Novel Endotype Xanthanase from Xanthan-Degrading Microbacterium sp. Strain XT11

This work characterized a novel endotype xanthanase, MiXen, and elucidated that the C-terminal carbohydrate-binding module of MiXen could drastically enhance the hydrolysis activity of the enzyme toward highly ordered xanthan. Both the sequence and structural analysis demonstrated that the catalytic domain and carbohydrate-binding module of MiXen belong to the novel branch of the GH9 family and CBMs, respectively. This xanthan cleaver can help further reveal the enzymolysis mechanism of xanthan and provide an efficient tool for the production of molecular modified xanthan with new physicochemical and physiological functions.

ABSTRACT

Under general aqueous conditions, xanthan appears in an ordered conformation, which makes its backbone largely resistant to degradation by known cellulases. Therefore, the xanthan degradation mechanism is still unclear because of the lack of an efficient hydrolase. Here, we report the catalytic properties of MiXen, a xanthan-degrading enzyme identified from the genus Microbacterium. MiXen is a 952-amino-acid protein that is unique to strain XT11. Both the sequence and structural features suggested that MiXen belongs to a new branch of the GH9 family and has a multimodular structure in which a catalytic (α/α)₆ barrel is flanked by an N-terminal Ig-like domain and by a C-terminal domain that has very few homologues in sequence databases and functions as a carbohydrate-binding module (CBM). Based on circular dichroism, shear-dependent viscosity, and reducing sugar and gel permeation chromatography analysis, we demonstrated that recombinant MiXen efficiently and randomly cleaved glucosidic bonds within the highly ordered xanthan substrate. A MiXen mutant free of the C-terminal CBM domain partially lost its xanthan-hydrolyzing ability because of decreased affinity toward xanthan, indicating the CBM domain assisted MiXen in hydrolyzing highly ordered xanthan via recognizing and binding to the substrate. Furthermore, side chain substituents and the terminal mannosyl residue significantly influenced the activity of MiXen via the formation of barriers to enzymolysis. Overall, the results of this study provide insight into the hydrolysis mechanism and enzymatic properties of a novel endotype xanthanase that will benefit future applications.

IMPORTANCE This work characterized a novel endotype xanthanase, MiXen, and elucidated that the C-terminal carbohydrate-binding module of MiXen could drastically enhance the hydrolysis activity of the enzyme toward highly ordered xanthan. Both the sequence and structural analysis demonstrated that the catalytic domain and carbohydrate-binding module of MiXen belong to the novel branch of the GH9 family and CBMs, respectively. This xanthan cleaver can help further reveal the enzymolysis mechanism of xanthan and provide an efficient tool for the production of molecular modified xanthan with new physicochemical and physiological functions.

INTRODUCTION

Xanthan, an anionic heteropolysaccharide with a molecular mass ranging from 1 × 10⁶ to 7 × 10⁶ Da, is produced by the plant-pathogenic bacterium Xanthomonas campestris pv. campestris (1). Xanthan contains a cellulosic chain as a backbone and a linear mannosyl-glucuronyl-mannosyl trisaccharide as a side chain attached at the C-3 position on the alternate glucosyl residue of the main chain (2, 3). The inner and terminal mannosyl residues of the side chains are often acetylated and pyruvylated, respectively, depending on both the xanthan-producing strains and the culture conditions (2). The complex structure endows xanthan with superior rheological properties, such as pseudoplasticity, high viscosity, and tolerance toward a wide range of pHs and temperatures, leading to its widespread application as the thickener and stabilizer in the food, pharmaceutical, and oil industries (4, 5).

To date, molecular modified xanthan with novel physicochemical and physiological functions has been sought for use in new applications. Although genetically engineered Xanthomonas mutants have been reported to produce variant xanthan products, their production levels are far from those needed for practical application (1, 6, 7). Preparation of such modified xanthan by chemical or physical degradation of the xanthan backbone is difficult because of the complex structure of the polymer (4). In addition, chemical or physical modification of xanthan is nonspecific and will also cause degradation of the xanthan side chains (8). Thus, enzyme-assisted methods that specifically attack the xanthan backbone while leaving the side chains intact seem to be a promising method for preparation of molecularly designed xanthan.

In aqueous solution, native xanthan exists in a double-stranded helical conformation along with an order-disorder transition (8). Previous studies have demonstrated that xanthan in a highly disordered conformation can be hydrolyzed by cellulases (9, 10). However, the accessibility of xanthan by enzymes could be dramatically reduced when the ordered fraction in xanthan increases (8, 11). It should be noted that the conformation of xanthan is highly dependent on changes in temperature, ionic strength, and molecular composition of xanthan (12, 13), particularly with respect to the presence of acetyl and/or pyruvate groups (14, 15), and completely disordered xanthan can only be obtained in solutions with very low ionic strength and at a high temperature (8), which is difficult to achieve under industrial conditions. Furthermore, considering that enzyme activities can also be affected by temperature and ionic strength, low enzyme activity may occur under conditions in which xanthan appears in a completely disordered conformation.

Currently, little is known about the hydrolysis mechanism of individual enzymes on the xanthan backbone and enzyme resistance of highly ordered xanthan remains a bottleneck for the production of modified xanthan. In this study, an endotype xanthanase, MiXen, from the xanthan-degrading bacterium Microbacterium sp. strain XT11, was cloned and heterologously expressed. The sequence/structure homology, biochemical characteristics, and product profiles of MiXen were studied, revealing novel hydrolysis properties concerning the high substrate specificity and hydrolysis efficiency of MiXen against xanthan with a highly ordered conformation.

RESULTS

Sequence analysis and homology modeling of MiXen reveals a multidomain structure.

Previously, an aerobic Gram-positive bacterium, Microbacterium sp. XT11, with the ability to degrade xanthan, was isolated from a soil sample collected from the Dalian Botanical Garden in China (16), and the complete genome sequence of strain XT11 was determined and analyzed (17). Within the genomic sequence of Microbacterium sp. XT11, the open reading frame encoding MiXen (GenBank accession number ALX66163.1) is 2,856 bp in length and has a GC content of 67.3%. The predicted molecular mass of MiXen is 101.1 kDa, and its calculated isoelectric point is 4.4. SignalP (version 4.1) analysis showed that the predicted signal peptide of MiXen was composed of 33 amino acid residues (Met¹ to Ala³³). Functional domains of MiXen were predicted by the InterProScan web server. MiXen consisted of a cellulase N-terminal Ig-like domain (Ala³⁴ to Glu¹²⁴), a putative family 9 glycoside hydrolase (GH9) catalytic domain (Asp¹²⁵ to Ser⁵⁸⁰), and a suspected carbohydrate-binding module (CBM) at the C terminus (Arg⁷⁰⁰ to Glu⁸⁴⁸).

