New Family of Carbohydrate-Binding Modules Defined by a Galactosyl-Binding Protein Module from a Cellvibrio japonicus Endo-Xyloglucanase

This study reveals carbohydrate-binding module family 88 (CBM88) as a new family of galactose-binding protein modules, which are found in series with diverse microbial glycoside hydrolases, polysaccharide lyases, and carbohydrate esterases. The definition of CBM88 in the carbohydrate-active enzymes classification (http://www.cazy.org/CBM88.html) will significantly enable future microbial (meta)genome analysis and functional studies.

ABSTRACT

Carbohydrate-binding modules (CBMs) are usually appended to carbohydrate-active enzymes (CAZymes) and serve to potentiate catalytic activity, for example, by increasing substrate affinity. The Gram-negative soil saprophyte Cellvibrio japonicus is a valuable source for CAZyme and CBM discovery and characterization due to its innate ability to degrade a wide array of plant polysaccharides. Bioinformatic analysis of the CJA_2959 gene product from C. japonicus revealed a modular architecture consisting of a fibronectin type III (Fn3) module, a cryptic module of unknown function (X181), and a glycoside hydrolase family 5 subfamily 4 (GH5_4) catalytic module. We previously demonstrated that the last of these, CjGH5F, is an efficient and specific endo-xyloglucanase (M. A. Attia, C. E. Nelson, W. A. Offen, N. Jain, et al., Biotechnol Biofuels 11:45, 2018, https://doi.org/10.1186/s13068-018-1039-6). In the present study, C-terminal fusion of superfolder green fluorescent protein in tandem with the Fn3-X181 modules enabled recombinant production and purification from Escherichia coli. Native affinity gel electrophoresis revealed binding specificity for the terminal galactose-containing plant polysaccharides galactoxyloglucan and galactomannan. Isothermal titration calorimetry further evidenced a preference for galactoxyloglucan polysaccharide over short oligosaccharides comprising the limit-digest products of CjGH5F. Thus, our results identify the X181 module as the defining member of a new CBM family, CBM88. In addition to directly revealing the function of this CBM in the context of xyloglucan metabolism by C. japonicus, this study will guide future bioinformatic and functional analyses across microbial (meta)genomes.

IMPORTANCE This study reveals carbohydrate-binding module family 88 (CBM88) as a new family of galactose-binding protein modules, which are found in series with diverse microbial glycoside hydrolases, polysaccharide lyases, and carbohydrate esterases. The definition of CBM88 in the carbohydrate-active enzymes classification (http://www.cazy.org/CBM88.html) will significantly enable future microbial (meta)genome analysis and functional studies.

INTRODUCTION

Plant biomass is regarded as a valuable source of renewable energy and biomaterials in light of global climate change and the continuous depletion of petroleum reserves (1–3). Nevertheless, plant biomass utilization is hindered by the intrinsic complexity of the plant cell wall, which arises from the sophisticated assembly of complex polysaccharides and other components. As part of the natural carbon cycle, bacteria and fungi have evolved efficient machineries to break down the tremendous diversity of polysaccharide structures via broad repertoires of carbohydrate-active enzymes (CAZymes) (4, 5). CAZymes, including glycoside hydrolases (GHs), polysaccharide lyases (PLs), carbohydrate esterases (CEs), and lytic polysaccharide mono-oxygenases (LPMOs), are often appended to noncatalytic carbohydrate-binding modules (CBMs) (reviewed in references 6 to 14), which potentiate activity by substrate targeting and proximity effects (15, 16).

Like CAZymes, CBMs are classified into families based on amino acid sequence identity (4; http://www.cazy.org/Carbohydrate-Binding-Modules.html). As of October 2020, 87 families of CBMs have been defined, comprising more than 233,000 sequences from GenBank. The majority of these families contain members that selectively bind plant cell wall polysaccharides (7). CBMs are further classified based on the mode of carbohydrate binding (7). Type A CBMs bind to the crystalline surfaces of insoluble polysaccharide substrates and typically possess flat binding faces. Type B CBMs bind single-glycan chains of soluble polysaccharides internally (i.e., endo-binding), often with cleft-shaped protein surfaces. Type C CBMs bind the termini of glycan chains or short oligosaccharides (i.e., exo-binding) via pocket-shaped binding sites. The number of CBM families is continuously expanding, with 15 new families having been discovered in the past 5 years alone (4, 5, 17–26). The identification and characterization of new CBMs remain a significant challenge (9), and like CAZymes (27), there are likely many potential CBM families lurking in microbial genomes.

