Laminarinase from Flavobacterium sp. reveals the structural basis of thermostability and substrate specificity

Laminarinase from Flavobacterium sp. strain UMI-01, a new member of the glycosyl hydrolase 16 family of a marine bacterium associated with seaweeds, mainly degrades β-1,3-glucosyl linkages of β-glucan (such as laminarin) through the hydrolysis of glycosidic bonds. We determined the crystal structure of ULam111 at 1.60-Å resolution to understand the structural basis for its thermostability and substrate specificity. A calcium-binding motif located on the opposite side of the β-sheet from catalytic cleft increased its degrading activity and thermostability. The disulfide bridge Cys31-Cys34, located on the β2-β3 loop near the substrate-binding site, is responsible for the thermostability of ULam111. The substrates of β-1,3-linked laminarin and β-1,3-1,4-linked glucan bound to the catalytic cleft in a completely different mode at subsite -3. Asn33 and Trp113, together with Phe212, formed hydrogen bonds with preferred substrates to degrade β-1,3-linked laminarin based on the structural comparisons. Our structural information provides new insights concerning thermostability and substrate recognition that will enable the design of industrial biocatalysts.

Glucans with β-1,3-linkages are widely distributed in nature and are found in bacteria, fungi, plants, and algae. β-1,3-linked glucan is the main constituent of botanical and fungal cell walls; it is a major structural polysaccharide 1 . Laminarin is a storage polysaccharide of marine macroalga 1, 2 . Its main chain consists of glucose with β-1,3-linkages and partial branches connected through β-1,6-linkages. Ratios of β-1,3and β-1,6-linkages are diverse, e.g., 7:1 in Laminaria digitata and L. hyperborea 3 and 3:2 in Eisenia bicyclis 4 . Laminarin has shown anti-apoptotic and anti-tumor activities 5,6 . Laminarin is also a potential source of fermentable sugars for bioethanol production 7 . Therefore, it has received attention in the design of biocatalysts for the development of a cost-competitive process for converting laminarin into fermentable sugar for the widespread utilization 8,9 . Some other extracellular polysaccharides, such as curdlan and lichenin, have also been demonstrated to be non-toxic, and have applications in the food and pharmaceutical industries 10 . Curdlan is unbranched and consists of glucosyl residues that are linked by β-D-1,3 bonds with its degree of polymerization being about 135 glucose residues 11 . Lichenin could form symbiotic relationships with algae in lichens has a ratio of β-D-1,4-glucopyranosyl to β-D-1,3-glucopyranosyl residues of 2.3:1 in Cetraria islandica 12 .
The Carbohydrate-Active Enzymes database (CAZy; http://www.cazy.org) provides a sequence-based family classification linking the sequence to the specificity and 3D structure of the enzymes 20 , in which most bacterial laminarinases and β-1,3-1,4-glucanases belong to the GH16 family. Some members of this family have been characterized, and crystal structures have been analyzed for enzymes from Thermotoga maritima 21 , hyperthermophile Pyrococcus furiosus 17,22 , Nocardiopsis sp. strain F96 23 , and Zobellia galactanivorans 24 . It is not easy to classify GH16 family enzymes aside from a classical sandwich-like β-jelly roll fold composed of two antiparallel β-sheets packed against each other 25,26 . Determining the crystal structures of laminarinases/glucanases provides useful examples of versatile yet specific protein-carbohydrate interactions, which could not be precisely predicted by sequence alignment alone.
Flavobacterium sp. UMI-01 is a novel bacterium recently identified from decayed brown algae; it has been grown in medium containing either alginate or laminarin as a sole carbon source 27 . Flavobacteria contains a high abundance of family GH16 laminarinases, underlying the environmental importance of decomposing algal biomass. Genomic sequence analysis revealed the presence of a candidate gene for GH16 β-1,3-laminarinase, which was designated as ULam111. This enzyme consists of 235 residues and has a molecular mass of approximately 27 kDa. Herein, we determined the crystal structure of ULam111 at 1.60-Å resolution. The structure revealed new insights concerning the thermostability and substrate recognition for the degradation of polysaccharides by the laminarinase family, which may be suitable for beer brewing and feed additives.

