Characterization of an Unknown Region Linked to the Glycoside Hydrolase Family 17 β-1,3-Glucanase of Vibrio vulnificus Reveals a Novel Glucan-Binding Domain

The glycoside hydrolase family 17 β-1,3-glucanase of Vibrio vulnificus (VvGH17) has two unknown regions in the N- and C-termini. Here, we characterized these domains by preparing mutant enzymes. VvGH17 demonstrated hydrolytic activity of β-(1→3)-glucan, mainly producing laminaribiose, but not of β-(1→3)/β-(1→4)-glucan. The C-terminal-truncated mutants (ΔC466 and ΔC441) showed decreased activity, approximately one-third of that of the WT, and ΔC415 lost almost all activity. An analysis using affinity gel containing laminarin or barley β-glucan revealed a shift in the mobility of the ΔC466, ΔC441, and ΔC415 mutants compared to the WT. Tryptophan residues showed a strong affinity for carbohydrates. Three of four point-mutations of the tryptophan in the C-terminus (W472A, W499A, and W542A) showed a reduction in binding ability to laminarin and barley β-glucan. The C-terminus was predicted to have a β-sandwich structure, and three tryptophan residues (Trp472, Trp499, and Trp542) constituted a putative substrate-binding cave. Linker and substrate-binding functions were assigned to the C-terminus. The N-terminal-truncated mutants also showed decreased activity. The WT formed a trimer, while the N-terminal truncations formed monomers, indicating that the N-terminus contributed to the multimeric form of VvGH17. The results of this study are useful for understanding the structure and the function of GH17 β-1,3-glucanases.


Introduction
Marine algae convert marine carbon into algal polysaccharides by photosynthesis. Algal polysaccharides are made up of a variety of glycans. The recycling of algal polysaccharides into carbon dioxide gives us a better understanding of the global marine carbon cycle [1]. Recently, the involvement of marine bacteria in this cycle has been gradually revealed [2]. Laminarin is a major glucose polymer found in marine environments [3]. Therefore, an understanding of the mechanisms underlying the degradation of large algal polysaccharides by enzymes and their modules is useful in order to produce sustainable and renewable raw materials for use in valuable compounds, feeds, and fuels [4].
The CAZy database provides the taxonomic distribution of the GH17 family (cazy. org/IMG/krona/GH17_krona.html, accessed on 23 February 2022), revealing that a large number of bacterial enzymes are found within the phylum Proteobacteria (recently renamed Pseudomonadota). Within Proteobacteria, GH17 enzymes are commonly found within the genus Pseudomonas, with its diverse members and metabolism. While many Vibrio species, which belong to the class Gammaproteobacteria, have GH16 enzymes, a limited number possess enzymes of the GH17 family. The genome of V. vulnificus has been sequenced and annotated, and sequence analysis has revealed that one GH16 enzyme (VvGH16) and one GH17 enzyme (VvGH17) exist adjacently in the genome. One GH3 enzyme is located close to the other two enzymes. On the other hand, three GH16 (VbGH16A, VbGH16B, and VbGH16C) and one GH17 (VbGH17A) enzyme of Vibrio breoganii 1C10 have been characterized [25].
Several CAZymes have various domains in addition to the catalytic domain, including carbohydrate-binding modules (CBMs) [26][27][28]. These domains are involved in carbohydrate binding. Tryptophan is an important amino acid residue in carbohydrate binding [26]. The VvGH17 C-terminus has several tryptophan residues. Therefore, we predicted that this region may have functions, such as carbohydrate binding, that can increase the catalytic efficiency or specificity. In this study, we characterized GH17 β-1,3-glucanase of V. vulnificus to clarify the unknown region of the protein and found that the N-and C-terminal regions were affiliated with the assembly of monomeric subunits into the multimeric form and the affinity for the substrate, respectively.