The soil-dwelling saprophyte Cellvibrio japonicus is a Gram-negative bacterium that is able to utilize nearly all plant cell wall polysaccharides and, as such, has been a historically important source for CAZyme discovery (28, 29). We recently characterized the ensemble of GHs constituting the xyloglucan utilization system of C. japonicus (30–35). During the course of these studies, we identified a protein module of unknown function in tandem with the endo-xyloglucanase CjGH5F encoded by gene locus CJA_2959 (Fig. 1A) (32). This module had been observed in diverse bacterial genomes and was denoted “X181” in the CAZy database (36; B. Henrissat, personal communication). In light of its size and association with CjGH5F and other CAZymes, we anticipated that X181 may comprise a noncatalytic CBM. To test this hypothesis, we produced the module as a recombinant fusion protein and examined its binding to cellulose and a range of amorphous plant cell wall polysaccharides. Our results indicate that X181 binds terminal galactosyl residues on polysaccharides and constitutes the first representative of the new family CBM88.

FIG 1.

Modular architecture of the native C. japonicus CJA_2959 gene product and different constructs used in the study. (A) The full-length gene product is composed of a signal peptide, an Fn3 domain, an X181 module, and a catalytic GH5_4 (CjGH5F) domain. (B) Recombinant proteins produced for characterization using the E. coli expression vector pET28a and the ligation-independent cloning (LIC) vectors pMCSG53, pMCSG69, and pMCSG-GST-TEV (67). MBP, GST, and sfGFP are connected to X181 or Fn3-X181 by a TEV cleavage site.

(This research was conducted by Mohamed A. Attia in partial fulfillment of the requirements for a Ph.D. from the University of British Columbia, Vancouver, BC, Canada [73].)

RESULTS

Bioinformatic analysis.

Gene locus CJA_2959 (GenBank accession no. ACE86198.1) encodes a multimodular protein comprising, in order, a secretion signal peptide (Met1 to Ala22, predicted signal peptidase I cleavage site) (37), a fibronectin type III (Fn3) domain (Gln23 to Cys120), an X181 module (Thr123 to Val205), and the CjGH5F catalytic module (Ser218 to Gln556) (Fig. 1A). A short Gly-Ser-Thr-rich sequence was also observed between the X181 and GH5 modules, which may constitute a flexible linker (38). CjGH5F was previously established as a strict endo-xyloglucanase following module dissection and independent recombinant production; however, the functions of the other modules were not previously characterized (32). Fn3 domains (PFAM PF00041) possess a small beta-sandwich structure and are frequently found in large animal cell and matrix adhesion proteins (e.g., fibronectin) (39) as well as in microbial GHs (40). Although little is known regarding specific functions of Fn3 domains, they generally appear to play structural roles as spacers between binding, catalytic, and/or membrane-anchoring motifs (39–41).

Recombinant protein production and purification.

To characterize the potential biochemical function of the X181 module independent of the CjGH5F catalytic module, we explored several truncated constructs containing a hexahistidine tag to facilitate purification following recombinant expression in Escherichia coli (Fig. 1B). Initial attempts to produce His₆-X181, as well as His₆-Fn3-X181, independently and in fusion with glutathione S-transferase (GST) (42) were unsuccessful in both E. coli BL21 and E. coli Rosetta strains (see Fig. S1 in the supplemental material). We observed an enhancement of the production of soluble Fn3-X181 in the Rosetta strain when fused to an N-terminal maltose-binding protein (MBP) (42) (Fig. S1). Unfortunately, this protein was proteolytically unstable during the purification process, making it impossible to obtain intact (data not shown).

To overcome this hurdle, we explored protein fusions with “superfolder” green fluorescent protein (SfGFP) (43) (Fig. 1B), which recently enabled us to produce and characterize two cellulose-specific CBMs from C. japonicus (31). Particular advantages of GFP as a fusion partner are the lack of nonspecific carbohydrate binding and the facile photometric detection (31, 44–48). As observed with the N-terminal MBP fusion described above, His₆-sfGFP-Fn3-X181 was highly prone to proteolysis during the purification process (Fig. S2). All efforts to minimize this proteolysis by decreasing expression and purification temperatures and by using protease inhibitors were fruitless (data not shown).

