Carbohydrate-binding modules (CBMs) are found within multi-modular polysaccharide degrading enzymes [glycoside hydrolases (GHs)]. CBMs play a critical role in the recognition of plant cell-wall polysaccharides and enhance the hydrolase activity of their cognate catalytic domains by increasing enzyme substrate proximity. Mimicking their role in Nature, we, in the present study, propose that CBMs may assist in vitro glycosynthase-catalysed polymerization reactions to produce artificial polysaccharides. Glycosynthases are GHs that have been engineered to catalyse glycoside bond formation for the synthesis of oligosaccharides, glycoconjugates and glycans. The degree of polymerization (DP) of the glycans generated is limited by the solubility of the polymeric product. In the present study, we have targeted the synthesis of artificial 1,3-1,4-β-glucans with a regular sequence using the glycosynthase E134S derived from a Bacillus licheniformis lichenase. We show that the addition of CBM11, which binds mixed-linked β-glucans, either as an isolated protein or fused to the glycosynthase E134S, has an effect on the DP of the polysaccharide products that is dependent on the rate of polymerization. The mechanism by which CBM influences the DP of the synthesized glycans is discussed.

INTRODUCTION

Non-catalytic carbohydrate-binding modules (CBMs) are components of multi-modular polysaccharide degrading enzymes where they play critical roles in the recognition of plant cell-wall polysaccharides such as cellulose, β-glucans or insoluble storage polysaccharides [1,2]. CBMs promote the association of the enzyme with the substrate and this proximity effect enhances the activity of their cognate catalytic module and provides efficient hydrolysis [3]. Removal of CBMs greatly reduces the activity of the truncated enzymes against insoluble substrates [4]. In addition, there are examples where CBMs are components of substrate-binding sites or even the active site, playing a critical role in substrate specificity [57]. To date, CBMs are classified in 71 families according to the CAZY database (Carbohydrate Active Enzymes, www.CAZY.org) [8]. Some CBM families, typically those that recognize crystalline polysaccharides, present invariant ligand specificity, whereas other families cover a broad range of carbohydrate specificities.

Mimicking the role of CBMs in polysaccharide hydrolysis in Nature, we propose that CBMs may assist in vitro glycosynthase-catalysed polymerization reactions. Glycosynthases, mutated glycoside hydrolases (GHs) lacking the catalytic nucleophile, have become powerful biocatalysts for the efficient synthesis of oligosaccharides and glycoconjugates [912]. They are able to catalyse glycosyl transfer from glycosyl fluoride donors, with opposite anomeric configuration to the natural substrate of the parental hydrolase, to an acceptor substrate with high yields of glycoside bond formation since products cannot be hydrolysed. Glycosynthase technology has been extended to polysaccharide synthesis [10,1315]. Self-condensation of a laminaribiosyl donor by the E134A glycosynthase derived from Bacillus licheniformis 1,3-1,4-β-glucanase (a GH from CAZY family GH16) led to insoluble polymers with a regular structure of alternating β-1,3 and β-1,4 linkages [13]. With different glycosyl fluoride donors (Glcβ4)mGlcβ3GlcαF (m=0–2), artificial mixed-linked β-glucans with different ratios and patterns of β-1,3 and β-1,4 linkages were obtained [16]. Polysaccharide morphology depends on the repeating unit. Thus, (4Glcβ3Glcβ)n and (4Glcβ4Glcβ4Glcβ3Glcβ)n generated polysaccharides that were crystalline spherulites, whereas (4Glcβ4Glcβ3Glcβ)n polymers formed amorphous precipitates. The degree of polymerization (DP) with the original E134A glycosynthase was approximately constant and independent of donor and enzyme concentrations leading to polymers of average molecular mass (Mw) of 10–12 kDa. However, a linear dependence of Mw on enzyme concentration was observed with the more active E134S glycosynthase mutant [16]. Polysaccharides with Mw of 30 kDa (DP=188) containing a small fraction of products up to 70 kDa were obtained, presenting a high polydispersity index (PDI). As the glycosynthase reaction takes place in solution, the polymerization can proceed until the products become insoluble and precipitate. The rate of polysaccharide formation and the amount of high-molecular-mass polymers produced led to oversaturated solutions and the resultant precipitation of the product entrapped the lower molecular mass polymers. Thus the insoluble polymer had a bimodal profile. Therefore, by increasing the rate of condensation relative to the rate of precipitation (i.e., more active glycosynthase), polymerization can be extended above the apparent solubility limit as long as the product can remain in solution.

With the goal of extending the DP, we, in the present study, propose that CBMs could enhance the solubility of the new polymers formed during the polymerization reaction by preventing inter-chain glycan–glycan interactions leading to increased turnover of the enzyme and insoluble polymers of higher molecular mass that, probably, have low PDIs. Therefore, CBMs able to bind 1,3-1,4-β-glucans were considered, exemplified by the family 11 CBM (CBM11) from Clostridium thermocellum Lic26A-Cel5E [17]. This multi-modular enzyme contains GH5 and GH26 catalytic domains that display β-1,4 and β-1,3-1,4 mixed-linked endoglucanase activity respectively and a CBM11 domain with a β-sandwich structure with a concave face that forms a substrate-binding cleft that displays affinity for both β-1,4 and β-1,3-1,4 mixed-linked glucans [17].

In the present study, we investigated the effect of CBM11 on the polymerization reaction catalysed by the E134S glycosynthase (Figure 1) to test the hypothesis that these modules can increase the DP of the product by reducing premature product precipitation.

Glycosynthase-catalysed polymerization to produce 1,3-1,4-β-glucans

EXPERIMENTAL

Substrates

Glycosyl fluoride donors Glcβ3GlcαF and Glcβ4Glcβ3GlcαF were prepared as previously reported [18,19] by treatment of peracetylated laminaribiose and 3-O-β-cellobiosyl-α-D-glucose with hydrogen fluoride in pyridine, purification of the peracetylated α-glycosyl fluorides by flash chromatography and de-O-acetylation with sodium methoxide in methanol.