Previous reports confirmed that structures of the catalytic domains of GH family 9 enzymes are characterized by an (α/α)₆-barrel fold with three acidic active-site residues (two aspartate residues and one glutamate residue) (18). The C terminus-free MiXen showed highest identity scores with three characterized GH9 enzymes, including GH9 endoglucanase LC-CelG from an uncultured bacterium (accession number AHL27900, 24% identity), cellobiohydrolase CbhA from Clostridium thermocellum (PDB entry 1RQ5, 21% identity), and endoglucanase AaCel9A from Alicyclobacillus acidocaldarius (accession number ACV59481, 19% identity). Sequence alignment of the C terminus-free MiXen with LC-CelG, CbhA, and AaCel9A revealed three conserved active-site residues (Asp192, Asp195, and Glu568) in the catalytic domain of MiXen (Fig. 1a). Secondary structural analysis predicted that the catalytic domain of MiXen belongs to a typical (α/α)₆-barrel fold, which contains 12 long α-helices forming the central (α/α)₆-barrel and 6 antiparallel strands forming two β-sheets (Fig. 1a). Taken together, these primary and secondary structural features suggest that MiXen belongs to the GH9 family.

FIG 1.

Sequence and structure properties of MiXen. (a) Alignment of amino acid sequences of MiXen, LC-CelG, AaCel9A, and CbhA. The accession numbers were ALX66163 for MiXen, AHL27900 for LC-CelG, ACV59481 for AaCel9A, and 1RQ5 for CbhA. The amino acid sequence of MiXen without a putative signal peptide and a C-terminal domain (residues 34 to 580) and the corresponding regions of the other three proteins are shown. The amino acid residues, which are conserved in all four proteins, are denoted with white letters and highlighted in black. The amino acid residues, which are conserved in two or three different proteins, are highlighted in gray. The ranges of the secondary structures of MiXen (βa–βg strands for the Ig-like domain, β1–β6 strands and α1–α13 helices for the catalytic domain) are shown above its sequence based on the predicted crystal structure of MiXen. The conserved residues that form the catalytic site (Asp192, Asp195, and Glu568) and substrate binding site (Tyr199, His511, and Arg513) are indicated by asterisks and plus signs, respectively, above the MiXen sequence. (b) Dimensional cartoon structure of the C-terminus-free MiXen was predicted using the I-TASSER protein structure homology-modeling on-line server (http://zhanglab.ccmb.med.umich.edu/I-TASSER/). The following colors were used to indicate distinct domains in MiXen: cellulase N-terminal Ig-like domain, violet; glycoside hydrolase family 9, green. (c) A structural overlay of the modeled MiXen with LC-CelG (PDB entry 3X17) was made to allow visual comparison. MiXen, green; LC-CelG, blue.

Homology modeling was also performed to further predict the typical GH9 structure of MiXen. The top three identified structural analogs of MiXen, LC-CelG (PDB entry 3X17), CbhA (PDB entry 1RQ5), and AaCel9A (PDB entry 3GZK), were used to construct models of MiXen without the C-terminal domain. Structures 3X17, 1RQ5, and 3GZK have a bond length root mean square deviation of around 0.77 Å, 2.79 Å, and 1.89 Å for the models, respectively, which would reflect the similarity of the protein cores. Figure 1b depicts a hypothetical structure of MiXen without the C-terminal domain. Overall, 68% of the input sequence was modeled at >65% coverage (structural analog, PDB entry 3X17), mainly covering a globular catalytic domain (54.2% of amino acid residues in α-helices) and a cellulase N-terminal Ig-like domain (69.1% amino acid residues in β-strands and coils). The structural superposition of the modeled MiXen with 3X17 revealed that the steric configuration of the MiXen catalytic domain is nearly identical to that of 3X17 with a typical GH9 structure (Fig. 1c). However, the low sequence identity of the MiXen catalytic domain to 3X17, 1RQ5, and 3GZK suggested that MiXen represents a novel branch of the GH9 family.

Recombinant expression and characterization of MiXen.

The gene encoding MiXen free of the N-terminal signal peptide was amplified from the genomic DNA of Microbacterium sp. XT11. The DNA products were gel purified and cloned into the expression vector. To investigate the function of the C-terminal domain of MiXen, two recombinants, MiXen-CD (MiXen mutant without the C-terminal domain) and MiXen-CT (the C-terminal domain of MiXen), were constructed, after which the proteins were heterologously expressed in E. coli BL21(DE3). Following the purification, proteins were examined to confirm the purity and molecular mass by SDS-PAGE. The purified soluble proteins MiXen, MiXen-CD, and MiXen-CT had a purity of ≥90% and migrated as single bands with approximate molecular masses of 100 kDa, 80 kDa, and 25 kDa, respectively (see Fig. S1 in the supplemental material).

To determine the apparent temperature/pH optima and stability of the enzymes, carboxymethyl cellulose (CMC) rather than xanthan was chosen as the model substrate. This was because (i) the conformation of xanthan is susceptible to temperature and pH and (ii) the backbone of xanthan is cellulose-like. MiXen presented its highest activity at 40°C to 45°C and retained more than 70% of its residual activity after incubation for 30 min at temperatures ranging from 20°C to 45°C (Fig. 2a). The optimal pH value for MiXen was 7.5 to 8.0 (Fig. 2b). The pH stability profile showed that MiXen could retain more than 80% of its residual activity after incubation in buffers with pHs ranging from 6 to 9 for 2 h (Fig. 2b). Notably, the maximum activity was obtained when the concentration of the NaH₂PO₄-Na₂HPO₄ buffer was 10 mM. When CMC degradation was conducted in deionized water, dramatically decreasing activity was observed (Fig. 2c). Therefore, the following enzymolysis experiments, including the determination of kinetic parameters, were conducted in 10 mM NaH₂PO₄-Na₂HPO₄ buffer at 40°C and pH 7.5.

