Diffraction Structural Biology Synchrotron Radiation Crystal Structure of Endo-1,4-b-glucanase from Eisenia Fetida

The saccharification process is essential for bioethanol production from woody biomass including celluloses. Cold-adapted cellulase, which has sufficient activity at low temperature (< 293 K), is capable of reducing heating costs during the saccharification process and is suitable for simultaneous saccharifica-tion and fermentation. Endo-1,4-glucanase from the earthworm Eisenia fetida (EF-EG2) belonging to glycoside hydrolase family 9 has been shown to have the highest activity at 313 K, and also retained a comparatively high activity at 283 K. The recombinant EF-EG2 was purified expressed in Pichia pastoris, and then grew needle-shaped crystals with dimensions of 0.02 Â 0.02 Â 1 mm. The crystals belonged to the space group P3 2 21 with unit-cell parameters of a = b = 136 A ˚ , c = 55.0 A ˚. The final model of EF-EG2, including 435 residues, two ions, seven crystallization reagents and 696 waters, was refined to a crystallographic R-factor of 14.7% (free R-factor of 16.8%) to 1.5 A ˚ resolution. The overall structure of EF-EG2 has an (/) 6 barrel fold which contains a putative active-site cleft and a negatively charged surface. This structural information helps us understand the catalytic and cold adaptation mechanisms of EF-EG2.


Introduction
Cellulose is the most abundant organic molecule on earth and is an excellent target for bioethanol production by biomass conversion. Bioethanol production from plant-derived lignocellulosic wastes would effectively reduce food costs relative to that from corn or cane juice (Dashtban et al., 2009). Most process concepts for bioethanol from lignocellulosic material start with a thermo-chemical hydrolysis of the hemicelluloses part (pretreatment), followed by an enzymatic hydrolysis of the cellulose part (sacccharification) and yeast-based fermentation of the resulting sugar. Fungi, bacteria and invertebrates produce enzymes named cellulases that are capable of degrading the linear biopolymers of anhydroglucopyranose connected by -1,4-glycosidic bonds into sugars. In the pretreatment process, the lignocellulosic materials are heated over 373 K in the alkaline or acidic solution to expose the cellulose. The use of thermophilic fungal species, such as Sporotrichum thermophile and Thielavia terrestris, can avoid a costly cooling step after the pretreatment process (Kovacs et al., 2009;Ingram et al., 2011). Instead, the reaction temperature must be kept above 333 K by a heating step to maintain its enzymatic activity at the saccharification process. Furthermore, such high temperature is unsuitable for the adaptation of the simultaneous saccharification and fermentation (SSF) process. In the SSF process, the enzymatic hydrolysis is performed together with the fermentation. The principal benefits of the SSF process are the reduced endproduct inhibition of the enzymatic hydrolysis, and the reduced investment costs. In addition, SSF is capable of the production of a high concentration of ethanol (> 20%). Thus, SSF is today important in the dry-milling process in the cornbased ethanol industry in the USA (Bothast & Schlicher, 2005), and is also an interesting process option for bioethanol production from lignocelluloses (Olofsson et al., 2008). Unfortunately, the enzymatic activity of most thermophilic organisms is almost lost at the optimal temperature (298-303 K) in the SSF process.
Recently, we cloned the gene for endo-1,4--glucanase from the earthworm Eisenia fetida (EF-EG2), which consisted of 1368 bp encoding 456 amino acid residues . The amino acid sequence of the gene shares sequence homology (> 50%) with endo-1,4--glucanases belonging to glycoside hydrolase (GH) family 9. Recombinant EF-EG2 PDB Reference: 3wc3 hydrolyzes soluble cellulose (carboxymethyl cellulose), but not insoluble (powdered cellulose) or crystalline (Avicel) cellulose substrates. Thin-layer chromatography analysis of the reaction products from 1,4--linked oligosaccharides of various lengths revealed a cleavage mechanism consistent with endoglucanases (not exoglucanases). The enzyme exhibited significant activity at 283 K (38% of the activity at optimal 313 K) and was stable at pH 5.0-9.0, with an optimum pH of 5.5. However, the catalytic mechanism and the cold adaptation mechanism of EF-EG2 are still unknown. Here we report the first crystal structures of EF-EG2. Structural information of EF-EG2 provided useful information for understanding its catalytic mechanism by comparing it with other GH family 9 enzymes. In addition, we discuss the cold adaptation mechanism of EF-EG2 with regard to its structural features.
After cultivation, the culture media was centrifuged at 8200g for 10 min at 277 K, before the supernatant was recovered. EF-EG2 was precipitated from the supernatant with ammonium sulfate (80% saturation) followed by centrifugation at 20000g for 30 min. The precipitate was dissolved in 20 mM Tris-HCl buffer (pH 7.5) containing a Protease Inhibitor Cocktail (Nacalai Tesque). EF-EG2 was purified using a Superdex 75 gel filtration column (GE Healthcare) equilibrated with 20 mM HEPES buffer (pH 7.5) containing 200 mM sodium chloride. The eluted fraction containing EF-EG2 was diluted with the same amount of 20 mM Tris-HCl buffer (pH 7.5), and applied to a RESOURCE Q anionexchange column (1 ml; GE Healthcare). After the column had been washed with 10 ml of 20 mM Tris-HCl buffer (pH 7.5) containing 50 mM sodium chloride, the EF-EG2 was eluted with a linear gradient from 20 mM Tris-HCl buffer (pH 7.5) containing 50 mM sodium chloride to the same buffer containing 400 mM sodium chloride with a flow rate of 1.0 ml min À1 for 20 min. Purified EF-EG2 was concentrated to 5.5 mg ml À1 for crystallization.