Results and Discussion
Overall structure of ULam111. The crystal structure of ULam111 was refined to 1.60-Å resolution. Two protein molecules and 670 water molecules were observed in the asymmetric unit of the ULam111 crystal. The R and R free values of the final model were 17.6 and 20.6%, respectively. In the Ramachandran plot, 97.4% of residues were included in the favored region, and 2.6% were in the allowed region. The refinement statistics are summarized in Table 1.
Characteristics of the active site. The catalytic residues included Glu118, Asp120, Glu123, and His137 ( Fig. 1b). On the basis of the conserved catalytic mechanism 21, 28 , Glu118 was assumed to be the nucleophile that directly attacks C1 of the sugar ring, and Glu123 was hypothesized to function as the proton donor. The interaction of Asp120 with His137 via a hydrogen bond was considered to stabilize the substrate.

Comparison of the ULam111 structure with other family enzymes. A structure homology search
using Dali 29 showed that the overall structure of ULam111 is highly similar to previously reported enzymes, including TmLam from Thermotoga maritima (PDB code, 3AZZ; Z-score, 34 Although ULam111 shared >43% sequence identity with TmLam, BglF and PfLamA, structural determination by molecular replacement using a structure model of TmLam was not successful, indicating that there may be conformational differences between ULam111 and some family enzymes. The superposition of ULam111 with TmLam, BglF, and PfLamA showed that these proteins contained the common β-jelly roll folds and a straight groove for substrate binding (Fig. 2). ULam111 shared the bulging loop of β2-β3 with TmLam and PfLamA. The comparison of ULam111 and ZgLamA showed that ULam111 lacked the flexible loop at the entrance to regulate the recognition of β-1,3-1,4-linked substrates. However, the disulfide bond of Cys31-Cys34 only existed in ULam111 (Figs 2b and S1). This loop may partly regulate substrate recognition, as it was located above the substrate-binding site and slightly decreased the degrading activity compared with the wild-type enzyme (Fig. 2c). The β-strands of β4 and β5 exhibiting a different spatial conformation were considered to serve as a structural feature because they were on the opposite side of the catalytic cleft.
Comparison of substrate binding site. Some β-strands of Sheet A twisted somewhat to form an electronegative-rich cleft where the substrates were implicated to bind (Fig. S2). It is obvious that ULam111 binds β-1,3-linked or β-1,3-1,4-linked substrates in a straight catalytic cleft and cleaves the glucosidic bond in an open form because there are no additional loops covering the active site (Fig. 2a). The sugar ring may form hydrogen bonds at the negatively charged binding site of ULam111 with hydrophilic residues, such as Asn33, Gly37, Asn38, and Arg71, and hydrophobic-stacking interactions with Trp98, Trp102, and Trp113, whereas most of the residues are conserved in the GH16 family enzymes (Fig. 3a). Notably, Phe212, located above the catalytic residues of positive substrate-binding site + 1, was not conserved (tryptophan in GH16 family enzymes) and showed different spatial orientations than other enzymes (Fig. 3b). It was reported to regulate substrate recognition in the release of glucose in carbohydrate hydrolysis 21 . Thermostability of ULam111. ULam111 contained a special calcium-binding motif on the opposite side of the β-sheet from the catalytic cleft (Fig. 4), which was composed of His12, Asn14, Gly54, and Asp229. The calcium ion was coordinated to the carbonyl and carboxylate O atoms of Asp229, the carbonyl O atom of Gly54, and the carbonyl O atom of Asn14. This motif is known to increase thermostability in GH16 family enzymes 30 . However, the calcium-binding motif is not conserved in this family. The residues of Asn14 and His12 in ULam111 were positioned near glutamic acid and aspartic acid in other enzymes (Fig. 4). The H12E/N14D substitution could increase the degradative activity toward laminarin and thermostability from 15 °C to 40 °C (Fig. 5). It is interesting that the calcium ion could significantly increase the activity and thermostability of wild-type and   Therefore, both the mutant and calcium ion were required to enhance thermostability.