FIG 2.

Biochemical characteristics of MiXen. (a) The temperature optimum (●) was measured by incubating 5 g/liter carboxymethyl cellulose (CMC) with 0.15 mg/ml MiXen for 20 min at the indicated temperatures. To determine the thermostability (■), the remaining enzyme activity was measured at 40°C after incubation at the indicated temperatures for 30 min. (b) The pH optimum (●) was determined by measuring the activity at the indicated pH values. For the pH stability (■), enzyme activity was measured at 40°C after incubation at the indicated pH for 2 h. (c) The effects of the NaH₂PO₄-Na₂HPO₄ buffer concentration on enzyme activities.

MiXen could attack xanthan with a highly ordered conformation.

As a key factor that affects the degree of xanthan enzymolysis, the disordered fraction (α) of the purified native xanthan sample was measured by circular dichroism. When xanthan was suspended in 10 mM NaH₂PO₄-Na₂HPO₄ buffer (pH 7.5) at 40°C, the disordered fraction was 0, which reflected that xanthan was folded in a highly ordered conformation under the enzyme assay conditions. As shown in Table 1, after 12 h of incubation with MiXen, the disordered fraction of xanthan digests increased to 0.425 ± 0.018. A smaller disordered fraction of 0.193 ± 0.023 was observed with another monomodular xanthanase, MiGH from Microbacterium sp. XT11 (19), indicating that MiXen degraded more highly ordered xanthan than MiGH. Moreover, the xanthan substrate maintained a highly ordered conformation after incubation with the commercial cellulases or xanthan lyase, which suggested that cellulases and xanthan lyase could not act on the xanthan molecule.

TABLE 1.

Hydrolysis properties of different enzymes on xanthan backbone

Enzyme(s)	Disordered fraction of xanthan digestsa (α)	Sp act (U/g)	Km (g/liter)	kcat (min−1)	kcat/Km (liters/g·min)
MiXen	0.425 ± 0.018	15.70 ± 0.98	0.65 ± 0.02	9.80 ± 0.20	15.08 ± 0.38
MiXen-CD	0.167 ± 0.013	5.12 ± 0.59	1.18 ± 0.02	3.37 ± 0.05	2.86 ± 0.05
MiXen-CT	0	0
MiGH	0.193 ± 0.023	9.20 ± 1.22	0.94 ± 0.01	7.29 ± 0.48	7.75 ± 0.62
Xanthan lyase	0	0
Cellulases	0	0.47 ± 0.25

^aThe disordered fractions of xanthan digests (α) were determined after incubation with enzymes at 40°C and pH 7.5 for 12 h.

The shear-dependent viscosity of xanthan digests was recorded in Fig. 3a. After 12 h of incubation with different enzymes, shear-thinning was observed for all xanthan digests being investigated. The MiXen-digested xanthan showed a low-shear-rate (0.015 s⁻¹) viscosity of about 75.0 Pa·s, which was lower than those observed by MiGH-digested xanthan (83.8 Pa·s), xanthan lyase-digested xanthan (93.6 Pa·s), cellulase-digested xanthan (94.0 Pa·s), and native xanthan (96.8 Pa·s). The decreased viscosity indicated that the backbone of highly ordered xanthan could be effectively cleaved by MiXen, whereas the cellulases and xanthan lyase could not react with the xanthan backbone.

FIG 3.

Hydrolysis performance of different enzymes against highly ordered xanthan. (a) Viscosity of native xanthan as a function of shear rate after being hydrolyzed by different enzymes. Xanthan degradation was conducted at 40°C and pH 7.5 for 12 h. The final concentration of the enzyme was 1.0 mg/ml. Xanthan incubated with bovine serum albumin was used as a negative control. All samples were prepared in duplicate, and the averages are presented. (b) Specific activities of enzymes toward different substrates. The final concentration of MiXen, MiXen-CD, MiXen-CT, MiGH, or xanthan lyase was 0.1 mg/ml. The concentration of mixed cellulases was 0.1 mg/ml for xanthan substrates and 0.01 mg/ml for CMC. Xanthan and CMC (5 g/liter) were incubated with enzymes at 40°C and pH 7.5 for 20 min, respectively. The data represent the averages ± standard deviations from at least three independent samples. *, P < 0.05. (c) GPC elution patterns of xanthan after incubation with different enzymes at 40°C for 12 h. The final concentration of the enzyme was 1.0 mg/ml.

The specific activities of MiXen and other selected enzymes against highly ordered xanthan were then investigated. As shown in Fig. 3b and Table 1, cellulases exhibited the highest activity toward CMC, while there was no significant difference between the specific activity of MiXen and MiGH toward CMC. However, when xanthan was used as the substrate, the MiXen activity was remarkably higher than the MiGH activity, while no activity was detected when cellulases or xanthan lyase was added. Moreover, the catalytic efficiency (k_cat/K_m) of MiXen was 2-fold higher than that of MiGH, and the K_m value of MiXen was significantly lower than that of MiGH, suggesting a higher substrate specificity of MiXen on xanthan with a highly ordered conformation.

The molecular mass distributions of xanthan digests produced by excessive amounts of enzymes were also detected by gel permeation chromatography (GPC) (Fig. 3c). After 12 h of incubation with MiXen, the xanthan enzymolysis products were divided into two fractions, nondegraded high-molecular-mass (HMW) xanthan (retention time [R_t] =18 to 21 min) and some intermediate degradation products (R_t = 21 to 24 min). Compared with MiXen, MiGH degraded HMW xanthan into intermediate degradation products with a wider range of molecular masses. Consistent with the observations in the viscosity and enzyme activity assays, cellulases failed to cleave the backbone of highly ordered xanthan.

MiXen exhibits endotype xanthanase activity.