Proteins

E134S mutant 1,3-1,4-β-glucanase from B. licheniformis was encoded by pET16b-E134S that contains an N-terminal His10-tag. It was expressed in Escherichia coli BL21 DE3 star cells and purified by immobilized metal affinity chromatography (IMAC) as reported [16]. E134S mutant in phosphate buffer solution (50 mM phosphate, pH 7.0, 0.1 mM CaCl2) was obtained with a yield of 23 mg/l of culture. CBM11 domain from C. thermocellum Lic26A-Cel5E enzyme was already cloned in pET21a-CBM11 [17] and it was expressed as C-terminal His6-tagged protein. E. coli BL21 DE3 star cells harbouring the plasmid were grown in 500 ml of Luria–Bertani (LB) medium (50 μg/ml ampicillin, 2% glucose) for 12 h at 37°C and 250 rpm until late exponential phase. Then, cells were cultured in LB medium containing 1 mM IPTG and 50 μg/ml ampicillin and induced for 5.5 h at 21°C. The cells were harvested and resuspended in 75 ml of 50 mM Tris/HCl buffer, pH 8.0, 300 mM NaCl, containing 1 mM PMSF. After lysis, CBM11 was purified by IMAC and eluted at 10 mM imidazole. CBM11 was obtained in 22 mg/l of culture yield.

The synthetic gene encoding the fusion protein E134S-CBM11 with E134S glycosynthase at the N-terminus, the (PT)7P linker and CBM11 domain at the C-terminus was synthesized by GenScript and cloned into vector pET16b. The amplified fragment of 1340 bp codes for the N-terminal His10-tagged protein of 396 amino acids (Supplementary Figure S1, Supporting Information). BL21 DE3 star cells were transformed for protein expression and after growth in 500 ml of LB medium (100 μg/ml ampicillin, 2% glucose) for 8 h at 37°C and 250 rpm, protein expression was induced by the addition of 1 mM IPTG. The cells were incubated at 20°C for 6 h, collected and resuspended in 50 mM phosphate, 200 mM NaCl, pH 7.0, 0.1 mM CaCl2. After cell disruption, the recombinant fusion protein was purified by IMAC as reported for E134S [20] and 93 mg/l of culture were obtained. Protein concentrations were determined for E134S mutant by UV spectrophotometry using ε280 of 3.53×105 M−1·cm−1 and for E134S-CBM11 protein fusion and CBM11 domain by Bradford using BSA as standard. Molecular masses of the proteins were assessed by MALDI-TOF MS (Microflex, Bruker). E134S and E134S-CBM11 were lyophilized for storage and re-dissolved prior to use. CBM11 was stored in solution at 6°C.

3D models of the E134S-CBM11 fusion protein

Secondary structures of each domain at the N-terminus or the C-terminus of the fusion protein were predicted using the Secondary Structure Prediction Server (Jpred) [21]. The fusion protein composed of the E134S glycosynthase domain and the CBM11 domain linked by a (PT)7P was modelled to assess that both domains may adopt proper orientations for ligand binding. 3D structure models were generated by means of homology modelling and simulated annealing. The templates were B. licheniformis β-glucanase (PDB code 1GBG; 98.6% sequence identity to E134S domain) and CBM11 of C. thermocellum Cel5E (PDB code 1V0A; 98.2% sequence identity to CBM11 domain). No structural template was used for the (PT)7 linker. A pool of 25 3D structures was generated with MODELLER v. 9.8 using a slow level refinement scheme based on simulated annealing molecular dynamics [22]. Supplementary Figure 2S shows the superimposition of the 25 generated structures showing the flexibility of the linker between both domains. A selected structure in which the substrate-binding clefts of both domains lie closer to each other is shown in Figure 4. Figures were generated using VMD (VMD software) [23].

HPSEC chromatograms of polymers (4Glcβ3Glcβ)n (A) and (4Glcβ4Glcβ3Glcβ)n (B) obtained under different conditions with E134S-CBM11

Figure 2
HPSEC chromatograms of polymers (4Glcβ3Glcβ)n (A) and (4Glcβ4Glcβ3Glcβ)n (B) obtained under different conditions with E134S-CBM11

Labels correspond to entries in Tables 1 and 2. 

Figure 2
HPSEC chromatograms of polymers (4Glcβ3Glcβ)n (A) and (4Glcβ4Glcβ3Glcβ)n (B) obtained under different conditions with E134S-CBM11

Labels correspond to entries in Tables 1 and 2. 

Enzyme kinetics for the E134S mutant and the E134S-CBM11 fusion protein

Trisaccharyl fluoride Glcβ4Glcβ3GlcαF donor (0.05–2.00 mM) and 4-nitrophenyl β-D-glucoside acceptor (20 mM) were dissolved in phosphate buffer (50 mM, pH 7.0) and CaCl2 (0.1 mM). Reactions were run in 96-well plates. After pre-incubation of the plate at 35°C for 5 min, reactions were initiated by addition of the enzyme (1.58 μM) and kept at 35°C. Aliquots were withdrawn at regular time intervals, diluted 1:10 with formic acid 2% (v/v) and analysed by HPLC (Agilent 1200 HPLC, Nova-Pak® C18 column, 4 μm, 3.9×150 mm from Waters, 1 ml/min, 14% MeOH in water, UV detector at 300 nm). Chromatographic peaks were assigned by co-injection with independent standards. Initial rates (vo) were obtained from the linear progress curve of product formation (normalized area compared with time) and expressed as vo/[E] in s−1. Kinetic parameters kcat, KM and kcat/KM were calculated by non-linear regression to the Michaelis–Menten equation.