ULam111 formed one intramolecular disulfide bond at Cys31-Cys34 in the loop between β2 and β3, which was considered to serve as a key structural feature in stabilizing the loop above the catalytic cleft of ULam111. At the same position, TmLam has the histidine and proline residues, while PfLamA adopts the isoleucine and proline residues (Fig. S1). The C31S/C34S mutant showed decreased thermostability in our detectable temperature range (especially from 15 to 30 °C). Therefore, the loop including Cys31-Cys34 played a pivotal role in regulating thermostability.
Phe212 (tryptophan in the GH16 family enzymes) showed a different spatial orientation than that of other enzymes (Fig. 3b). Trp232 in TmLam may form hydrophobic interactions with a flexible GASIG loop in the closed form, contributing to regulation of exo-cleavage activity and preferred release of glucose in carbohydrate hydrolysis 21 . Trp270 in PfLamA showed a dual spatial orientation to regulate substrate specificity. To investigate the role of Phe212 in substrate recognition, we mutated Phe212 to tryptophan. The result showed that F212W mutant of ULam111 led to the decreased degrading activity (V max and k cat /K m ) compared with the wild-type enzyme toward all of the tested substrates ( Table 2). The decreased ratio of k cat /K m toward curdlan (β-1,3) was the  Table 2. Kinetic Parameters of ULam111 Wild Type and F212W on Several Substrates. All assays were repeated three times, and the data are shown as mean ± S.D. highest among those toward all tested substrates. F212W mutant reduced the degrading activity toward laminarin to glucose within the initial 30 min (Fig. S4), which implied that phenylalanine has less steric hindrance than tryptophan and was the optimal residue for substrate recognition.

Materials and Methods
Cloning, expression, and purification. A gene encoding the candidate β-1,3-glucanase, termed ULam111 (GenBank accession no. LC202090), was found in the draft genome sequence of the Flavobacterium sp. strain UMI-01 as previously analyzed 27 . Genomic DNA was prepared from strain UMI-01 as previously described 27 and was used as a template for genomic polymerase chain reaction (PCR) with the Q5 High-Fidelity DNA Polymerase DNA polymerase (New England Biolabs, Ipswich, MA) and a pair of specific primers, F1 (5′-GTTCGGCTAAAACACTCGAAGCTG-3′) and F2 (5′-ATCAAATGCAATCTAAATTCCGTG-3′), for the 5′-and 3′-untranslated regions, respectively. PCR was conducted using temperature settings of 95 °C for 5 min followed by 30 cycles of 95 °C for 15 s, 55 °C for 15 s, and 72 °C for 45 s. The final step for extension was 72 °C for 2 min. Amplified DNA was subcloned into the pTac-1 vector (BioDynamics, Tokyo, Japan) and sequenced with a genetic analyzer 3130xl (Applied Biosystems, Foster City, CA). DNA encoding residues 18-251 were amplified with a primer set of F2 (5′-AGGTAATACACCATGACTAAAGGAAAAAAACTGGT-3′) and R2 (5′-CACCTCCACCGGATCCTTGATACACCTTAATATAGTC), and PCR was conducted as described above. The ULam111 gene was cloned into a modified pCold I vector (Takara Bio, Shiga, Japan) as previously described 27 between the Nco I and BamH I sites using an In-fusion HD cloning kit (Clontech, Mountain View, CA). The expressed ULam111 protein was fused to a modified octahistidine (His 8 ) tag (HHHHHHHH) at the C-terminus connected by a linker of GSGGGGGGGG. The plasmid vector was transformed into Escherichia coli Rosetta gami 2(DE3) (Merck Millipore, Billerica, MA) for protein expression.