MiXen resembles (24% sequence identity) the catalytic part of a GH9 from uncultured bacterium (PDB entry 3X17), which has been shown to be an endotype glucanase that exerts significant activities against CMC but no activity toward p-nitrophenyl cellobioside (18). Gel permeation chromatography analysis also provided information about the xanthan-degrading pattern of MiXen with time (Fig. 4). As the incubation time increased, more HMW xanthan was converted into intermediate degradation products. After 6 h of incubation with 1 mg/ml MiXen, no more degradation of the high-molecular-mass xanthan was observed, and no completely degraded low-molecular-mass (LMW) xanthan digests (R_t = 28 to 33 min) were detected. Moreover, further increases in enzyme concentration did not lead to the production of additional xanthan digests (data not shown). The GPC result suggests that MiXen could not cleave at the termini of the xanthan backbone. To further assess the randomness of cleavage of the xanthan backbone by MiXen, the relationship between the production of reducing groups and increase in fluidity was monitored. As shown in Fig. 5, there was a proportionate and steep increase in fluidity with increases in reducing sugar, which demonstrated that MiXen could randomly cleave glucosidic bonds within the xanthan backbone to produce intermediate xanthan digests, suggesting an endotype hydrolytic attack of MiXen.

FIG 4.

GPC elution patterns of xanthan after incubation with 1.0 mg/ml MiXen for different times.

FIG 5.

Relationship between fluidity and production of reducing sugar groups during hydrolysis of xanthan by MiXen.

The C-terminal domain helps MiXen bind to highly ordered xanthan.

Compared with the wild-type protein, the truncated protein free of CBM (MiXen-CD) showed dramatically decreased xanthan-degrading ability. The higher low-shear-rate viscosity (89.8 Pa·s) and smaller disordered fraction (0.167 ± 0.013) of MiXen-CD-digested xanthan suggested a weakened interaction with xanthan (Fig. 3a and Table 1). The specific activity of MiXen-CD on xanthan did not differ significantly from that on CMC and was nearly 2-fold lower than the MiXen activity on xanthan (Fig. 3b). As shown in Fig. 3c, a similar elution pattern was observed when xanthan was digested by truncated MiXen. Compared to MiXen, MiXen-CD showed slightly weaker hydrolysis performance via degrading less xanthan into intermediate degradation products. Michaelis-Menten constants were also determined. As shown in Table 1, the catalytic efficiency (k_cat/K_m) of MiXen was 5-fold higher than that of MiXen-CD. In addition, the K_m value of the truncated protein was significantly higher than that of MiXen, representing a decreased affinity for xanthan. Taken together, these data indicate that the C-terminal domain plays a key role in xanthan hydrolysis by enhancing the substrate affinity of the enzymes.

The low-shear-rate viscosity and disordered fraction of MiXen-CT-digested xanthan were 94.5 Pa·s and 0, respectively (Fig. 3a and Table 1). No activity was observed when xanthan was incubated with MiXen-CT. All of the results indicate that the C-terminal domain of MiXen could not cleave the backbone of xanthan. The surface plasmon resonance (SPR) was then designed to measure the interaction between MiXen-CT and the xanthan substrate. As shown in Fig. 6, the C-terminal domain of MiXen exhibited a strong affinity toward xanthan with a low dissociation constant (K_D) value at 0.651 μg/ml, indicating that the C-terminal domain assists MiXen in hydrolyzing highly ordered xanthan via adsorption on the substrate.

FIG 6.

SPR analysis of the interaction of MiXen-CT and native xanthan. Different traces correspond to increasing xanthan concentrations. The dissociation constant (K_D) values were calculated with a one-site steady-state binding model.

Modification of xanthan side chain affects the hydrolysis activity of MiXen by changing the xanthan secondary structure.

To determine the structural transition of xanthan upon changes in the xanthan side chain, xanthan samples with different side chain modifications, including pyruvate-free xanthan (PFX), acetyl-free xanthan (AFX), acetyl- and pyruvate-free xanthan (APFX), and terminal mannosyl residue-free xanthan (TMFX), were prepared. Tables 2 and 3 provide an overview of chemical compositions of different xanthan samples and the corresponding fraction of disordered conformation (α). Except for TMFX, the disordered fraction of which was detected as 0.100 ± 0.015, all other xanthan samples showed no disordered fraction under enzymolysis conditions. After 12 h of incubation with MiXen, the disordered fraction of all xanthan digests increased. The TMFX digests had the largest disordered fraction, 0.537 ± 0.021, probably because of the partially disordered conformation of the initial substrate. The NX digests also gained a relatively larger disordered fraction (0.425 ± 0.018) than the AFX, PFX, and APFX digests. These results suggest that, after being incubated with MiXen, more disordered molecules were produced from NX and TMFX than from AFX, PFX, and APFX.

TABLE 2.

Summarized properties of xanthan with different modifications

Xanthan samplea	Content (wt/wt, %)		Disordered fractionb (α)
	Acetyl	Pyruvate
NX	13.2	9.6	0
AFX	0	8.2	0
PFX	10.2	5.0	0
APFX	0	5.7	0
TMFX	11.8	6.4	0.100 ± 0.015

^aXanthan samples include native xanthan (NX), pyruvate-free xanthan (PFX), acetyl-free xanthan (AFX), acetyl- and pyruvate-free xanthan (APFX), and terminal mannosyl residue-free xanthan (TMFX).

^bThe fractions of disordered conformation (α) were determined at 40°C and pH 7.5.

TABLE 3.

Summarized characteristics of MiXen against different xanthan substrates

Characteristic	Value for xanthan substrateb
	NX	AFX	PFX	APFX	TMFX
Disordered fraction of xanthan digestsa (α)	0.425 ± 0.018	0.183 ± 0.025	0.242 ± 0.013	0.250 ± 0.016	0.537 ± 0.021
Sp act (U/g)	15.7 ± 0.98	13.0 ± 0.54	11.2 ± 0.35	10.2 ± 0.27	19.4 ± 1.05
Km (g/liter)	0.65 ± 0.02	1.05 ± 0.03	0.85 ± 0.06	1.00 ± 0.10	0.52 ± 0.01
kcat (min−1)	9.80 ± 0.20	9.59 ± 0.67	8.56 ± 0.14	6.24 ± 0.17	12.87 ± 0.21
kcat/Km (liters/g·min)	15.08 ± 0.38	9.13 ± 0.78	10.07 ± 0.20	6.24 ± 0.21	24.75 ± 0.49

^aThe disordered fractions of xanthan digests (α) were determined after incubation with MiXen at 40°C and pH 7.5 for 12 h.