Crystallization
Initial screening for EF-EG2 crystallization was performed by the sitting-drop vapour-diffusion method at 293 K using 96well Intelliplates (Hampton Research) and a Hydra II Plus One (Matrix Technology). Each sitting drop was prepared by mixing 0.3 ml of the protein solution and of reservoir solution; the resulting drop was equilibrated against the reservoir solution. The initial search for crystallization conditions was performed using the following screening kits: Crystal Screen I and II (Hampton Research), Wizard Screen I and II (Emerald Biostructures) and Precipitant Synergy (100, 67 and 33% of its primary concentration; Emerald Biostructures). After 10 d, small crystals were obtained in some conditions. Crystallization conditions were further optimized by changing the precipitant based on the conditions from Wizard Screen II No. 31 (200 mM sodium chloride, 1 M sodium citrate, 100 mM Tris-HCl pH 7.0). The optimization of crystallization conditions was performed by using the hanging-drop vapour-diffusion method at 293 K. Needle-shaped crystals with dimensions of 0.02 Â 0.02 Â 1 mm were grown from a drop consisting of 2.0 ml each of the protein solution and the reservoir solution containing 200 mM sodium chloride, 600 mM sodium citrate and 67 mM Tris-HCl pH 7.0.

Data collection and refinement
For X-ray diffraction measurements under cryogenic conditions, EF-EG2 crystals were rinsed with the well solution containing 20% (v/v) glycerol as a cryo-protectant and then flash-cooled in a cold nitrogen gas stream. X-ray diffraction data from the crystal were collected at 100 K using an ADSC Quantum 315r CCD detector (Area Detector Systems Co., CA, USA) and synchrotron radiation (0.98 Å wavelength) at beamline BL17A at Photon Factory, KEK (Tsukuba, Japan). The oscillation angle was 1.0 and the exposure time was 2.0 s per frame. In total, 180 diffraction images were recorded at a camera distance of 155.4 mm and were processed using HKL2000 (Otwinowski & Minor, 1997) to 1.5 Å resolution. The crystal belonged to the space group P3 2 21 with unit-cell dimensions a = b = 136, c = 55.0 Å . The Matthews coefficient was 2.9 Å 3 Da À1 assuming that the presence of one molecule in the asymmetric unit corresponded to the solvent content of 57%.

Figure 1
Overall structure of EF-EG2. Inner and outer helices of the (/) 6 barrel, other short helices and -strands are colored in cyan, blue, light purple and magenta, respectively. Crystallization reagents and putative active residues are shown as purple and magenta stick models, respectively. Two ions, calcium and sodium, are shown as green and purple sphere models, respectively. All figures were prepared using the PyMOL Molecular Graphics System (DeLano Scientific, San Carlos, CA, USA).

Figure 2
Binding sites of calcium (a) and sodium (b) ions in EF-EG2 structure. The green and purple contours show the F o À F c (7.0) map calculated without calcium and sodium ions, respectively. Each binding site is composed of eight and six oxygen atoms chelating to a calcium and sodium ion with distances of 2.3-2.5 and 2.3-2.7 Å (represented by black dashed lines), respectively. in the catalytic activities of these enzymes is still unclear. A sodium ion, probably induced by the reservoir solution containing 200 mM sodium chloride and 600 mM sodium citrate, exists in the loop region between 1 and 2 with a triangular antiprism geometry formed by three waters (2.3, 2.4 and 2.7 Å ), a main-chain carbonyl group (O-Leu44: 2.4 Å ) and two side-chain carboxyl groups (O1-Asp43, O1-Asp55: 2.4 and 2.4 Å ) (Fig. 2b). These aspartic acids are conserved in endoglucanase Cel9G from Clostridium cellulolyticum (Mandelman et al., 2003) and also participate in the recognition of magnesium ions included in the crystallization medium. In addition, several clear electron densities were interpreted as a citrate (precipitant), a Tris (buffer) and five glycerol molecules (cryo-protectant).