C-terminal sfGFP fusions were explored in parallel (Fig. 1B). Recombinant production of the simpler His₆-X181-GFP fusion was unsuccessful, with extremely low yields in the soluble fractions, in contrast to His₆-Fn3-X181-sfGFP. Although most of the His₆-Fn3-X181-sfGFP fusion protein was produced as insoluble inclusion bodies, we were nonetheless able to obtain modest yields of intact protein from the bacterial lysate (typically ca. 6  mg per liter of culture medium after purification [Fig. S2]; calculated mass, 48,674 Da, observed by electrospray ionization-mass spectrometry [ESI-MS], 48,672 Da [Fig. S3]). Hence, His₆-Fn3-X181-sfGFP was used for subsequent characterization. It is notable that this construct is the most homologous to the native protein, with the sfGFP directly replacing the GH5_4 module in series (Fig. 1).

Carbohydrate-binding specificity.

To explore the potential carbohydrate-binding ability of the X181 module in the context of its natural fusion to the endo-xyloglucanase CjGH5F, we employed a panel of common plant cell wall polysaccharides, including cellulose and several amorphous matrix glycans (hemicelluloses). We have previously observed that C. japonicus produces a GH74 endo-xyloglucanase in tandem with a cellulose-specific CBM2 member (31). In an analogous pulldown assay, recombinant His₆-Fn3-X181-sfGFP was incubated with an aqueous suspension of Avicel (microcrystalline cellulose). SDS-PAGE of the supernatant and pellet fractions versus sfGFP and sfGFP-CjCBM2 (31), as negative and positive controls, respectively, revealed no significant cellulose-binding affinity of the X181 module (Fig. 2).

FIG 2.

Binding capacity of sfGFP, Fn3-X181-sfGFP, and sfGFP-CjCBM2 for Avicel. Recombinant sfGFP (27.7 kDa; lanes 1 and 2), His₆-Fn3-X181-sfGFP (48.7 kDa; lanes 3 and 4), and sfGFP-CjCBM2 (40.7 kDa; lanes 5 and 6) were incubated with Avicel, and unbound proteins were removed by centrifugation (lanes 1, 3, 5, and 7). Bound proteins were released from Avicel by boiling in SDS-PAGE loading dye (lanes 2, 4, 6, and 8). sfGFP was used as a negative control, while sfGFP-CjCBM2 (31) was used as a positive control.

Binding to the soluble plant cell wall polysaccharides barley β-glucan, tamarind galactoxyloglucan, konjac glucomannan, carob (locust bean) galactomannan, guar galactomannan, and beechwood xylan, as well as the artificial hydroxyethyl cellulose, was screened by native affinity gel electrophoresis (Fig. 3). Strikingly, retardation of migration was observed in gels containing tamarind galactoxyloglucan, carob galactomannan, and guar galactomannan, which share terminal galactosyl (t-Gal) residues on branches as a commonality. To exclude the possibility that X181 is catalytically active, His₆-Fn3-X181-sfGFP was incubated with tamarind galactoxyloglucan and subjected to the sensitive bicinchoninic acid (BCA)-reducing sugar assay, which was negative.

FIG 3.

Affinity gel electrophoresis for Fn3-X181-sfGFP against different polysaccharide substrates. Recombinant protein bands were visualized by fluorescence (A) and Coomassie blue staining (B) following electrophoresis at 90 V for 10 h. The negative-control sfGFP completely migrated through under these conditions. (C) Gel following 5.5 h of electrophoresis, which demonstrates no binding of the negative-control sfGFP toward tamarind galactoxyloglucan (fluorescence visualization). BBG, barley β-glucan; HEC, hydroxyethyl cellulose; KGM, konjac glucomannan; XyG, tamarind galactoxyloglucan; GalMan, galactomannan.