The E. coli transformants were incubated at 37 °C until the optimal density at 600 nm (OD 600 ) reached 0.6-0.8. Isopropyl β-D-1-thiogalactopyranoside (IPTG) was added at a final concentration of 0.5 mM, and the culture was further incubated for 16 h at 15 °C. After harvesting, the cells were disrupted by sonication in the resuspending buffer containing 50 mM sodium phosphate (pH 7.8), 10 mM imidazole, 100 mM NaCl and 1% Triton X-100. Cell debris was removed by centrifugation at 40,000 g. ULam111 was trapped on TALON resin (Clontech), which is a key purification step because ULam111 was partly trapped on Ni-NTA Superflow resin (QIAGEN, Hilden, Germany) in the resuspending buffer. After washing with buffer 1 [50 mM sodium phosphate (pH 7.8), 20 mM imidazole, 300 mM NaCl] and buffer 2 [20 mM Tris-HCl (pH 7.5), 100 mM NaCl], the His 8 -tagged protein was eluted with elution buffer [20 mM Tris-HCl (pH 7.5), 300 mM imidazole, 300 mM NaCl]. The eluted solution was dialyzed in buffer 2 and further purified using a Resource Q (GE Healthcare, Chicago, IL) column. The fractions containing purified ULam111 were dialyzed against buffer 2 and concentrated to 10 mg ml −1 for crystallization using a Vivaspin-20 (10 000 MWCO).
To obtain the selenomethionine-labeled ULam111 (ULam111 SeMet ), cells were transferred into M9 medium supplemented with 50 mg ml −1 selenomethionine (SeMet), when the OD 600 reached 0.5 33,34 . The expression and purification of ULam111 SeMet were the same as the native protein described above.
Site-directed mutagenesis was performed by PCR with a QuikChange kit (Stratagene, La Jolla, CA) and pCold-ULam111 plasmid as a template. The mutations were confirmed by DNA sequencing. ULam111 mutants were expressed and purified using the method described above for wild-type ULam111.
Crystallization and data collection. Crystallization experiments were performed using the sitting-drop vapor diffusion method at 20 °C. Crystallization drops were prepared by mixing 1 μL of the protein solution with 1 μL of a variety of reservoir solutions. Crystals of native ULam111 were obtained with a reservoir solution containing 0.1 M sodium citrate (pH 5.6), 30% (w/v) polyethylene glycol (PEG) 4000 and 0.2 M ammonium acetate. Crystals of ULam111 SeMet were obtained with a reservoir solution containing 0.1 M sodium acetate (pH 4.6), 30% (w/v) polyethylene glycol monomethyl ether (PEG MME 2000), and 0.2 M ammonium sulfate.
The X-ray diffraction data for the native ULam111 and ULam111 SeMet crystals were collected on the BL-5A beamline at the Photon Factory (Tsukuba, Japan). X-ray diffraction data were collected at a resolution of 1.60 Å for native ULam111 and a resolution of 2.2 Å for ULam111 SeMet . All diffraction data were indexed, integrated, and scaled with the XDS program 35 . The data collection statistics are summarized in Table 1.
Structural determination. The initial phase of ULam111 SeMet was obtained using single anomalous dispersion (SAD) 36 . After selenium atom search and phase calculations with PHENIX AutoSol Wizard in the PHENIX program suite 37 , model building was automatically carried out with PHENIX AutoBuild Wizard 37 . Manual rebuilding and refinement of ULam111 SeMet were performed with COOT 38 and PHENIX.REFINE 35 . The structure of native ULam111 was determined by the molecular replacement method using the MOLREP program 39,40 with the ULam111 SeMet structure as the initial model. Manual rebuilding and refinement of native ULam111 was performed with COOT 36 and PHENIX.REFINE 37 , respectively. The geometry of the final structure was evaluated with the program Rampage 41 . The coordinates of ULam111 have been deposited into the Protein Data Bank (PDB) with the accession number (5WUT).

Structural analysis.
Structural analysis was carried out using a set of programs: Dali 29 was used for the search of similar structures from the database 42 , DaliLite 43 was used for the superposition of molecules, ESpript 44 was used for the preparation of alignment figures, and Pymol (http://pymol.sourceforge.net/) for the depiction of structures.
Kinetic studies of ULam111 wild-type or F212W were performed in a solution containing 10 mM NaPi (pH 6.0), 100 mM NaCl, 0.1 mg ml −1 BSA, 1.0 mM CaCl 2 , 1 μg ml −1 wild-type or F212W enzymes, and various concentrations of substrates at 30 °C. Kinetic parameters, K m , V max , and k cat were calculated using the Michaelis-Menten equation with GraphPad Prism 7.0 (GraphPad software, La Jolla, CA) by employing nonlinear regression.