Active site
Out of the above crystallization reagents, a Tris (TRS) and three glycerol molecules (GOL1, GOL2 and GOL3) were confirmed with sufficient electron densities at the open acidic cleft located at the N-terminal site of the inner barrel [Figs. 3(a) and 3(b)]. The volume of this cleft was calculated to be 320 Å 3 (about 25 Å long, 4-6 Å wide and 6-8 Å deep) using POCASA (http://altair.sci.hokudai.ac.jp/g6/index-e.html) (Yu et al., 2010). This acidic cleft has been widely known as an active-site cleft in GH family 9 enzymes. Fourteen structures (six enzymes) out of 28 structures (11 enzymes) deposited in the PDB have been determined as complex structures with glucopyranose molecules (Sakon et al., 1997;Parsiegla et al., 2002;Mandelman et al., 2003;Schubot et al., 2004;Pereira et al., 2009;Eckert et al., 2009;Moré ra et al., 2011), and in all 14 structures glucopyranose molecules locate in this cleft. Recognition residues of Tris and glycerols are shown in Fig. 3(c). Several aromatic residues, His144, Tyr226, Trp270, Trp320, His378 and Tyr427, along the cleft were situated within 4 Å of the Tris and glycerol molecules. In addition, some hydrogen-bonding interactions contribute to recognition of these molecules. Three hydrogen bonds are constructed, between hydroxyl oxygen atoms of GOL2 and the side-chain nitrogen atom of Trp270 and Arg324, with an N-O distance of 2.9 Å . The hydroxyl oxygen atom of GOL1 is 3.4 and 2.8 Å distant from the side-chain nitrogen and oxygen atoms of His378 and Glu431, respectively. TRS is recognized by the side-chain carboxyl oxygen atom of Glu431 via direct hydrogen bond (O-O distance: 2.6 Å ) and by those of the Asp74 and Asp77 via water-mediated hydrogen bonds. These glutamic and aspartic acid residues, Glu431, Asp74 and Asp77, are conserved throughout GH family 9 enzymes, and are thought to be crucial residues for its catalytic activity. These three acidic residues are located in the upper limb of the cleft (Fig. 3a).
To understand the catalytic mechanism of EF-EG2, the crystal structure of an inactive mutant (E795Q) of cellobiohydrolase CbhA from Clostridium thermocellum in complex with cellotetraose (CTT) (Schubot et al., 2004) was superimposed on the EF-EG2 structure at three C atoms of putative catalytic residues (Fig. 3c). In superposed structures, the location of TRS, GOL1 and GOL2 loosely corresponded to the glucose units in subsites À1, +1 and À2, respectively. This consistency is not surprising because we already confirmed experimentally that glycerols (cryo-protectants) and Tris or HEPES molecules (buffer components) occupy a similar position as the substrate, N-acetyl-d-glucosamine residues, in the crystal structures of chitinase C from Ralstonia sp. A-471 (Arimori et al., 2013). This superposed model structure gave us useful information about the catalytic mechanism of EF-EG2. The side-chain carboxyl group of Glu431 may exist within the possible distance [red dashed line in Fig. 3(c)] to donate a proton to the scissile bond as a general acid in the inverting mechanism. On the other hand, a water molecule, which is located at the same position as the catalytic water in the superposed structure of the CbhA-CTT complex, may be situated close to the C1 atom to be suitable for nucleophilic attack [red dashed line in Fig. 3(c)]. One of two aspartic acid residues situated within hydrogen-bonding distance may activate catalytic water as a general base. Furthermore, the aromatic residues around Tris and glycerol molecules are conserved as aromatic amino acids in CbhA. Some aromatic residues in the EF-EG2 structure may also play a role as possible partners for stacking interactions with the substrate. These hypotheses on the catalytic and substrate recognition mechanism of EF-EG2 will have to be tested through a combination of structural and mutation studies.

