Purification, crystallization and crystallographic analysis of Clostridium thermocellum endo-1,4-β-d-xylanase 10B in complex with xylohexaose

The N-terminal moiety of C. thermocellum endo-1,4-β-d-xylanase 10B, comprising a carbohydrate-binding module (CBM22-1) and a GH10 E337A mutant domain, has been crystallized in complex with xylohexaose. The crystals belong to the trigonal space group P3₂21, contain a dimer in the asymmetric unit and diffract to beyond 2.0 Å resolution.

Abstract

The cellulosome of Clostridium thermocellum is a highly organized multi-enzyme complex of cellulases and hemicellulases involved in the hydrolysis of plant cell-wall polysaccharides. The bifunctional multi-modular xylanase Xyn10B is one of the hemicellulase components of the C. thermocellum cellulosome. The enzyme contains an internal glycoside hydrolase family 10 catalytic domain (GH10) and a C-terminal family 1 carbohydrate esterase domain (CE1). The N-terminal moiety of Xyn10B (residues 32–551), comprising a carbohydrate-binding module (CBM22-1) and the GH10 E337A mutant, was crystallized in complex with xylohexaose. The crystals belong to the trigonal space group P3₂21 and contain a dimer in the asymmetric unit. The crystals diffracted to beyond 2.0 Å resolution.

1. Introduction

Cellulose and hemicelluloses are major constituents of plant cell walls (Bayer et al., 2004 ▶; see also the Carbohydrate-Active Enzymes website http://www.cazy.org/). Xylanase Xyn10B (formally XynY; EC 3.2.1.8) from Clostridium thermocellum is an endo-1,4-β-d-xylan hydrolase involved in the degradation of xylan, one of the most abundant hemicelluloses. Xyn10B comprises two family 22 carbohydrate-binding modules (CBM22-1 and CBM22-2; formerly X6a and X6b, respectively), which flank the glycoside hydrolase GH10 catalytic module, a dockerin sequence and a C-terminal family 1 carbohydrate esterase catalytic module (CE1; Fontes, Hazlewood et al., 1995 ▶). Structures have been elucidated of CBM22-2 (Charnock et al., 2000 ▶), the dockerin module in complex with a cognate cohesin (Carvalho et al., 2003 ▶) and CE1 (Prates et al., 2001 ▶), providing important insights into the molecular role of Xyn10B in hemicellulose hydrolysis. CE1 (residues 792–1077) displays the α/β-hydrolase fold with a classical Ser-His-Asp catalytic triad. CE1 has been shown to cleave the ferulate groups involved in the cross-linking of arabinoxylans to lignin (Tarbouriech et al., 2005 ▶; Prates et al., 2001 ▶). The dockerin module (residues 730–791), which is a repeat of approximately 24 amino acids, was solved in complex with the second cohesin module of C. thermocellum cellulosomal scaffolding protein, termed CipA. Both helices of the dockerin helix–turn–helix motif may interact with the cohesin β-sheet, revealing a general dual binding mode of dockerins to cohesins (Carvalho et al., 2003 ▶, 2005 ▶, 2007 ▶). Cellulosome assembly results from the flexible interaction of dockerin modules located on the enzymes with one of the nine cohesin domains of the CipA scaffold (Bayer et al., 2004 ▶). Therefore, the cellulosome of C. thermocellum comprises nine enzymatic sub­units in addition to CipA. CBM22-2 (residues 560–720) has a classic β-jelly-roll fold with a cleft containing three aromatic residues and two polar conserved residues (Charnock et al., 2000 ▶). Trp53, Tyr103 and Glu138 play a crucial role in xylan recognition, while Tyr136 and Arg25, although involved in ligand binding, also maintain the structural integrity of the cleft (Xie et al., 2001 ▶). The catalytic site of the GH10 module comprises Glu337 (proton donor) and Glu480 (nucleophile) (Fontes, Hall et al., 1995 ▶). A mutant construct (CBM22-­1–GH10 E337A) was made to capture the substrate xylohexaose in the inactivated site. Here, we report the preliminary crystal structural characterization of the N-terminus of Xyn10B (residues 32–551), comprising the modules CBM22-1 and GH10 E337A, in complex with xylohexaose.

2. Materials and methods

2.1. Bacterial strains, plasmids and growth conditions

The Escherichia coli strains used in this study were XL10-Gold and BL21 (DE3) (Stratagene). The plasmids used for cloning the truncated derivative of Xyn10B were pGEM T-easy (Promega) and pET-­21a (Novagen). Recombinant E. coli strains were cultured at 310 K in Luria–Bertani broth supplemented with ampicillin (100 µg ml⁻¹). Xylohexaose was obtained from Megazyme (Bray, Ireland).