Bioinformatic Analysis of VvGH17
VvGH17 is composed of 615 amino acids (AAs) comprising 1-22 AAs as signal peptides, 23-86 AAs as an unknown N-terminal region (Uk-N), 87-415 AAs as the GH17 domain, and 416-615 AAs as an unknown C-terminal region (Uk-C) (Figure 1a). Secondary structure prediction showed that Uk-N had a random coil structure, while Uk-C was composed of a β-sheet structure. The structure of VvGH17 was predicted using AlphaFold2 [29] ( Figure 1b). The GH17 domain and C-terminus of Uk-C were predicted to have a (β/α) 8 barrel structure and a β-sandwich structure, respectively. It was expected that the Uk-C structure possessed some function. Therefore, we attempted to characterize the impact of Uk-N and Uk-C on the catabolic properties of the enzymes. Colors are related to the truncated mutation region in this study.

Biochemical Properties of VvGH17
VvGH17 was produced in an Escherichia coli expression system and isolated using a TALON affinity resin (Takara Bio, Otsu, Japan). The purified enzyme (20 mg) was obtained from one liter of medium and showed a single band of approximately 63 kDa on SDS-PAGE. The optimal temperature and pH were 50 • C and 5.0-6.5, respectively (Figure 2a,b). The temperature required for the half inactivation of the hydrolysis activity of VvGH17 at 30 min was 47 • C ( Figure 2c). The activity of VvGH17 decreased by approximately 80% in the presence of 0.5 M NaCl and retained the activity up to 4.0 M (Figure 2d). The specific activity of VvGH17 in optimal conditions was 65.5 U/mg. VvGH17 hydrolyzed curdlan (insoluble β-(1→3)-glucan) and laminarin, mainly producing laminaribiose, glucose, and laminaritriose; it did not hydrolyze barley β-glucan, which as β-(1→3)and β-(1→4)linkages ( Figure 2e). The products of laminarin hydrolysis by VvGH17 were monitored from 0 to 60 min using gel filtration; the results revealed that VvGH17 hydrolyzed laminarin via an endolytic mechanism (Figure 2f). The effect of temperature on VvGH17 stability. A mixture containing 50 mM MES buffer (pH 6.0) and 0.2 mg/mL VvGH17 was incubated at the indicated temperature for 30 min and placed on ice for 10 min. Then, the enzyme activity was assayed in a mixture containing 50 mM MES buffer (pH 6.0), 1% (w/v) laminarin, and 0.02 mg/mL VvGH17 at 45 • C for 10 min. (d) The effect of NaCl on VvGH17 activity. The enzyme reaction was conducted in a mixture containing 50 mM MES buffer (pH 6.0), 1% (w/v) laminarin, 0.02 mg/mL VvGH17, and 0-4.0 M NaCl at 45 • C for 10 min. (e) Thin layer chromatography (TLC) analyses of the hydrolysis products obtained using VvGH17. One microliter of each reaction mixture was applied for TLC analysis. Mk-marker of glucose and laminaripentaose; '−'-without VvGH17; '+'-with VvGH17. 1 -curdlan; 2 -laminarin; 3 -β-glucan. (f) Gel filtration chromatography analysis for the hydrolysis of laminarin by VvGH17.

C-terminal-Truncated Mutant
C-terminal-truncated mutants of VvGH17 were constructed for the characterization of Uk-C. The position from 87-415 AA was demonstrated as the conserved domain of GH17. Therefore, Uk-C was defined as the position from 416-615 AA in VvGH17, and the three C-terminal-truncated mutants (∆C466, ∆C441, and ∆C415) were constructed ( Figure 3a). Recombinant proteins of the three mutants were successfully expressed, and we evaluated the enzyme kinetics ( Figure 3b, Table 1). The enzyme kinetics k cat /K m of the WT toward laminarin was 93.0 mM −1 s −1 , and 32.8 and 30.5 mM −1 s −1 for ∆C466 and ∆C441, respectively, which are approximately one-third of the k cat /K m values in the WT. The k cat /K m values of ∆C415 were less than 2% of those in the WT. To confirm whether the loss of activity in ∆C415 was derived from folding, the secondary structure was compared using circular dichroism (CD) spectroscopy ( Figure 3c). The difference in the CD spectrum (deg cm 2 dmol −1 ) between 210 and 230 nm may be a result of the deletion of the C-terminus in VvGH17. From these results, the truncation of the C-terminus in Uk-C (AA 442-615) resulted in a decrease in the catalytic efficiency of VvGH17, and the truncation of the whole Uk-C (416-615 AA) caused the loss of the majority of its activity, suggesting that this is an essential region.  Table 1. Enzyme kinetics of VvGH17 and mutants using laminarin as a substrate.