Cold adaptation
Structural features of cold-adapted enzymes have been summarized in a review by Siddiqui & Cavicchioli (2006). The article concludes that high structural flexibility, particularly around the active site, is translated into low-activation enthalpy, low-substrate affinity and high specific activity at low temperatures. The high flexibility is also accompanied by a trade-off in stability, resulting in heat lability. There was no remarkable structural flexibility at the active site because ten residues, Asp74, Asp77,His144,Tyr226,Trp270,Trp320,His378,Tyr427 and Glu431 [shown in Fig. 3(c)] which appear to concern enzymatic activity have almost the same average Bfactor (7.3 Å 2 ) as that of the overall structure (8.3 Å 2 ). In their review, overall structural features correlated with high flexibility. These structural features are: (i) high surface charge, particularly negative charge; (ii) few electrostatic interactions, particularly few arginine-mediated interactions; (iii) few hydrophobic interactions, particularly few aromatic-aromatic interactions; (iv) secondary structure elements, such as weak intrahelical charge-dipole interactions, and the existence of a proline residue as a helix breaker; (v) surface loops with few prolines and many glycines; (vi) other factors, such as bridges via metal ion and disulfide bonds.
Because of these structural features, we compared the structure of EF-EG2 with that of NtEgl from N. takasagoensis, which live in subtropical zones. NtEgl has an optimal activity at 340 K and loses 47% activity at 303 K (Kesavulu et al., 2012). There was no significant difference between both structures with regard to features (ii)-(vi) listed above.
However, there was a slight difference in molecular surface charge (Fig. 4). The substrate binding face is negatively charged in both structures (left-hand figures); in contrast, the distribution of negative charge at the opposite face (righthand figures) of EF-EG2 is larger than that of NtEgl. The negatively charged amino acids (Asp and Glu) of EF-EG2 occupy over two-thirds of accessible surface area (ASA) from the total charged amino acids (ASA Asp,Glu : 2900 Å 2 ; ASA Arg,Lys : 1300 Å 2 ) whereas those of NtEgl occupy about half (ASA Asp,Glu : 2400 Å 2 ; ASA Arg,Lys : 2300 Å 2 ). In addition, those of the Clostridium thermocellum cellulase CelT (CtCelT) which has optimal activity at 343 K and retains 72% activity at 303 K (Kesavulu et al., 2012) occupy about half (ASA Asp,Glu : 2900 Å 2 ; ASA Arg,Lys : 3200 Å 2 ). CtCelT has a slightly higher activity at 303 K than NtEgl, but almost loses activity at 293 K. Thus, it seems that the highly negatively charged surface of EF-EG2 contributes to its cold adaptation under 293 K. The increase in surface negative charge has also been described for cold-adapted cellulase belonging to GH family 5 from the Antarctic bacterium Pseudoalteromonas haloplanktis which retains sufficient activity at 297 K (Garsoux et al., 2004).
At low temperature, the energy cost of breaking the hydrogen-bonding network is very high because of the high viscosity and high surface tension of water (Kumar & Nussinov, 2004). Therefore, the energetic cost may be cancelled out by the surface charge of acidic amino acids interacting with water molecules, accordingly maintaining overall structural flexibility. In addition, the localization of acidic residues in the surface may produce charge-charge repulsions causing destabilization of overall structure. This low structural stability caused by a lop-sided negatively charged surface in cold-adapted enzymes may mainly contribute to maintain its activity at low temperature.
We thank the staff of the Photon Factory, Japan, for their assistance with X-ray data collection (proposal Nos. 11G088 and 13G122). Electrostatic potentials of (a) EF-EG2 and (b) NtEgl. Solvent accessible surfaces are contoured from À3kT (red) to +3 kT (blue).

Figure 3
Putative active site of EF-EG2. (a) Solvent accessible surface showing the binding cleft. The surface is colored according to the local electrostatic potential as calculated by APBS (Baker et al., 2001) from À3kT (red) to +3kT (blue). (a) is drawn from the identical direction of Fig. 1 and is a close-up view of the left-hand side of Fig. 4(a). (b) Electron densities of putative active residues and crystal reagents at the binding cleft. The dark green contour shows the F o À F c (3.5) map calculated without TRS, GOL1, 2, 3 molecules and the side-chain atoms Asp74, Asp77 and Glu431. (c) Superposition of residues Asp74, Asp77 and Glu431 from EF-EG2 (magenta) on Asp383, Asp386 and Gln795 from CbhA (cyan). Residues within 4 Å from TRS, GOL1, GOL2 (in EF-EG2 structure, purple) and CTT (in CbhA structure, yellow) are shown as white and gray stick models, respectively. Hydrogen-bonding interactions (within 3.4 Å ) in the EF-EG2 structure are represented by black dashed lines. Bond lengths that may be useful for elucidating the catalytic mechanism of EF-EG2 are indicated by red dashed lines.