2.2. Protein expression and purification

The DNA fragment encoding CBM22-1–GH10 was amplified by PCR from C. thermocellum YS genomic DNA using Pfu Turbo (Stratagene). Primers (5′-CTCGCTAGCGATTATGAAGTGGTTCATG-3′ and 5′-CACCTCGAGGGCCGGATTGTTACCGTC-3′) in­corporated NheI and XhoI restriction sites (in bold) at the 5′ and 3′ ends of the PCR product, respectively. The resulting DNA fragment was cloned into pGEM T-easy, generating pBP1, and sequenced to ensure that no mutations had occurred during amplification. The plasmid pBP1 was digested with NheI and XhoI and the excised gene was cloned into the similarly restricted expression vector pET-21a such that the recombinant protein contained an N-terminal His₆ tag. The CBM22-1–GH10 nucleophile mutant (Glu→Ala), termed CBM22-1–GH10 E337A, was generated using the PCR-based QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer’s instructions. The primers used to generate the nucleic acid mutant were GACGTTGTAAATGCGGCAGTAAGTGATGATGC and GCATCATCACTTACTGCCGCATTTACAACGTC (nucleotide substitutions in bold). The mutated DNA sequence was sequenced by MWG (Germany) to ensure that only the appropriate mutation had been incorporated into the nucleic acid. To express the recombinant CBM22-1–GH10 E337A protein, E. coli strain BL21 (DE3) harbouring the pET-21a plasmid was cultured in Luria–Bertani broth at 310 K to mid-exponential phase (A ₅₅₀ = 0.6). Isopropyl β-d-1-thiogalacto­pyranoside was then added to a final concentration of 1 mM and the cultures were incubated for a further 5 h. Cells were collected by centrifugation and the cell pellet was resuspended in 50 mM sodium HEPES buffer pH 7.5 containing 1 M NaCl and 10 mM imidazole. CBM22-1–GH10 E337A was purified by immobilized metal-ion affinity chromatography as described previously (Najmudin et al., 2006 ▶). The protein was buffer-exchanged, using a PD-10 Sephadex G-25M gel-filtration column (Amersham Bio­sciences), into 50 mM HEPES buffer pH 7.5 containing 200 mM NaCl (buffer A) and concentrated to 20 mg ml⁻¹ with Amicon 10 kDa molecular-weight centrifugation membranes. Gel filtration with a HiLoad 16/60 Superdex 75 column (Amersham Biosciences) was used to further improve protein purity and the protein was eluted at 1 ml min⁻¹ in buffer A (Fig. 1 ▶). The purified protein was concentrated as described above and washed three times with water containing 2 mM CaCl₂. The final protein concentration was adjusted to 60 mg ml⁻¹. The protein sequence of the construct corresponds to amino-acid residues 32–551 (UniProt accession No. P51584) with the E337A mutation. With the additional 20 amino-acid residues (MGSSHHHHHHSS­GLVPRGSH) at the N-­terminus the expected final molecular weight is 60 kDa.

Figure 1.

A Coomassie Brilliant Blue-stained 10% SDS–PAGE gel evaluation of protein purity during purification. Lane M, molecular-weight markers (kDa); lane 1, cell debris after sonication; lane 2, cell-free extract after sonication; lane 3, purified CBM22-1–GH10 E337A after Ni-affinity chromatography; lane 4, purified CBM22-­1–GH10 E337A after gel-filtration chromatography. Although the protein is a highly pure single polypeptidic chain of 60 kDa, two bands were observed in the gel.

2.3. Crystallization

Crystallization conditions were screened by the hanging-drop vapour-phase diffusion method using an in-house modified version of the sparse-matrix method of Jancarik & Kim (1991 ▶) and the commercial screens Crystal Screen, Crystal Screen 2 and PEG/Ion Screen from Hampton Research (California, USA). Drops consisting of 1 µl of 20, 40 and 60 mg ml⁻¹ CBM22-1–GH10 E337A and 1 µl reservoir solution were prepared and equilibrated against 1 ml reservoir solution at 292 K. Protein crystals were obtained at 292 K within 4–6 d using the following conditions: 1 M sodium acetate, 0.1 M HEPES pH 7.5, 0.05 M CdSO₄ and 10 mM xylohexaose (Fig. 2 ▶). The crystals were flash-cooled in liquid nitrogen after soaking in a cryoprotectant solution [the crystallization mother liquor containing 30%(v/v) glycerol] for a few seconds.

Figure 2.

Crystals of CBM22-1–GH10 E337A obtained by hanging-drop vapour diffusion in the presence of 1 M sodium acetate, 0.1 M HEPES pH 7.5 and 0.05 M CdSO₄. The largest crystals are approximately 0.2 mm in their longest dimension.