Enzymatic polymerization reactions

Reaction mixtures (0.4 ml) consisted of donor substrate (290 mM laminaribiosyl fluoride or 150 mM trisaccharidyl fluoride), E134S (20–200 μM) or E134S-CBM11 fusion protein (50–120 μM) and CBM11 (0–1200 μM) in phosphate buffer (100 mM, pH=7.2), CaCl2 (0.1 mM). The reactions were run at 37°C and 250 rpm in an orbital shaker for 1 or 3 days, as indicated. A precipitate was formed during the reaction and the presence of unreacted donor or hydrolysis product was checked by TLC (CH3CN–H2O=8:2). The precipitated product was isolated by centrifugation at 15 000 g for 3 min and it was thoroughly washed with cold water. Finally, the product was freeze-dried to yield water-insoluble polymers as white powders. Supernatants were also lyophilized to recover soluble oligomers.

Polymers analysis by HPSEC, NMR and SEM

High performance size exclusion chromatography (HPSEC) analyses to determine molecular mass profiles were performed on an Agilent 1100 HPLC system equipped with a refractive index detector using a PSS Gram column [8.0×300 mm, 100 Å (1 Å=0.1 nm), 10 μm] and a PSS GRAM pre-column (9.0×50 mm, 100 Å, 10 μm) thermostated at 50°C with DMSO and lithium bromide (5 g/l) as eluent at a flow rate of 0.5 ml/min. The calibration curve was obtained with dextrans as standards (American Polymer Standards Corporation DXT1–DXT55 kDa). Freeze dried polymers and standards were dissolved in DMSO with lithium bromide (5 g/l) and filtered. From the chromatograms, Mp (molecular mass of the peak maximum), Mw, Mn (number average molecular mass), DP and PDI were calculated [24]. 13C-NMR spectra of polymers (4Glcβ3Glcβ)n and (4Glcβ4Glcβ3Glcβ)n were identical with those previously reported in the absence of CBM [16]. For SEM experiments, the dried product was fixed on a graphite tape, coated with gold/palladium by ion-sputtering and observed at an accelerating voltage of 10 kV and working distance of 16 mm using a JEOL JSM-5310 microscope.

RESULTS AND DISCUSSION

Glycosynthase reactions for polysaccharide synthesis were explored with the E134S 1,3-1,4-β-glucanase mutant (E143S) in the presence of CBM11 as a discrete entity and when fused to E143S (CBM11-E143S).

Effect of the CBM11 protein domain on polymerization reactions catalysed by the glycosynthase E134S

The glycosynthase E134S was shown previously to synthesize high molecular-mass polysaccharides reflecting its catalytic efficiency [13,16]. This glycosynthase variant was thus used to explore the effect of CBM11 on polysaccharide synthesis. E134S was expressed as a N-terminus His-tagged protein and CBM11 as a C-terminus His-tagged protein and obtained with the expected molecular masses of 27 and 19.6 kDa respectively.

Polymerization reactions were carried out with α-laminaribiosyl fluoride (Glcβ3GlcαF) and 3-O-β-cellobiosyl-α-D-glucopyranosyl fluoride (Glcβ4Glcβ3GlcαF) donors at different E134S-CBM concentration ratios (Tables 1 and 2). As expected, insoluble polymers were obtained with mixed β-1,3 and β-1,4 linkages corresponding to (4Glcβ3Glcβ)n and (4Glcβ4Glcβ3Glcβ)n when using the donors Glcβ3GlcαF and Glcβ4Glcβ3GlcαF respectively, consistent with the specificity of the enzyme.

Table 1
Polymerization reactions of donor Glcβ3GlcαF by E134S-CBM11

Conditions: [Donor]=290 mM. E134S: 0.21 units (for the reference glycosynthase reaction Glcβ4Glcβ3GlcαF + GlcpNP, [CBM11]=0–360 μM), phosphate buffer, pH 7.0, 0.1 mM CaCl2, 37°C. Products analysed after 1 or 3 days of reaction.

 Molar ratio Glcβ3GlcαF 1 day 
Entry E134S-CBM11 Mw* (kDa) n (DP) Mn (kDa) Yield (%) PDI§ Mp (kDa) 
1:0 16 50 (100) 11 45 1.6 42, 16 
1:1 17 53 (105) 10 48 1.7 43, 15 
1:2 23 70 (140) 18 52 1.4 40, 21 
1:3 20 62 (123) 11 69 1.5 44, 21 
  Glcβ3GlcαF 3 days 
1:0 17 53 (105) 56 2.0 44, 17 
 Molar ratio Glcβ3GlcαF 1 day 
Entry E134S-CBM11 Mw* (kDa) n (DP) Mn (kDa) Yield (%) PDI§ Mp (kDa) 
1:0 16 50 (100) 11 45 1.6 42, 16 
1:1 17 53 (105) 10 48 1.7 43, 15 
1:2 23 70 (140) 18 52 1.4 40, 21 
1:3 20 62 (123) 11 69 1.5 44, 21 
  Glcβ3GlcαF 3 days 
1:0 17 53 (105) 56 2.0 44, 17 

*Mw, weight average molecular mass.

n, number of donor units in the polymer; DP, degree of polymerization expressed as glucosyl units.

Mn, number average molecular mass.

§PDI, polydispersity index.

Mp, molecular mass of the peak maximum.

Table 2
Polymerization reaction of donor Glcβ4Glcβ3GlcαF by E134S-CBM11

Conditions: [Donor]=150 mM; for polymers 6–8, E134S: 2.6 units (for the reference glycosynthase reaction Glcβ4Glcβ3GlcαF + GlcpNP, [CBM11]=0–640 μM), for polymers 9–11, E134S: 0.21 units (for the reference glycosynthase reaction Glcβ4Glcβ3GlcαF + GlcpNP, [CBM11]=0–1200 μM), phosphate buffer, pH 7.0, 0.1 mM CaCl2, 37°C.