To further elucidate the specificity of the X181 module in the context of the full-length CJA_2959 gene product, we used isothermal titration calorimetry (ITC) to measure the binding affinity of Fn3-X181-sfGFP for tamarind galactoxyloglucan and for the corresponding limit-digest products of CjGH5F (Fig. 4 and Table 1). The latter comprises a mixture of oligosaccharides based on a Glc₄ backbone, namely, XXXG (Glc₄Xyl₃), XLXG (Glc₄Xyl₃Gal), XXLG (Glc₄Xyl₃Gal), and XLLG (Glc₄Xyl₃Gal₂) in the approximate ratio of 2:1:3:3 (32; see reference 49 for xyloglucan oligosaccharide nomenclature). Commensurate with the affinity gel electrophoresis results, ITC revealed a high association constant for the galactoxyloglucan polysaccharide (K_a = 7.17 × 10³ M⁻¹, based on a molar equivalent concentration of the Glc₄-based oligosaccharide units). The affinity for the limit-digest products was correspondingly lower (K_a = 1.11 × 10³ M⁻¹), as might be expected to facilitate product release. The combined thermodynamic data indicate that substrate binding is enthalpically driven, as has been observed for the majority of CBMs studied to date (50–52).

FIG 4.

Isothermal titration calorimetry of Fn3-X181-sfGFP against different ligands. (A) Tamarind seed xyloglucan. (B) Glc₄-based tamarind seed xyloglucan oligosaccharides. The top graph in each pair shows the raw heat during titration, while the bottom graph shows the integrated heat after correction.

TABLE 1.

Summary of the thermodynamic parameters for Fn3-X181-sfGFP obtained by isothermal titration calorimetry

Carbohydrate	Ka (M−1)	ΔG° (kcal mol−1)	ΔH° (kcal mol−1)	TΔS° (kcal mol−1)
XyGa	(7.17 ± 0.52) × 103	−5.3	−10 ± 0.4	−4.7
XyGO	(1.11 ± 0.04) × 103	−4.2	−6.5 ± 0.1	−2.3

^aBinding thermodynamics of tamarind seed galactoxyloglucan (XyG) is based on the equivalent molar concentration of Glc₄-based xyloglucan oligosaccharide (XyGO) units.

DISCUSSION

CBMs are commonly found appended via flexible linkers to glycoside hydrolases in unique multimodular architectures (6–14, 38). Here, we identified and functionally characterized a unique protein module from the CJA_2959 gene product in the saprophyte C. japonicus, which establishes the new family CBM88 in the CAZy classification (4). This CBM exclusively binds the t-Gal-containing plant cell wall polysaccharides galactoxyloglucan and galactomannan. The recognition of t-Galβ-(1,2)-Xylα-(1,6) branches on galactoxyloglucan by the CBM is congruent with its partnership with the specific endo-xyloglucanase CjGH5F (32) (Fig. 1A).

More broadly, CBM88 members are primarily found in predicted multimodular CAZymes from Bacteroidetes and Gammaproteobacteria, comprising diverse GH5 (predicted endo-xyloglucanases or endo-mannanases) (53), GH16 (predicted endo-glucanases or endo-galactanases) (54), GH26 (predicted endo-mannanases) (55), GH53 (predicted endo-galactanases) (56), GH74 (predicted endo-xyloglucanases) (57), PL11 (predicted rhamnogalacturonan lyases) (58), and carbohydrate esterase catalytic modules (see http://www.cazy.org/CBM88.html for an actively updated list of sequences and source organisms) (4). These families are united by a specificity for galactans or galactosyl-branched polysaccharides, which further rationalizes the observed t-Gal specificity of CBM88. In particular, GH74 is a family of specific endo-(galacto)xyloglucanases (4, 5, 57), whereas GH26 is primarily comprised of endo-(galacto)mannanases (4, 5, 55). Interestingly, CBM88 modules are also found in sequence with GH8, GH10, GH30 (59), and GH43 (60) modules, which are classically associated with xylan hydrolysis (5), suggesting potential alternative binding specificities within the family.

We note here that our inability to produce the CBM88 module independently of the Fn3 domain formally precludes us from ruling out a role of the latter in carbohydrate binding. On the other hand, Fn3 domains are generally not known to have this function (39–41). Furthermore, CBM88 is otherwise not preceded by an Fn3 domain in the vast majority of other multimodular protein sequences currently known, with the CJA_2959 gene product studied here being a rare exception (http://www.cazy.org/CBM88.html) (4).