^bXanthan substrates include native xanthan (NX), pyruvate-free xanthan (PFX), acetyl-free xanthan (AFX), acetyl- and pyruvate-free xanthan (APFX), and terminal mannosyl residue-free xanthan (TMFX).

Because the effects of acetyl and/or pyruvate elimination on the xanthan structure could not be significantly observed through order-disorder transitions, the variations in secondary structure among different modified xanthans were investigated directly using atomic force microscopy (AFM). As shown in Fig. 7a, purified native xanthan (NX) exposed visible branches, giving it a tree-like structure, and the mean height of the NX sample was estimated to be 0.66 ± 0.05 nm (Fig. 7f), suggesting a double-stranded secondary structure. In contrast to those of NX, strands of PFX, AFX, and APFX became more agglomerated and complex and exhibited many sharp kinks (Fig. 7b to d). Higher strand heights of the three samples were also observed (PFX, 0.84 ± 0.07 nm; AFX, 0.83 ± 0.09 nm; APFX, 0.85 ± 0.10 nm) (Fig. 7f), which was probably due to the high abundance of locally folded polymers. Taken together, these changes would make xanthan strands more enzyme resistant. Unlike the other samples, terminal mannosyl residue-free xanthan (TMFX) exposed more random-coil structures, with the lowest mean height being 0.48 ± 0.07 nm (Fig. 7e and f), suggesting more single-stranded polymers, which is consistent with a previous report (20). The mean height of xanthan samples after incubation with MiXen was then calculated. As shown in Fig. 7f, compared with that of the undigested samples, the mean height of all degraded xanthan samples decreased, indicating that some locally folded polymers or other complex secondary structures have been transformed into disordered structures. Compared with those of digests produced from AFX, PFX, and APFX, TMFX and NX digests showed relatively lower mean heights of 0.36 ± 0.08 nm and 0.48 ± 0.08 nm, respectively, suggesting more disordered molecules formed. These results were in accordance with those observed in the circular dichroism assay.

FIG 7.

AFM analysis of different xanthan samples. All xanthan samples were suspended in 10 mM NaH₂PO₄-Na₂HPO₄ buffer (pH 7.5) at a final concentration of 1 g/liter. (a to e) AFM topography images of native xanthan (NX), pyruvate-free xanthan (PFX), acetyl-free xanthan (AFX), acetyl- and pyruvate-free xanthan (APFX), and terminal mannosyl residue free xanthan samples (TMFX), respectively. Figure contrast was automatically adjusted with XEP software (versions 1.7.70.3; Park Systems, South Korea) to enable better visualization. (f) Boxplots of the measured heights of the different xanthan samples before (blank) and after (gray) incubation with 1.0 mg/ml MiXen at 40°C for 12 h. The sample sizes were 15 (NX), 15 (PFX), 20 (AFX), 18 (APFX), and 17 (TMFX).

The specific activities and Michaelis-Menten constants were measured after the enzyme MiXen was incubated with different xanthan modifications. Table 3 summarizes the hydrolysis parameters of MiXen. The activities of MiXen on AFX, PFX, and APFX greatly decreased compared to that on the NX sample, and the simultaneous absence of acetyl and pyruvate groups caused the maximum decline of enzyme activity. The maximum MiXen activity was obtained in the TMFX-hydrolyzing system. The Michaelis-Menten constants of MiXen were also determined for different xanthan samples. The K_m values of MiXen on AFX (K_m = 1.05 g/liter), PFX (K_m = 0.85 g/liter), and APFX (K_m = 1.00 g/liter) were significantly higher than that of MiXen on NX (K_m = 0.65 g/liter), indicating decreased substrate affinities. However, a slightly lower K_m value was calculated for TMFX (K_m = 0.52 g/liter). The catalytic efficiencies (k_cat/K_m) of MiXen on AFX, PFX, and APFX were lower than that of MiXen on NX, and the highest k_cat/K_m value was calculated when TMFX was used as the substrate. The enzymatic parameters of MiXen toward different xanthan samples were consistent with the xanthan structural properties observed using circular dichroism and AFM.

DISCUSSION

Similar to cellulose, the full degradation of which depends on synergistic interactions of glycoside hydrolases (21), exhaustive degradation of the cellulose-like backbone of xanthan should be the result of the collective activity of glycoside hydrolases. To date, many studies have investigated xanthan degradation using mixed cellulase systems prepared from cellulolytic microorganisms (8, 10, 11). However, few studies have been conducted to elucidate the hydrolysis properties of individual enzymes on xanthan. A xanthanase in Microbacterium sp. strain XT11, MiXen, is presumed to be a key xanthan-degrading enzyme because (i) the gene encoding MiXen is located in a xanthan-degrading gene cluster in strain XT11 (17) and (ii) sequence and structural features suggest that MiXen belongs to the glycoside hydrolase family 9 (Fig. 1), most members of which have been characterized as cellulases. Interestingly, because of its low sequence and structural homologies to previously identified enzymes (<25%), MiXen may represent a novel branch of the GH9 family. Additionally, the secondary structural prediction for MiXen revealed a multidomain structure, with a function-unknown C-terminal module that might confer some unique hydrolysis properties to MiXen.

Circular dichroism revealed that, in a solution with high ionic strength (Na⁺ concentration, >10 mM) at 40°C, native xanthan exists in a highly ordered conformation (α = 0) (Table 2), which was previously confirmed to be completely resistant to the commercial cellulases from SEAB (France) (11). In the present study, MiXen showed higher specific activity and affinity toward highly ordered xanthan than other selected enzymes and exhibited efficient hydrolysis performance by greatly increasing the disordered fraction and decreasing the viscosity of xanthan (Fig. 3 and Table 1), suggesting that such highly ordered xanthan exhibited less resistance to MiXen. In addition to having cellulase activity, many GH9 cellulases have been reported to display side activities on related polysaccharides, such as glycans (22), xylans (23), or xyloglucans (24), but they most prefer soluble (CMC and cellodextrins) or insoluble (Avicel) cellulose (25). Unlike other known GH9 cellulases, the specific activity of MiXen on xanthan was much higher than that on CMC (Fig. 3), whereas the K_m value against xanthan was significantly lower than that against CMC, indicating that xanthan is the favorite substrate for MiXen. A considerable portion of the high-molecular-mass xanthan remained after 36 h of incubation with an excessive amount of MiXen, which was probably because of the multiple stranded networks of xanthan helices (26). Nevertheless, MiXen should still be considered a good cleaver compared to the other enzymes selected in this study and previously reported cellulase mixtures that could not attack highly ordered xanthan (α = 0.02), even after 48 h of incubation (8).