  Glcβ4Glcβ3GlcαF 1 day (E=2.6 U) 
  Insoluble fraction Soluble fraction 
Entry Molar ratio E134S-CBM Mw* (kDa) n (DP) Mn (kDa) Yield (%) PDI Mp (kDa) Mw (kDa) n (DP) Mn (kDa) Yield (%) PDI 
1:0 18 37 (111) 11 82 1.9 20, 40 20 41 (123) 3.4 
1:1 18 37 (111) 11 79 1.7 21, 40 13 27 (80) 3.2 
1:4 17 35 (105) 11 51 1.5 18 16 33 (99) 46 1.6 
  Glcβ4Glcβ3GlcαF 3 days (E=0.21 U) 
1:0 19 39 (117) 11 82 1.7 22, 40 – – – – 
10 1:4 22 45 (136) 13 92 1.8 20, 40 – – – – 
11 1:6 18 37 (111) 11 92 1.6 18 – – – – – 
  Glcβ4Glcβ3GlcαF 1 day (E=2.6 U) 
  Insoluble fraction Soluble fraction 
Entry Molar ratio E134S-CBM Mw* (kDa) n (DP) Mn (kDa) Yield (%) PDI Mp (kDa) Mw (kDa) n (DP) Mn (kDa) Yield (%) PDI 
1:0 18 37 (111) 11 82 1.9 20, 40 20 41 (123) 3.4 
1:1 18 37 (111) 11 79 1.7 21, 40 13 27 (80) 3.2 
1:4 17 35 (105) 11 51 1.5 18 16 33 (99) 46 1.6 
  Glcβ4Glcβ3GlcαF 3 days (E=0.21 U) 
1:0 19 39 (117) 11 82 1.7 22, 40 – – – – 
10 1:4 22 45 (136) 13 92 1.8 20, 40 – – – – 
11 1:6 18 37 (111) 11 92 1.6 18 – – – – – 

*Polymer parameters as defined in Table 1.

Using the disaccharide donor and low enzyme activity, insoluble polymer yields did not exceed 70% after 1 and 3 days in all cases (polymers 1–5, Table 1). The soluble fraction contained polymers and laminaribiose from donor hydrolysis. Molecular mass distributions of the insoluble polymers presented bimodal profiles (Figure 2A). Polymer 1 obtained without CBM11 showed two peaks with a molecular mass Mp of 16 kDa (DP =100) and 42 kDa (DP=260) respectively, corresponding to Mw of 16 kDa with a DP of ∼100 glucosyl units. In the presence of CBM11 (polymers 2–4), Mp values were shifted up to 21 and 44 kDa respectively. The resulting average Mw value of polymer 4 was ∼20 kDa, which corresponds to insoluble polymers of DP=123 glucosyl units. Moreover, the presence of CBM11 had an important effect on morphology (Figure 3). Polymer 1 revealed the characteristic spherulitic morphology of polymers synthesized from laminaribiosyl fluoride, with an average diameter of 7–10 μm [13]. The spherulite formation is a spontaneous self-assembling process during the enzymatic polymerization since spherulites are directly observed by SEM after filtration and lyophilization of the precipitate. Polymer 2, produced with an equimolar concentration of E134S and CBM, presented a similar morphology. However, when increasing the ratio of CBM, fewer spherulites were observed in polymer 3 and polymer 4 contained no spherulites but an amorphous morphology. Therefore, the CBM extends the glycosynthase-catalysed polymerization yielding up to 25% larger polymers probably due to disruption of the interaction between nascent polysaccharides resulting in increased solubility of the products. In addition to obtaining larger insoluble polymers, the CBM interferes with polysaccharide assembly to form spherulites. This in vitro effect resembles the disruption of the structure of crystalline polysaccharides by some CBMs. This function was first documented for the N-terminal family 2a CBM of Cel6A from Cellulomonas fimi [25]. The CBM mediated non-catalytic disruption of the crystalline structure of cellulose and enhanced the degradative capacity of the catalytic module. This function has not been described for CBM11 with respect to β-glucanase hydrolytic activity.

SEM micrographs of freeze-dried polymers (4Glcβ3Glcβ)n

Figure 3
SEM micrographs of freeze-dried polymers (4Glcβ3Glcβ)n

(AD) Correspond to polymers 1–4 (Table 1).

Figure 3
SEM micrographs of freeze-dried polymers (4Glcβ3Glcβ)n

(AD) Correspond to polymers 1–4 (Table 1).

Using the trisaccharyl donor, insoluble polymers 6–11 were obtained (Table 2). As expected, due to the higher reactivity of the trisaccharyl donor compared with the disaccharyl donor, yields of insoluble polysaccharides by E134S were higher reaching 82% yield after 1 or 3 days either at high or at low enzyme activity (reactions 6 and 9 respectively). The molecular mass distribution of polymer 6 presented a bimodal profile with Mp values of 20 and 40 kDa (Figure 2B). The Mw was 18 kDa (PDI=1.9), which corresponds to insoluble polymers with a DP of ∼110 glucosyl units. The liquid fraction after the reaction contained a small amount of soluble polymers, with no donor trisaccharide (Glcβ4Glcβ3GlcαF or Glcβ4Glcβ3Glc) evident, indicating that transglycosylation was complete and no hydrolysis occurred. Interestingly, with a 1:4 E134-CBM ratio the percentage of soluble polymers increased to 46% after 24 h (entry 8) and by 3 days, most of the polymers were insoluble (entry 10). With E134S-CBM ratios from 1:0–1:4, the amount (yield) of insoluble polysaccharides decreased with a corresponding increase in the yield of soluble polysaccharides. After longer incubation times (3 days) most of the polysaccharides became insoluble (yields of 90%). The transient solubility of the synthesized polymers, after 1 day, did not affect the molecular mass of the polymers. With excess of CBM (entries 10 and 11), insoluble polymers had Mw values of ∼20 kDa, which is similar to polymer 9 produced in the absence of CBM. Most probably the trisaccharide donor was quickly self-condensed by the enzyme, polymers were elongated very fast and, although they remained soluble (oversaturated) due to the presence of CBM11 for 1 day, the reaction was already complete. The appearance of insoluble polysaccharides after 3 days reflected the gradual displacement of CBMs from the polymers, which was mediated by interactions between the glucan chains.