Our attempts to crystallize the sfGFP fusion (61) and the Fn3-X181 protein following tobacco etch virus (TEV) cleavage have currently been unsuccessful, thus precluding definitive structure-function analysis. Yet, we anticipate that CjCBM88 may be a type C CBM that directs the endo-xyloglucanase to the polysaccharide through the recognition of side chain termini. Such an end-on binding mode is reminiscent of the interaction of a Hungateiclostridium thermocellum CBM62 with t-Gal residues on both galactoxyloglucan oligosaccharides (PDB ID 2YFZ) and galactomannan oligosaccharides (PDB IDs 2YB7 and 2YG0) (52). This binding mode is notably distinct from that observed in interactions of an exemplar type B CBM65 with xyloglucan, which involve binding to the β-(1,4)-glucan backbone and α-(1,6)-xylosyl side chains (PDB ID 2YPJ) (51). Indeed, backbone recognition is common among other type B CBMs (e.g., CBM44) (62), type A CBMs (e.g., CBM2 and CBM3) (63), and cell surface glycan-binding proteins (64, 65), which demonstrate various degrees of preference for the branched xyloglucan chains versus the homologous β-(1,4)-glucan, cellulose.

To conclude, we anticipate that the definition of CBM88 in the CAZy classification will aid bioinformatic analysis of (meta)genomic sequences through refined modular dissection of complex CAZymes (4, 36). We also look forward to future structural biology studies to provide molecular insight into the details of carbohydrate recognition by this nascent CBM family.

MATERIALS AND METHODS

Bioinformatic analysis.

The modular architecture of the full-length protein Fn3-X181-CjGH5F, encoded by C. japonicus gene locus CJA_2959 (GenBank accession no. ACE86198.1) (29), was obtained from BLASTP analysis and alignment with representative GH modules from the CAZy database (4) using ClustalW (66). The boundaries of the X181 module were generously provided by Bernard Henrissat (head of the CAZy Database, AFMB-CNRS, Marseille, France) (36).

Plasmid construction.

All PCR primers are listed in Table 2. The open reading frames (ORFs) for X181 and Fn3-X181 were PCR amplified from C. japonicus Ueda107 genomic DNA. Amplified products were double digested with NheI and XhoI, gel purified, and then ligated to the respective site of pET28a so that they were fused to an N-terminal hexahistidine (His₆) tag. For GFP fusion, overlapping extension PCR was used to fuse the sfGFP sequence to either the N or C terminus of the Fn3-X181 and X181 domain cDNA, with a tobacco etch virus (TEV) protease cleavage site in between. TEV-Fn3-X181, Fn3-X181-TEV, and X181-TEV were PCR amplified from C. japonicus genomic DNA, while sfGFP was amplified from a pBAD vector harboring the sfGFP gene (GenBank accession no. AGT98536.1) (generously provided by Lawrence McIntosh, University of British Columbia). Purified sfGFP was mixed individually with the purified TEV-Fn3-X181, Fn3-X181-TEV, and X181-TEV, and the respective fusion products sfGFP-TEV-Fn3-X181, Fn3-X181-TEV-sfGFP, and X181-TEV-sfGFP were obtained by PCR amplification. The purified sfGFP-TEV-Fn3-X181, Fn3-X18-TEV-sfGFP, and X18-TEV-sfGFP DNA fragments were cloned in pET28a as previously described for Fn3-X181 after double digesting with NheI and XhoI and ligating the digested products in the respective sites of the pET28a vector. The amplified Fn3-X181 product was also individually ligated in SspI-linearized pMCSG53, pMCSG69, and pMCSG-GST-TEV (a vector derived from pMCSG52) using ligation-independent cloning (LIC) strategy (67) so that the ORF Fn3-X181 was fused to N-terminal His₆, MBP-His₆, and His₆-GST, respectively. Successful cloning was confirmed by PCR and plasmid DNA sequencing. Q5 high-fidelity DNA polymerase was used for all the PCR amplifications.

TABLE 2.