2.4. Data collection and processing

Data were collected on beamline ID14-EH1 at the ESRF (Grenoble, France) using a Quantum-4 charge-coupled device detector (ADSC) with the crystal cooled at 100 K using a Cryostream (Oxford Cryosystems Ltd). All data sets were processed using the programs MOSFLM (Leslie, 1992 ▶) and SCALA (Kabsch, 1978 ▶) from the CCP4 suite (Collaborative Computational Project 4, Number 4, 1994 ▶). The crystal belonged to the trigonal space group P3₁21 or P3₂21 as indicated by POINTLESS (Evans, 2006 ▶). Plots of the acentric and centric moments and the cumulative intensity distribution from the SCALA output indicated that the crystal form was indeed P3₁21 or P3₂21 and not the the lower symmetry twin parental P32. An anisotropy analysis plot calculated by TRUNCATE (French & Wilson, 1978 ▶) showed the data to be isotropic. Calculation of the Matthews coefficient indicated the possibility of a range of oligomers from a dimer to a pentamer in the asymmetric unit (Table 1 ▶; Matthews, 1968 ▶). The CCP4 programs ECALC, POLARRFN and RFCORR (Collaborative Computational Project 4, Number 4, 1994 ▶) and AMoRE (Navaza, 2001 ▶) were used to resolve this ambiguity. ECALC converts F values to E values and gives beautifully sharp maps. POLARRFN was used to calculate a fast rotation function in polar angles at different radii (20, 25, 30, 35 Å). The self-rotation function (SRF) indeed has clear noncrystallographic symmetry (NCS) peaks on the 180° section (Fig. 3 ▶); the axes are in the xy plane at 20° and 40° from the x axis (and obviously repeated by the crystallographic symmetry) and it is possible that there is also one at 60° under the crystallographic twofold. However, there is no twofold peak along the z axis, which would appear to rule out a 222 tetramer. There are also clear peaks corresponding to rotations about the z axis at χ = 40°, 80° and 160° (and also obviously at 120° for the crystallo­graphic threefold). These rotations probably relate molecules in different asymmetric units. Thus, every peak can be explained by having just a single NCS twofold. An NCS trimer or tetramer would give many more peaks. AMoRE was used to compute both the cross-rotation function (XRF) and SRF maps in Eulerian space and they were fed into RFCORR. RFCORR relates pairs of peaks in the XRF to the peaks found in the SRF. Only two of the XRF solutions were consistent with the SRF, leaving no doubt that there is just a single NCS twofold, indicating the presence of a dimer in the asymmetric unit with an ∼75% solvent content.

Table 1. Matthews coefficient calculations.

No. of molecules per ASU	Matthews coefficient (Å3 Da−1)	Solvent content (%)	Probability
2	4.74	74.07	0.01
3	3.16	61.10	0.17
4	2.37	48.14	0.68
5	1.90	35.17	0.13

Figure 3.

Plot of a 180° section of a self-rotation function from POLARRFN. The plot was calculated with a radius of 35 Å.

The programs Phaser (McCoy et al., 2005 ▶) and MOLREP (Vagin & Teplyakov, 2000 ▶) were used in attempts to solve the native structure by molecular replacement using an ensemble of GH10 domains [PDB codes 1xyz (Dominguez et al., 1995 ▶), 1nq6 (Canals et al., 2003 ▶), 1r86 (Bar et al., 2004 ▶) and 1n82 (Solomon et al., 2007 ▶); sequence identities 30–35%] and an ensemble of CBM22-2 domains [PDB codes 1dyo (Charnock et al., 2000 ▶), 1h6x and 1h6y (Xie et al., 2001 ▶); sequence identity 22%]. Both programs found good solutions for two GH10 domains in space group P3₂21, but not for the CBM22-1 domains. Automated model building using ARP/wARP (Cohen et al., 2004 ▶) and the solutions from molecular replacement and the X-ray data located 300 amino acids with a sequence coverage of 28% and an estimated correctness of the model of approximately 65%. Completion of the structure is ongoing.

Table 2. X-ray crystallography data-collection statistics.

Values in parentheses are for the lowest/highest resolution shells.

Data set	CBM22-1–GH10 E337A
X-ray source	ID14-EH1, ESRF
Wavelength (Å)	0.9340
Space group	P3221
Unit-cell parameters
a = b (Å)	173.2
c (Å)	131.8
Resolution limits (Å)	65.8–2.0
No. of observations	2939405 (100685/232571)
No. of unique observations	153122 (5107/22116)
Multiplicity	19.2 (19.7/10.5)
Completeness (%)	99.9 (99.7/99.7)
〈I/σ(I)〉	17.3 (51.2/1.7)
Rmerge†	12.0 (4.5/>100)
Rp.i.m.‡	2.8 (1.1/44.1)

^† R _merge = , where I_i(hkl) is the intensity of the ith measurement of reflection hkl and 〈I(hkl)〉 is the mean value of I_i(hkl) for all i measurements.

^‡ R _p.i.m. = and is a measure of the quality of the data after averaging the multiple measurements.