttilaÈ, M. (1996). Eur. J. Biochem. 235, Wessels, J. G. H. (1997). Adv. Microb. Physiol. 38, References 248±255. 1±45. Askolin, S., Nakari-SetaÈlaÈ, T. & Tenkanen, M. Otwinowski, Z. & Minor, W. (1997). Methods WoÈsten, H. A. B. (2001). Annu. Rev. Microbiol. 55, (2001). Appl. Microbiol. Biotechnol. 57, 124± Enzymol. 276, 307±326. 625±646. 130. Paananen, A., Vuorimaa, E., Torkkeli, M., WoÈsten, H. A. B. & de Vocht, M. L. (2000). HakanpaÈaÈ, J., Paananen, A., Askolin, S., Nakari- PenttilaÈ, M., Kauranen, M., Ikkala, O., Biochim. Biophys. Acta, 1469, 79±86. SetaÈlaÈ, T., Parkkinen, T., PenttilaÈ, M., Linder, Lemmetyinen, H., Serimaa, R. & Linder, M. WoÈsten, H. A. B., van Wetter, M.-A., Lugones, M. B. & Rouvinen, J. (2004). J. Biol. Chem. 279, B. (2003). Biochemistry, 42, 5253± L. G., van der Mei, H. C., Busscher, H. J. & 534±539. 5258. Wessels, J. G. H. (1999). Curr. Biol. 9, 85±88.

Acta Cryst. (2004). D60, 1903±1905 Askolin et al.  HFBI 1905