In conclusion, the addition of CBM11 to the polymerization reaction by E134S has several consequences. Polymerization of the slow-reacting disaccharyl donor gave insoluble polymers in 50%–70% yields with 25% higher DP in the presence of CBM, where spherulite formation was clearly prevented. Polymerization reactions with the fast-reacting trisaccharyl donor gave insoluble polymers in higher yields; the presence of CBM slowed the kinetics of precipitation but had no effect on the molecular mass of the final polymers.

Polymerization reactions catalysed by the E134S-CBM11 fusion protein

Fusion protein design

Since CBMs are often appended to GHs, a fusion protein consisting of a chimeric enzyme on which the glycosynthase was appended to a single CBM11 module was designed. We hypothesize that the CBM in the fusion protein may accommodate the polymer as it is synthesized and achieves longer polysaccharides before their precipitation.

The linker between E134S and CBM11 was the first consideration in the fusion protein design. Based on previous studies of Gustavsson et al. [26] and Kavoosi et al. [27], the (PT)7P linker composed of prolines and threonines was selected according to proteolytic stability in E. coli and spatial positioning of the fused domains. Then, structural predictions of the fusion protein with E134S or CBM at the N-terminus were performed and compared with the crystal structures of the single proteins, 1GBG and 1VOA, PDB accession codes of E134S and CBM11 respectively. Whereas the E134S glycosynthase protein domain retained its secondary structure whether located at the N- or C-terminus in the fusion protein, when located at the N-terminus the conformation of CBM11 was predicted to be significantly disrupted; specifically β-strands 1 and 2 were replaced by an α-helix (from residues 7 to 15) and the length of β-strand 7 was reduced. In both predictions the linker is in an extended conformation. Therefore, in the fusion protein the E134S glycosynthase and CBM11 were located at the N- and C-termini, respectively. A 3D structural model was built to assess that both domains may adopt proper orientations for ligand binding. The different models generated show high flexibility of the linker between both domains confirming the proper design of the linker (Supplementary Information, Supplementary Figures S1 and S2). A selected structure in which the substrate-binding clefts of both domains lie closer to each other is shown in Figure 4. The fusion protein was expressed as an N-terminal His10-tagged protein in good yield with a molecular mass of 47089 Da (confirmed by size exclusion chromatography and MALDI-TOF MS, demonstrating that no inter-domain interaction occurred during protein expression or purification).

3D structural model of the E134S-CBM11 fusion protein
Figure 4
3D structural model of the E134S-CBM11 fusion protein

A representative structure with the shortest distance between E134S and CBM11 substrate pockets.

Figure 4
3D structural model of the E134S-CBM11 fusion protein

A representative structure with the shortest distance between E134S and CBM11 substrate pockets.

Specific glycosynthase activity

The fusion protein E134S-CBM11 was characterized by performing the glycosynthase reaction between the donor Glcβ4Glcβ3GlcαF and the acceptor GlcPNP. Transglycosylation rates at constant acceptor concentration (20 mM) and different donor concentrations (0–2 mM) showed that tetrasaccharide product formation followed saturation kinetics with an apparent kcat value of 16.6 min−1 and KM of 0.25 mM (Table 3). Compared with E134S, the kcat of the fusion protein was 2-fold lower and the KM value was ∼3-fold higher, resulting in a 7-fold decrease in kcat/KM.

Table 3
Kinetic parameters for the glycosynthase reactions Glcβ4Glcβ3GlcαF + GlcpNP catalysed by E134S and E1434S-CBM11 fusion protein

Conditions: 0.1–2.0 mM Glcβ4Glcβ3GlcαF donor, 50–100 mM phosphate buffer, pH 7.2, 0.1 mM CaCl2 and 35°C. For E134S, 7.8 mM GlcPNP acceptor and 0.1 μM enzymes were used. For E134S-CBM11, 20 mM GlcpNP acceptor and 1.58 μM enzymes were used.

Enzyme kcat (min−1KM (mM) kcat/KM(M−1·s−1
E13440.8±0.6 0.09±0.01 7.55×103 
E134S-CBM11 16.6±0.98 0.25±0.06 1.10×103 
Enzyme kcat (min−1KM (mM) kcat/KM(M−1·s−1
E13440.8±0.6 0.09±0.01 7.55×103 
E134S-CBM11 16.6±0.98 0.25±0.06 1.10×103 

Polymerization reactions by E134S-CBM11

Polymerization reactions of the trisaccharyl donor with E134S-CBM11 and E134S were compared, using enzyme activities standardized against the glycosynthase reaction between Glcβ4Glcβ3GlcαF and GlcPNP.