Based on the investigations described above, we investigated what makes MiXen active specifically on highly ordered xanthan and how it is distinct from the other GH9 cellulases. Earlier studies of modular cellulases focused on bacterial enzymes possessing cellulose-binding CBM, such as family 17, family 28, and family 2a (27–29). Moreover, previous biochemical studies showed that CBM increases the amount of enzyme bound on cellulose, indicating that the main role of CBM is to increase the affinity to cellulose (30, 31). Similar to a large proportion of known GH9 enzymes, MiXen possesses an ancillary C-terminal module in addition to the catalytic domain and an Ig-like domain, which was confirmed to be a carbohydrate-binding module by SPR analysis (Fig. 6). However, it should be noted that both sequence and structural similarities of CBM in MiXen were lower (<10%) than those of other known cellulase CBMs, indicating the C-terminal domain of MiXen is a representative of a novel branch of CBMs. Consistent with previous demonstrations that removal of CBM significantly reduced the activity of cellulases toward insoluble substrates (32, 33), the CBM domain-truncated MiXen-CD lost its substrate specificity for xanthan while also exhibiting dramatically decreased degradation activity, substrate affinity, and catalytic efficiency (Fig. 3 and Table 1). In addition, the CBM domain showed no activity against xanthan, indicating that CBM is simply a helper for substrate recognition. Taken together, these results confirmed the hypothesis that the CBM domain assists MiXen in hydrolyzing highly ordered xanthan by recognizing and binding to the substrate.

It is known that the final degree of enzymolysis could be influenced by the secondary xanthan structure (8), and that changes in xanthan structure depend on different factors, including temperature, pH, ionic strength, and modifications of the xanthan side chain (20, 34, 35). In this study, xanthan degradation was conducted under set temperature, pH, and ionic strength; thus, the effect of xanthan side chain modification on MiXen activity via changes in xanthan structure was further investigated. The removal of pyruvate and/or acetyl groups from xanthan, along with the increasing electrostatic interactions and/or rigidity, led to higher strand heights and the formation of high entanglement points and locally folded polymers (Fig. 7), which is in agreement with the results of previous studies (20, 26, 36). These changes indicated that substituent-free xanthan samples were in a highly ordered state and formed barriers to enzymolysis, explaining the dramatically decreased hydrolysis capacity of MiXen on AFX, PFX, and APFX observed in this study. Notably, compared to native xanthan, which has side chains that revolve around the main-chain skeleton by reverse winding (37), TMFX had a slightly larger disordered fraction (α = 0.1) and more flexible structure, which provided more accessibility to enzymatic degradation. Thus, MiXen showed the highest hydrolysis ability against TMFX.

Conclusions.

MiXen is a xanthan hydrolase obtained from a xanthan-degrading bacterium, Microbacterium sp. XT11, and probably belongs to a novel branch of the GH9 family. This enzyme can efficiently and randomly cleave glucosidic bonds within the backbone of highly ordered xanthan, and the novel C-terminal carbohydrate-binding module of MiXen plays a key role in xanthan recognition and binding. Moreover, this study demonstrated that the modification of xanthan side chains influences the hydrolysis activity of MiXen by changing xanthan secondary structures.

MATERIALS AND METHODS

Strains and culture conditions.

Xanthan-degrading strain Microbacterium sp. XT11 (deposited at the China Center for Type Culture Collection [CCTCC AB2016011]) was isolated by the laboratory of X. Li (16) and used in this study for the cloning of endoxanthanase-encoding genes. Escherichia coli DH5α was used in all cloning experiments. Escherichia coli BL21(DE3) was capable of the expression of endoxanthanase.

Microbacterium sp. XT11 was cultivated at 30°C in xanthan medium (3 g/liter xanthan, 0.5 g/liter glucose, and 3 g/liter yeast extract dissolved in mineral salt solution, pH 7.0). The mineral salt solution contained (0.05 g/liter K₂HPO₄, 0.8 g/liter NaCl, 0.025 g/liter MgSO₄·7H₂O, and 0.70 g/liter KNO₃). E. coli strains were grown at 37°C in Luria-Bertani (LB) medium (10 g/liter tryptone, 5.0 g/liter yeast extract, and 10 g/liter sodium chloride, pH 7.0) supplemented with 100 mg/liter ampicillin or kanamycin if necessary.

Sequence analysis and modeling of enzymes.

Protein annotation was conducted using the Carbohydrate Active Enzyme ANnotation web server (dbCAN; http://cys.bios.niu.edu/dbCAN2/blast.php) (38). Functional domains of the protein were predicted using the InterProScan web server (http://www.ebi.ac.uk/interpro/search/sequence-search) (39). To obtain reference data for MiXen, the online automated protein structure homology-modeling server I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER/) was employed to predict the protein structure (40). The amino acid sequence of MiXen (GenBank accession number ALX66163.1) without the signal peptide (amino acids 1 to 33) was used for modeling.

Cloning, expression, and purification of endoxanthanase constructs.

The coding gene of MiXen without the signal peptide was amplified by PCR using the genomic DNA template extracted from Microbacterium sp. XT11. The forward primer was 5′-CCCGAATTCGCCACCATCGACAAGGTCACGG-3′ (EcoRI restriction site underlined), and the reverse primer was 5′CGCAAGCTTTCAGCCCACAGCGACT-3′ (HindIII restriction site underlined). To construct the C-terminal domain of MiXen and the mutant missing the C terminus, plasmid pET-MiXen was used as the PCR template. The DNA fragment encoding the N terminus and the catalytic domain was deleted using the forward primer 5′-CCCGAATTCGCGTCGGGCAGGTGACCAAC-3′, and the resulting mutant was named MiXen-CT. The DNA fragment encoding the C-terminal domain was deleted using the reverse primer 5′-CGCAAGCTTGGCATGGTTGTTGTACCAATTCG-3′, and the resulting mutant was named MiXen-CD. To amplify the gene encoding another previously reported monomodular xanthanase from Microbacterium sp. XT11 (19), the amino acid sequence was first determined by liquid chromatography-tandem mass spectrometry (LC-MS/MS), which revealed that this xanthanase is identical to a superfamily glycoside hydrolase, MiGH (GenBank accession number WP_067195711), in the genome annotation database of strain XT11 (see Fig. S2 in the supplemental material). The primers used to amplify the MiGH coding genes were 5′-CCCGAATTCATGCGCCCCACCATCG-3′ (EcoRI restriction site underlined) and 5′-CGCAAGCTTTCAGACGACGGTGTTCTG-3′ (HindIII restriction site underlined).