Affinity Gel Analysis of C-Terminal-Truncated Mutants of VvGH17
The truncation of the VvGH17 C-terminus revealed that this region affects the enzyme kinetics (k cat and K m ). This indicates that Uk-C has the potential for carbohydrate binding. To investigate the Uk-C function further, affinity gel analysis was performed ( Figure 4). The WT and three C-terminal-truncated mutants showed two bands with and without substrates. The two bands were confirmed as monomers and oligomers (trimer) of the enzyme, as discussed in Section 2.7. Bovine serum albumin (BSA) was used as a marker protein for the mobility shift assay. No affinity toward curdlan was found in the tested enzymes, compared to the gel without substrate. The mobility of the WT monomer was clearly shifted from below the BSA band (without substrate) to upper the BSA band in the gels, confirming its affinity toward laminarin and β-glucan.

Affinity Gel Analysis of Uk-C and Point Mutants of VvGH17
The affinity of Uk-C and the C-terminus of VvGH17 toward laminarin and β-glucan was revealed. To confirm the important amino acids for substrate binding, point mutants of Uk-C and VvGH17 were constructed. Affinity toward substrates was also evaluated by the mobility as compared with BSA ( Figure 5). The mobility of Uk-C in the gel containing laminarin and β-glucan was decreased compared to the gel without substrate. This indicated that Uk-C had a binding ability for laminarin and β-glucan. Tryptophan is an important amino acid for carbohydrate binding [26]. Therefore, we mutated four tryptophans in Uk-C to alanines (W472A, W499A, W542A, and W567A), and the affinity was evaluated by mobility shift assays. The WT and four mutants showed the same mobilities without a substrate. The mobilities of W472A, W499A, and W542A in the gel containing laminarin and β-glucan differed from those of the WT and W567A. The decreased mobility of W472A, W499A, and W542A indicated reduced binding ability, suggesting that the three tryptophans are essential amino acids for substrate binding.

Prediction of Uk-C Structure and Function
We attempted to clarify the relationship between the predicted three-dimensional structure of Uk-C (AA 441-615 of VvGH17) and the binding ability of the mutants. The predicted Uk-C had two domains: the N-terminus of Uk-C (AA 416-462 of VvGH17) was predicted to be a linker between GH17 and the binding region, and the C-terminus of Uk-C (AA 463-615 of VvGH17) was predicted to be a β-sandwich structure with a possible carbohydrate-binding ability (Figure 1b). Three tryptophan residues (Trp472, Trp499, and Trp542) were located in the putative substrate binding region in the β-sandwich structure ( Figure 6). On the other hand, the predicted structure suggested that Trp567 was located outside of the putative substrate binding region. The results of the mutation experiments agreed with the predicted structure.