Insoluble polysaccharides synthesized by E134S were obtained in 60% yield after 1 day (Table 4, polymers 12 and 14). Since the enzyme concentration used was relatively low, soluble polymers and un-reacted donor substrate were recovered. The molecular-mass distribution presented a bimodal profile (Figure 5A) at Mp values of 17 kDa and ∼7 kDa. The Mw value of 15 kDa corresponds to insoluble polymers with DP of 90 glucosyl units. Interestingly, reactions catalysed by E134S-CBM11 gave insoluble polysaccharides in 68% yield at low enzyme activity but in quantitative yield at high enzyme activity (Table 4, entries 13 and 15). Molecular mass distribution of polymer 15 presented a bimodal profile (Mp of 39 and 21 Da) resulting in a Mw value of 20 kDa composed of 121 glucosyl units (Figure 5A). In these conditions E134S-CBM11 generated slightly longer insoluble polysaccharides compared to E134S (≈33%; 14). The morphology of polymer 15 (Figure 6B) showed the typical amorphous precipitate of the (4Glcβ4Glcβ3Glcβ)n structures obtained by E134S [16].

Table 4
Polymerization reactions of donor Glcβ4Glcβ3GlcαF by E134S compared with E134S-CBM11 fusion protein

Conditions: [Donor]=150 mM, phosphate buffer, pH 7.0, 0.1 mM CaCl2, 37°C, 24 h.

Entry Enzyme Activity (U)* Mw (kDa) n (DP) Mn (kDa) Yield (%) PDI Mp (kDa) 
12 E1340.38 15 30 (90) 57 1.7 17, 4 
13 E134S-CBM11 0.35 18 37 (110) 68 1.9 39, 19 
14 E1340.94 15 30 (90) 60 1.7 17, 7 
15 E134S-CBM11 0.85 20 40 (121) 10 100 2.0 39, 21 
Entry Enzyme Activity (U)* Mw (kDa) n (DP) Mn (kDa) Yield (%) PDI Mp (kDa) 
12 E1340.38 15 30 (90) 57 1.7 17, 4 
13 E134S-CBM11 0.35 18 37 (110) 68 1.9 39, 19 
14 E1340.94 15 30 (90) 60 1.7 17, 7 
15 E134S-CBM11 0.85 20 40 (121) 10 100 2.0 39, 21 

*Enzyme activity for the reference glycosynthase reaction Glcβ4Glcβ3GlcαF + GlcpNP.

†Polymer parameters as defined in Table 1.

HPSEC chromatograms of (4Glcβ4Glcβ3Glcβ)n polymers
Figure 5
HPSEC chromatograms of (4Glcβ4Glcβ3Glcβ)n polymers

(A) Polysaccharide 14 obtained with E134S and polysaccharide 15 obtained with the fusion E134S-CBM11 protein. (B) Polysaccharides 16 and 18 obtained with the fusion protein + CBM11. Labels correspond to entries in Tables 4 and 5. 

Figure 5
HPSEC chromatograms of (4Glcβ4Glcβ3Glcβ)n polymers

(A) Polysaccharide 14 obtained with E134S and polysaccharide 15 obtained with the fusion E134S-CBM11 protein. (B) Polysaccharides 16 and 18 obtained with the fusion protein + CBM11. Labels correspond to entries in Tables 4 and 5. 

SEM micrographs of freeze-dried polymers (4Glcβ4Glcβ3Glcβ)n
Figure 6
SEM micrographs of freeze-dried polymers (4Glcβ4Glcβ3Glcβ)n

(A) Polymer 11, (B) Polymer 15, (C) Polymer 17 and (D) Polymer 20 (Tables 2, 4 and 5).

Figure 6
SEM micrographs of freeze-dried polymers (4Glcβ4Glcβ3Glcβ)n

(A) Polymer 11, (B) Polymer 15, (C) Polymer 17 and (D) Polymer 20 (Tables 2, 4 and 5).

When CBM11 was added to the polymerization reactions performed by E134S-CBM11, no significant change in product profile was observed with respect to DP, Mp, Mw and morphology (Table 5; Figures 5B, 6C and 6D).

Table 5
Polymerization reactions by E134S-CBM11 fusion protein with added CBM11 (Glcβ4Glcβ3GlcαF donor)

Conditions: [Donor]=150 mM, enzyme 0.5 units (for the reference glycosynthase reaction Glcβ4Glcβ3GlcαF + GlcpNP, [CBM]=0–720 μM), phosphate buffer, pH 7.0, 0.1 mM CaCl2, 37°C, 24 or 72 h.

 Molar ratio E134S- Glcβ4Glcβ3GlcαF 1day 
Entry CBM11/CBM11 Mw1 (kDa) n (DP) Mn (kDa) Yield (%) PDI Mp (kDa) 
16 1:1 17 35 (105) 91 1.9 39, 18 
17 1:2 17 35 (105) 12 87 1.9  
18 1:3 16 33 (99) 81 1.9  
  Glcβ4Glcβ3GlcαF 3 days 
19 1:3 18 37 (111) 80 1.9 39, 18 
20 1:6 17 35 (105) 91 2.0  
 Molar ratio E134S- Glcβ4Glcβ3GlcαF 1day 
Entry CBM11/CBM11 Mw1 (kDa) n (DP) Mn (kDa) Yield (%) PDI Mp (kDa) 
16 1:1 17 35 (105) 91 1.9 39, 18 
17 1:2 17 35 (105) 12 87 1.9  
18 1:3 16 33 (99) 81 1.9  
  Glcβ4Glcβ3GlcαF 3 days 
19 1:3 18 37 (111) 80 1.9 39, 18 
20 1:6 17 35 (105) 91 2.0  

*Polymer parameters as defined in Table 1.

All insoluble polymers obtained by self-condensation of the trisaccharyl donor were quite similar when yields were over 80%. Polysaccharides with Mw of 17–20 kDa (DP=100–120) were obtained when the reaction was catalysed by the E134S mutant or by E134S-CBM11, with or without additional CBM11 (Tables 2 and 5). The rates of polymerization were high, thus there was not sufficient time for inter-chain interactions to occur, which leads to polymer precipitation around these Mw values. By contrast, when yields did not reach 80%, due to a lower rate of polymerization, addition of CBM11 or the fusion protein enabled longer insoluble polysaccharides to be generated, compared with the E134S-catalysed reaction. In the present study, the CBM11 prevented the nascent polymers from inter-chain interactions enabling continued polymerization of these molecules in solution before the reaction was terminated by product precipitation.