All of the PCR products were digested with EcoRI and HindIII and then ligated into the EcoRI- and HindIII-cut pET28a vector. A 6× His tag sequence was added to the 5′ terminus of DNA products for downstream protein purification. The generated recombinant plasmids pET-MiXen, pET-MiXen-CD, pET-MiXen-CT, and pET-MiGH were transformed into electrocompetent E. coli BL21(DE3) by electroporation at 2.5 kV, 25 μF, and 200 Ω in a 0.2-cm-gap electroporation cuvette using a Bio-Rad Gene Pulser. The recombinants were selected on LB plates containing 100 mg/liter kanamycin.

For the expression of MiXen, MiXen-CD, MiXen-CT, and MiGH, E. coli transformants were grown at 37°C in LB medium supplemented with 30 mg/liter kanamycin until the optical density at 600 nm (OD₆₀₀) reached 0.5. For induction, isopropyl-β-d-thiogalactopyranoside (IPTG) was added at a final concentration of 0.5 mM, after which the cells were further cultured at 16°C for 12 h. Cells next were harvested by centrifugation, and the cell pellet was suspended and sonicated 20 times on ice for 10 s with 30-s intervals between cycles using an ultrasonic processor (Branson Digital Sonifier; Branson, Inc., USA). Following removal of cell debris by centrifugation, the supernatant was ready for subsequent purification. His-tagged proteins were purified using a nickel-nitrilotriacetic acid (Ni-NTA) purification system as previously described (41), and the N-terminal His tag was then cleaved using thrombin and removed using Ni-NTA Sepharose. The proteins were subsequently fractionated on a Superdex 75 gel filtration column (GE Healthcare, Uppsala, Sweden) to remove residual thrombin. The purity of the fractions was assessed by SDS-PAGE. The purified protein concentration was measured using a Bio-Rad protein assay kit.

Enzyme assay.

The specific activity of the enzyme was determined by incubating 5 g/liter xanthan or carboxymethyl cellulose (CMC) with appropriately diluted enzymes in 10 mM NaH₂PO₄-Na₂HPO₄ buffer (pH 7.5). The final concentration of MiXen, MiXen-CD, MiXen-CT, MiGH, or xanthan lyase (42) used in the assay was 0.1 mg/ml. The concentrations of mixed cellulases (Sigma) from Aspergillus niger were 0.1 mg/ml for xanthan substrates and 0.01 mg/ml for CMC. The reaction was conducted at 40°C for 20 min and then terminated by boiling the mixture for 5 min. Under these conditions, the formation of reducing-end sugars was proportional to the reaction time, and less than 5% of the substrate was consumed. The liberation of reducing-end sugars was assayed by the 2,2’-bicinchoninate method (43). One unit of enzyme activity was defined as the amount of enzyme that produced 1 μmol reducing-end sugar per min. All assays were conducted in triplicate, and the mean values were presented.

To assess the optimum temperature, enzyme activity was measured at temperatures ranging from 20°C to 70°C according to the standard procedure described above. The thermostability of the enzyme was determined by assaying the residual enzyme activity after the enzyme was incubated at various temperatures (20°C to 60°C) for 30 min. The optimal pH for activity was assayed by measuring the enzyme activity at 40°C in different buffers, ranging from pH 2 to 10. To measure the pH stability, activity assays were performed using enzyme samples that were preincubated at 4°C for 2 h in different buffers, ranging from pH 2 to 10. Buffers were prepared at a 10 mM concentration at 0.5 pH intervals, including glycine-HCl buffer (pH 2.0 to 4.0), acetic acid-sodium acetate buffer (pH 4.0 to 6.0), NaH₂PO₄-Na₂HPO₄ buffer (pH 6.0 to 8.0), and glycine-NaOH buffer (pH 8.0 to 10.0). The effect of buffer concentration on enzyme activity was examined by determining activity in different concentrations of NaH₂PO₄-Na₂HPO₄ buffer. All reactions to optimize activity were conducted using a final enzyme concentration of 0.15 mg/ml.

Under the optional conditions, kinetic parameters (K_m and V_max) were determined by incubating the enzyme with different concentrations of xanthan solution (0.1, 0.25, 0.5, 0.75, 1.0, 1.5, 2.0, 2.5, 4.0, and 5.0 g/liter). According to the time course for the formation of reducing-end sugars, the initial velocity was calculated as the slope of a tangent at the origin. K_m and V_max were calculated using the Michaelis-Menten equation.

Hydrolysis of xanthan substrates (5 g/liter) was accomplished in 10 mM NaH₂PO₄-Na₂HPO₄ buffer containing an excess of enzyme at 1 mg/ml at 40°C for the specified times.

All assays were conducted in triplicate, and the mean values are presented.

Surface plasmon resonance.

SPR measurement was performed on a BIAcore T200 instrument with a sensor chip with dextran matrix that was precoated with streptavidin (GE Healthcare). The chip was activated with 420-s injections of 0.05 M N-hydroxysuccinimide and 0.2 M ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride. MiXen-CT diluted with running buffer was injected and captured onto the channel surface to achieve an immobilization level of about 2,000 RU (response units). For the binding assay, the reaction temperature was controlled at 25 ± 0.01°C and the flow rate was set to 30 μl/min. The response obtained from the detection channel (xanthan) was normalized by subtracting the signal simultaneously acquired from the control channel, which could eliminate nonspecific binding and buffer-induced bulk refractive index changes. The running buffer was NaH₂PO₄-Na₂HPO₄ (10 mM, pH 7.5). Different concentrations (1.25, 2.5, 5, 10, and 20 nM) of xanthan were prepared, and their binding with MiXen-CT was measured by SPR assay. Briefly, the sample solutions were injected onto the chip surface for 120 s. Following each binding reaction, a further dissociation time of 600 to 3,000 s was applied, and as the regeneration buffer, 5 M NaCl solution was injected for 60 s to allow the signal to return to baseline.