N-Terminal-Truncated Mutants of VvGH17
The catalytic domain of VvGH17 was mapped from AA 87 to 415, and AA 1-22 of VvGH17 were predicted as a signal peptide. The function of the N-terminus in VvGH17 (AA 23-86) was unclear. Therefore, we constructed two N-terminal-truncated mutants (∆N50 and ∆N65) and evaluated their activity (Figure 7). The k cat /K m values of ∆N50 and ∆N65 toward laminarin were 67.5 and 24.7 mM −1 s −1 , respectively. The predicted three-dimensional structure of VvGH17 showed that the region of AA 51-65 was composed of the bottom of the (β/α) 8 barrel structure (Figure 1b). Consequently, the loss of this region in the ∆N65 mutant led to structural instability, resulting in decreased catalytic efficiency (K cat /k m ) ( Table 1). Two bands from the WT and C-terminal-truncated mutants were observed in native PAGE, as shown in Section 2.4. To confirm the assembly of monomeric subunits into the multimeric form, blue native PAGE was performed for the WT and N-and C-terminaltruncated mutants (Figure 7c). The WT and ∆C466 mutant clearly showed two bands (monomers and putative trimers from the molecular mass), while ∆N50 and ∆N65 showed a single band corresponding to the monomer.
VvGH17 produced laminaribiose as the main product regardless of soluble and insoluble β-(1→3)-glucans, showing its potential for oligosaccharide production. The amino acid identity between VvGH17 and VbGH17A from V. breoganii 1C10 was low (42% identity). VbGH17A contains a signal peptide, the GH17 catalytic domain, and an unknown region of the C-terminus from amino acid (AA) 411 to 634 [18]. The catalytic domains of VvGH17 and VbGH17A shared 56% identity; however, the identity between the C-termini was low (21%). The hydrolysis products of VbGH17A were oligosaccharides, which were larger than a degree of polymerization (DP) of 4. V. breoganii 1C10 has four endo-type β-(1→3)-glucanases, which presumably show synergic activity during the hydrolysis of β-(1→3)-glucans. On the other hand, V. vulnificus has two endo-type β-(1→3)-glucanases. We expected several enzymes; however, there was only GH16, which had 57% identity with VbGH16A and produced mainly DP 3 and 4. The GH3 enzymes of Vibrio sp. have been shown to have activity toward laminaribiose [30]. Therefore, V. vulnificus may metabolize β-(1→3)-glucan by producing small DP oligosaccharides using two endo-type enzymes and then hydrolyzing them using the GH3 enzyme.

Bioinformatic Analysis of VvGH17
The GH17 gene from Vibrio vulnificus (hypothetical protein AOT11_01225) was obtained from GenBank (accession no. ASM98089.1). The putative conserved domain was searched using BLASTP [41]. The signal peptide was predicted using the SignalP 4.1 server [42]. Secondary structure prediction was performed using the PSIPRED server [43] and the structure was predicted by AlphaFold2 [29]. Homolog proteins of the C-terminus of VvGH17 (416-615 AA) were searched by BLASTP with the standard algorithm, excluding Vibrio vulnificus (taxid: 672); we also removed query covers of less than 40%. A phylogenetic tree was constructed by pairwise sequence alignment.

Construction, Expression, and Purification of VvGH17
The expression plasmid of the gene putatively encoding β-(1→3)-glucanase (VvGH17) was constructed as follows: a codon-optimized mature Vvgh17 gene was synthesized (Eurofins Genomics) for expression in E. coli harboring NdeI and HindIII sites at 5 and 3 , respectively. Then, the Vvgh17 gene was cloned into the NdeI-HindIII site of pET28a to construct an expression vector of pET28a(VvGH17). The recombinant protein was produced in E. coli BL21-RIL (DE3) cells (Agilent Technologies, Palo Alto, CA, USA) harboring pET28a(VvGH17) and was purified as previously described [44]. The protein concentrations were determined by absorbance at 280 nm using the molar extinction coefficients for VvGH17 [45].

VvGH17 Standard Activity Assay
VvGH17 activity was determined at 45 • C for 10 min with an appropriate amount of enzyme, 1% (w/v) laminarin, and 50 mM 2-morpholinoethanesulfonic acid (MES; pH 6.0). The amount of reducing sugars was determined using the dinitrosalicylic acid method (DNS) [46]. One unit of activity was defined as the amount of enzyme that liberated reducing sugars equivalent to 1.0 µmol glucose per minute. The optimal temperature of VvGH17 was measured as follows: a reaction mixture containing 1% (w/v) laminarin and 50 mM MES (pH 6.0) was incubated at 22-70 • C for 10 min. The optimal pH of VvGH17 was measured as follows: a reaction mixture containing 1% (w/v) laminarin and 100 mM Britton-Robinson buffer (a mixture containing sodium acetate buffer, sodium phosphate buffer, and glycine-NaOH buffer; pH 4.0-10.0) was incubated at 45 • C for 10 min. Temperature stability was determined by measuring the residual activity after incubation in 50 mM MES (pH 6.0) at 30-57 • C for 30 min. The effect of NaCl was determined using a mixture containing 1% (w/v) laminarin, 50 mM MES (pH 6.0), and 0-4.0 M NaCl at 45 • C for 10 min. The V max and K m with laminarin (0.5-40 mg/mL) were determined by the standard Michaelis-Menten equation using nonlinear regression (Origin Software, Lightstone Corp., Tokyo, Japan). All the activity assays were performed in triplicate.