CONCLUSION

We explored the effect of a CBM on glycosynthase-catalysed polymerization to generate mixed-linked 1,3-1,4-β-glucans with regular sequences. The observed effect is dependent on the rate of polysaccharide formation. When the rate of polymerization is high, polymer yields are higher than 80% and CBM11, either as a discrete protein or appended to E134S, had no effect on the DP. In contrast, when the rate of polymerization was low and polymer yields were below 80%, the CBM11 facilitated oversaturation of the polymers and longer insoluble polysaccharides were obtained. Moreover, in the case of the alternating polysaccharide (4Glcβ3Glcβ)n, the presence of the CBM disrupts the crystallinity of the insoluble polymer obtained yielding amorphous precipitates instead of the characteristic spherulite morphology.

For the system studied, the observed effect on DP is not very large but significant, probably reflecting the non-processive behaviour of endo-1,3-1,4-β-glucanase. For a truly processive glycosynthase it might be expected that the synthesized oligomeric chain protruding from the active site binds the CBM which may keep the polymer in solution during chain extension. For a non-processive enzyme, polymer extension occurs by association and dissociation of intermediate products into the active site and the concentration of the intermediate polymers in solution is higher than the concentration of CBM (which is in the range of enzyme concentration). The concentration of CBM should have been much higher, at least equimolar to the polymer formed, to keep the product in solution. Therefore the solubilization effect provided by the CBM is not as extensive as would be expected for a processive enzyme. The effect on DP increase, in the present study, observed for slow polymerization reactions agrees with this hypothesis.

In conclusion, CBM assists glycosynthase-catalysed polymerizations to achieve longer polysaccharides, potentially being a general strategy in the design of efficient glycosynthases aimed at the production of artificial polysaccharides

AUTHOR CONTRIBUTION

Magda Faijes and Antoni Planas designed the work. Victoria Codera performed the experimental work. Victoria Codera, Harry Gilbert, Magda Faijes and Antoni Planas analysed the data and wrote the manuscript.

We aknowledge Dr Xevi Biarnés (IQS) for the modeling analysis of the fusion protein. VC acknowledges a predoctoral fellowship from the Institut Químic de Sarrià.

FUNDING

This work was supported by the MINECO [grant number BIO2013-49022-C2-1-R (to A.P.)].