Preparation and chemical analysis of xanthan samples.

For xanthan purification, 1 g/liter food-grade xanthan solution in distilled water was precipitated by adding an equal volume of ice-cold absolute ethanol in the presence of NaCl (100 g/liter). After stirring for 2 h at 4°C, the precipitate was collected by filtration and washed with 95% ethanol (44). The purified native xanthan was lyophilized for the following test.

The chemical modification of xanthan was prepared as previously described. Briefly, 5 g/liter purified xanthan solution in 5 mM trifluoroacetic acid was heated at 100°C for 1.5 h to remove the pyruvic acetyl groups, which yielded PFX (45). AFX was prepared by saponification treatment of NX with 1 M NaOH at 60°C for 1 h to remove the acetyl groups (15). APFX was produced by 5 mM trifluoroacetic acid treatment followed by saponification. To obtain TMFX samples, 10 g/liter purified xanthan solution was incubated with 1 mg/ml xanthan lyase, prepared by our laboratory (42), at 40°C for 12 h. All modified xanthan polymers were dialyzed against demineralized water for 24 h and then lyophilized.

The acetyl content of xanthan was measured according to the method described by Hestrin et al. (46) using acetylcholine as a reference. To determine the pyruvate content of xanthan, 0.01 g xanthan was hydrolyzed in 5 ml HCl (1 M) at 100°C for 3 h, after which the free pyruvate was determined as described by Sloneker et al. (47) using sodium pyruvate as the reference.

Circular dichroism.

The ellipticities (millidegrees) of 1 g/liter xanthan or xanthan digests were monitored at 219 nm using a Jasco J-1500 spectropolarimeter (Jasco Corp., Tokyo, Japan) with a response time of 1 s, sensitivity of 100 millidegrees, and bandwidth of 2 nm. The transition profiles of NX, AFX, PFX, APFX, and TMFX were determined in 10 mM NaH₂PO₄-Na₂HPO₄ buffer (pH 7.5) at 40°C using a Jasco MCB-100 controller. Quartz cuvettes with an optical path of 1 cm were used. The obtained curves were normalized by the best-fit parameters, and the fractions of disordered conformation (α) were determined by transition profiles using the following equation:

$$\text{α}=1-\frac{{\text{θ}}_t-{\text{θ}}_U}{{\text{θ}}_F-{\text{θ}}_U}$$

where θ_t is ellipticity at a given condition, θ_U is ellipticity of a completely disordered structure, and θ_F is ellipticity of a completely ordered structure (48). The minimum and maximum ellipticities of each type of xanthan were determined as described by Kool et al. (8).

Atomic force microscopy imaging of xanthan samples.

To allow imaging of individual well-separated xanthan molecules, 5-μl aliquots of undigested or digested xanthan samples (10 mg/liter) were pipetted onto freshly cleaved mica sheets and dried under a gentle flow of dry nitrogen gas. Topographical and error signal mode imaging of the samples was then conducted using a commercial AFM (XE-Bio; Park Systems, South Korea) in noncontact mode with NCHR monolithic silicon cantilevers (Park Systems, South Korea). To provide a disturbance-free environment, an e-Stable mini vibration isolation table and an acoustic enclosure (Park Systems, South Korea) were used during the measurements. The scanning and imaging process was conducted using the AFM controller software XEP (versions 1.7.70.3). The image resolution was set to 256 by 256 pixels for overview images (10 by 10 μm²), and imaging was conducted under ambient conditions with a scan rate of 0.5 Hz.

Raw AFM data were processed and analyzed using the XEI software package (version 1.8.0; Park Systems, South Korea), while xanthan structural analysis was conducted using the software package XEP (version 1.8.4; Park Systems, South Korea). Using XEI, the raw data were flattened and error lines were corrected. Furthermore, the height of individual xanthan strands was estimated by XEI. The height of an individual strand was evaluated by averaging the height estimated from at least five cross sections. To estimate the mean height, the overall mean of one sample was calculated.

Gel permeation chromatography analysis.

The molecular masses of xanthan and its digests were determined by gel permeation chromatography (GPC) using an Agilent 1260 infinity system (Agilent Technologies, Santa Clara, CA, USA). A set of three gel columns (Agilent Technologies, Santa Clara, CA, USA) was used in series with PL aquagel-OH 60, PL aquagel-OH MIXED-M, and PL aquagel-OH 30 separation columns (8 μm by 300 mm). The column temperature was set to 40°C, and the samples (20 μl, 1 g/liter) were eluted with 0.1 M NaNO₃ at a flow rate of 0.8 ml/min (8, 49). The eluate was then monitored using refractive index detection. Molecular masses were estimated with the help of pullulan molecular mass standards (Polymer Laboratories, Palo Alto, CA, USA). The relative molecular mass distribution was calculated by dividing the integrated refractive index (RI) peak area of each fraction by the total RI peak area measured from 18 to 33 min.

Viscosity measurements.

The viscosities of xanthan or xanthan digests were measured against a variety of shear rates using a Discovery hybrid rheometer (TA Instruments), with the following settings: geometry, 40-mm rotating drum; temperature, 40°C; stop head gap, 1,000.0 μm; geometry gap, 1,000.0 μm; torque, 5.0 to 1,000.0 μN · m. A total of 1.0 ml of xanthan enzyme products was taken, and the viscosity was measured against the shear rate for 60 s. All samples were prepared in duplicate.

Statistical analysis.

All analyses were performed in triplicate, and all data were expressed as the means ± standard errors of the means. Analyses were conducted using SPSS 11 (SPSS Inc., Chicago, IL), and a P value of <0.05 was considered significant.

Accession number(s).

The amino acid sequences of MiXen and MiGH have been submitted to GenBank under accession numbers ALX66163.1 and WP_067195711, respectively.