Analysis of Hydrolysis Products by TLC and Gel Filtration Chromatography
The products of VvGH17 hydrolysis were analyzed by TLC using a silica gel 60 plate (Merck). The substrates (curdlan, laminarin, and β-glucan: 10 mg/mL) were hydrolyzed with 0.01 U/mL of VvGH17 for 24 h, and the reaction was terminated by heating at 100 • C for 10 min. The hydrolysis products (1 µL) were developed in ethyl acetate, acetic acid, and water (2:2:1, v/v/v); sugars were detected by spraying a solution of 10% (v/v) sulfuric acid in ethanol and then heating at 100 • C for 10 min.
The distribution of the hydrolysis products of laminarin was analyzed using highperformance liquid chromatography (HPLC) with a Superdex Peptide 10/300 GL column (GE Healthcare UK Ltd., Little Chalfont, UK) and a Corona Charged Aerosol Detector (Thermo Scientific Inc., Chelmsford, MA, USA). Laminarin (10 mg/mL) was hydrolyzed with 0.01 U/mL of VvGH17 from 0 to 60 min, and the reaction was terminated by heating at 100 • C for 10 min. The samples were eluted using water with a flow rate of 0.3 mL/min.

CD Spectroscopy
The secondary structures of VvGH17 were determined by CD spectroscopy using a J-720WI spectrometer (Jasco Corp. Tokyo, Japan). The proteins were dissolved at a final concentration of 0.1 mg/mL in 50 mM MES buffer (pH 6.0). The spectra were acquired at 37 • C using a 0.2 cm cuvette. The molar ellipticities (per residue) were calculated using the equation [θ] = 100(θ)/(lcN), where [θ] is the molar ellipticity per residue, (θ) is the observed ellipticity in degrees, l is the optical path length in centimeters, c is the molar concentration of the protein, and N is the number of residues in the protein.

Polyacrylamide Gel Electrophoresis (PAGE) Analysis
The assays for the binding activity of the proteins were performed by affinity gel electrophoresis, according to the procedure described by Zhang et al. [47]. A stacking gel containing 3 wt% polyacrylamide in 1.5 M Tris-HCl buffer (pH 8.3), a native gel with 12 wt% polyacrylamide containing 0.1 wt% polysaccharides (curdlan, laminarin, and barley β-glucan), and a control gel without polysaccharides were prepared. Each protein (1 µg) was loaded onto the gel, and the gels were electrophoresed at 4 • C and 100 V for 3 h. The gels were then stained with Coomassie brilliant blue G-250 for protein visualization.

Conclusions
In this study, we characterized the unknown domains of the GH17 β-(1→3)-glucanase of V. vulnificus. The WT formed a trimer, but the N-terminal truncations formed monomers. Therefore, the N-terminus contributes to the assembly of monomeric subunits into the multimeric form of VvGH17. The C-terminal region showed an affinity for β-(1→3)-glucan and β-(1→3)/β-(1→4)-glucan. The C-terminus was predicted to have a β-sandwich structure, and three tryptophan residues (Trp472, Trp499, and Trp542) were located at the substrate binding site using mutational analysis. A BLAST search revealed that the C-terminal region of GH17 was conserved among Gammaproteobacteria. The results of this study are useful for understanding bacterial GH17 enzymes and oligosaccharide preparation.