Abbreviations

     
  • CBM

    carbohydrate-binding module

  •  
  • DP

    degree of polymerization

  •  
  • GH

    glycoside hydrolase

  •  
  • IMAC

    immobilized metal affinity chromatography

  •  
  • LB

    Luria—Bertani

  •  
  • Mn

    number average molecular mass

  •  
  • Mp

    molecular mass of the peak maximum

  •  
  • Mw

    weight average molecular mass

  •  
  • PDI

    polydispersity index

References

References
1
Gilbert
H.J.
Knox
J.P.
Boraston
A.B.
Advances in understanding the molecular basis of plant cell wall polysaccharide recognition by carbohydrate-binding modules
Curr. Opin. Struct. Biol.
2013
, vol. 
23
 (pg. 
669
-
677
)
[PubMed]
2
Boraston
A.B.
Bolam
D.N.
Gilbert
H.J.
Davies
G.J.
Carbohydrate-binding modules: fine-tuning polysaccharide recognition
Biochem. J.
2004
, vol. 
781
 (pg. 
769
-
781
)
3
Gilbert
H.J.
The biochemistry and structural biology of plant cell wall deconstruction
Plant Physiol.
2010
, vol. 
153
 (pg. 
444
-
455
)
[PubMed]
4
Hervé
C.
Rogowski
A.
Blake
A.W.
Marcus
S.E.
Gilbert
H.J.
Knox
J.P.
Carbohydrate-binding modules promote the enzymatic deconstruction of intact plant cell walls by targeting and proximity effects
Proc. Natl. Acad. Sci. U.S.A.
2010
, vol. 
107
 (pg. 
15293
-
15298
)
[PubMed]
5
Cuskin
F.
Flint
J.E.
Gloster
T.M.
Morland
C.
Baslé
A.
Henrissat
B.
Coutinho
P.M.
Strazzulli
A.
Solovyova
A.S.
Davies
G.J.
, et al. 
How nature can exploit nonspecific catalytic and carbohydrate binding modules to create enzymatic specificity
Proc. Natl. Acad. Sci. U.S.A.
2012
, vol. 
109
 (pg. 
20889
-
20894
)
[PubMed]
6
Meekins
D.A.
Raththagala
M.
Husodo
S.
White
C.J.
Guo
H.-F.
Kötting
O.
Vander Kooi
C.W.
Gentry
M.S.
Phosphoglucan-bound structure of starch phosphatase Starch Excess4 reveals the mechanism for C6 specificity
Proc. Natl. Acad. Sci. U.S.A.
2014
, vol. 
111
 (pg. 
7272
-
7277
)
[PubMed]
7
Montanier
C.
Money
V.A.
Pires
V.M.R.
Flint
J.E.
Pinheiro
B.A.
Goyal
A.
Prates
J.A.M.
Izumi
A.
Stalbrand
H.
Morland
C.
, et al. 
The active site of a carbohydrate esterase displays divergent catalytic and noncatalytic binding functions
PLoS Biol.
2009
, vol. 
7
 (pg. 
687
-
697
)
8
Lombard
V.
Golaconda Ramulu
H.
Drula
E.
Coutinho
P.M.
Henrissat
B.
The carbohydrate-active enzymes database (CAZy) in 2013
Nucleic Acids Res.
2014
, vol. 
42
 (pg. 
490
-
495
)
9
Wang
L.-X.
Huang
W.
Enzymatic transglycosylation for glycoconjugate synthesis
Curr. Opin. Chem. Biol.
2009
, vol. 
13
 (pg. 
592
-
600
)
[PubMed]
10
Faijes
M.
Planas
A.
In vitro synthesis of artificial polysaccharides by glycosidases and glycosynthases
Carbohydr. Res.
2007
, vol. 
342
 (pg. 
1581
-
1594
)
[PubMed]
11
Shaikh
F.A.
Withers
S.G.
Teaching old enzymes new tricks: engineering and evolution of glycosidases and glycosyl transferases for improved glycoside synthesis
Biochem. Cell Biol.
2008
, vol. 
86
 (pg. 
169
-
177
)
[PubMed]
12
Perugino
G.
Trincone
A.
Rossi
M.
Moracci
M.
Oligosaccharide synthesis by glycosynthases
Trends Biotechnol.
2004
, vol. 
22
 (pg. 
31
-
7
)
[PubMed]
13
Faijes
M.
Imai
T.
Bulone
V.
Planas
A.
In vitro synthesis of a crystalline (1–>3,1–>4)-beta-D-glucan by a mutated (1–>3,1–>4)-beta-D-glucanase from Bacillus
Biochem. J.
2004
, vol. 
380
 (pg. 
635
-
641
)
[PubMed]
14
Gullfot
F.
Ibatullin
F.M.
Sundqvist
G.
Davies
G.J.
Brumer
H.
Functional characterization of xyloglucan glycosynthases from GH7, GH12, and GH16 scaffolds
Biomacromolecules
2009
, vol. 
10
 (pg. 
1782
-
1788
)
[PubMed]
15
Spadiut
O.
Ibatullin
F.M.
Peart
J.
Gullfot
F.
Martinez-Fleites
C.
Ruda
M.
Xu
C.
Sundqvist
G.
Davies
G.J.
Brumer
H.
Building custom polysaccharides in vitro with an efficient, broad-specificity xyloglucan glycosynthase and a fucosyltransferase
J. Am. Chem. Soc.
2011
, vol. 
133
 (pg. 
10892
-
10900
)
[PubMed]
16
Pérez
X.
Faijes
M.
Planas
A.
Artificial mixed-linked β-glucans produced by glycosynthase-catalyzed polymerization: tuning morphology and degree of polymerization
Biomacromolecules
2011
, vol. 
12
 (pg. 
494
-
501
)
[PubMed]
17
Carvalho
A.L.
Goyal
A.
Prates
J.A.M.
Bolam
D.N.
Gilbert
H.J.
Pires
V.M.R.
Ferreira
L.M.A.
Planas
A.
Romão
M.J.
Fontes
C.M.G.A.
The family 11 carbohydrate-binding module of Clostridium thermocellum Lic26A-Cel5E accommodates beta-1,4- and beta-1,3-1,4-mixed linked glucans at a single binding site
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
34785
-
34793
)
[PubMed]
18
Malet
C.
Planas
A.
From beta-glucanase to beta-glucansynthase: glycosyl transfer to alpha-glycosyl fluorides catalyzed by a mutant endoglucanase lacking its catalytic nucleophile
FEBS Lett.
1998
, vol. 
440
 (pg. 
208
-
212
)
[PubMed]
19
Malet
C.
Vallbs
J.
Bou
J.
Planas
A.
A specific chromophoric substrate for activity assays of
J. Biotechnol.
1996
, vol. 
48
 (pg. 
209
-
219
)
[PubMed]
20
Andrés
E.
Aragunde
H.
Planas
A
Screening glycosynthase libraries with a fluoride chemosensor assay independent of enzyme specificity
Biochem. J.
2014
, vol. 
458
 (pg. 
1
-
23
)
[PubMed]
21
Cuff
J.A.
Clamp
M.E.
Siddiqui
A.S.
Finlay
M.
Barton
G.J.
JPred: a consensus secondary structure prediction server
Bioinformatics
1998
, vol. 
14
 (pg. 
892
-
893
)
[PubMed]
22
Eswar
N.
Webb
B.
Marti-Renom
M.-A.
Madhusudhan
M.S.
Eramian
D.
Shen
M.
Pieper
U.
Sali
A.
Comparative protein structure modeling using modeller
Curr. Protoc. Protein Sci.
2006
, vol. 
65
 (pg. 
712
-
7254448
)
23
Humphrey
W.
Dalke
A.
Schulten
K.
VMD: visual molecular dynamics
J. Mol. Graph.
1996
, vol. 
14
 (pg. 
33
-
38
)
[PubMed]
24
Carraher
C.E.J.
Seymour/Carraher's Polymer Chemistry
2007
7th edn
CRC Press
25
Din
N.
Damude
H.G.
Gilkes
N.R.
Miller
R.C.
Warren
R.A.J.
Kilburn
D.G.
C1-Cx, revisited: Intramolecular synergism in a cellulase
Proc. Natl. Acad. Sci. U.S.A.
1994
, vol. 
91
 (pg. 
11383
-
11387
)
[PubMed]
26
Gustavsson
M.
Lehtio
J.
Denman
S.
Teeri
T.T.
Hult
K.
Martinelle
M.
Stable linker peptides for a cellulose-binding domain–lipase fusion protein expressed in Pichia pastoris Fusion proteins composed of a cellulose-binding domain
Protein Eng.
2001
, vol. 
14
 (pg. 
711
-
715
)
[PubMed]
27
Kavoosi
M.
Creagh
A.L.
Kilburn
D.G.
Haynes
C.A.
Strategy for selecting and characterizing linker peptides for cbm9-tagged fusion proteins expressed in Escherichia coli
Biotechnol. Bioeng.
2007
, vol. 
98
 (pg. 
599
-
610
)
[PubMed]

Supplementary data