PCR primers used in this study

Primer	Sequence	Recombinant protein(s)
X181-NheI-F	5′-GACCGCTAGCATGACTGCGGTGACGCCCTATATCAAC-3′	X181
X181-XhoI-R	5′-GGTCCTCGAGTTAGACCGTTACGCTGAAAACCTGGC-3′	X181
FN3-NheI-F	5′-GACCGCTAGCATGCAGAATTGCGGCAGCGGTGGCG-3′	FN3-X181
X181-XhoI-R	5′-GGTCCTCGAGTTAGACCGTTACGCTGAAAACCTGGC-3′	FN3-X181
GFP-NheI-F	5′-GACCGCTAGCATGGTTAGCAAAGGTGAAGAAC-3′	GFP-TEVoh
GFP-TEVoh-R	5′-GAAAATAAAGATTCTCGCTGCCTTTATACAGTTC-3′	GFP-TEVoh
GFPoh-TEV-FN3-F	5′-CTGTATAAAGGCAGCGAGAATCTTTATTTTCAGGGCCAGAATTGCGGCAGCGGTGGCG-3′	GFPoh-TEV-FN3-X181
X181-XhoI-R	5′-GGTCCTCGAGTTAGACCGTTACGCTGAAAACCTGGC-3′	GFPoh-TEV-FN3-X181
GFP-NheI-F	5′-GACCGCTAGCATGGTTAGCAAAGGTGAAGAAC-3′	GFP-FN3-X181
X181-XhoI-R	5′-GGTCCTCGAGTTAGACCGTTACGCTGAAAACCTGGC-3′	GFP-FN3-X181
FN3-NheI-F	5′-GACCGCTAGCATGCAGAATTGCGGCAGCGGTGGCG-3′	FN3-X181-TEV-GFPoh
X181-TEV-GFPoh-R	5′-CTTCACCTTTGCTAACGCCCTGAAAATAAAGATTCTCGACCGTTACGCTGAAAACCTGG-3′	FN3-X181-TEV-GFPoh
TEVoh-GFP-F	5′-CTTTATTTTCAGGGCGTTAGCAAAGGTGAAGAAC-3′	TEVoh-GFP
GFP-XhoI-R	5′-GGTCCTCGAGTTAGCTGCCTTTATACAGTTCATC-3′	TEVoh-GFP
FN3-NheI-F	5′-GACCGCTAGCATGCAGAATTGCGGCAGCGGTGGCG-3′	FN3-X181-GFP
GFP-XhoI-R	5′-GGTCCTCGAGTTAGCTGCCTTTATACAGTTCATC-3′	FN3-X181-GFP
X181-NheI-F	5′-GACCGCTAGCATGACTGCGGTGACGCCCTATATCAAC-3′	X181-GFP
GFP-XhoI-R	5′-GGTCCTCGAGTTAGCTGCCTTTATACAGTTCATC-3′	X181-GFP
FN3-X181-LIC-F	5′-TACTTCCAATCCAATGCCATGCAGAATTGCGGCAGCGGTGGC-3′	FN3-X181, GST-FN3-X181, MBP-FN3-X181
FN3-X181-LIC-R	5′-TTATCCACTTCCAATGTTATCAGACCGTTACGCTGAAAAC-CTGGC-3′	FN3-X181, GST-FN3-X181, MBP-FN3-X181

Gene expression and protein purification.

To produce the recombinant proteins His₆-X181, His₆-Fn3-X181, His₆-sfGFP-Fn3-X181, His₆-X181-sfGFP, and His₆-Fn3-X181-sfGFP in E. coli, constructs were individually transformed into chemically competent Rosetta (DE3) cells. Colonies were grown on LB solid media containing kanamycin (50 µg ml⁻¹) and chloramphenicol (33 µg ml⁻¹). One colony of the transformed E. coli cells was inoculated in 10 ml of LB medium containing the same antibiotics and grown overnight at 37°C (200 rpm). The whole overnight culture was used to inoculate 1 liter of terrific broth medium containing the proper antibiotics. Cultures were grown at 37°C (200 rpm) until A₆₀₀ was equal to 0.6. Gene expression was induced by adding isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 0.1 mM. After induction, cultures were grown at 16°C (200 rpm) for 20 to 22 h. Cultures were then centrifuged, and pellets were resuspended in 5 ml of E. coli lysis buffer containing 20 mM HEPES, pH 7.0, 500 mM NaCl, 40 mM imidazole, 5% glycerol, 1 mM dithiothreitol (DTT), and 1 mM phenylmethylsulfonyl fluoride (PMSF). Cells were then lysed by sonication, and the clear supernatant was separated by centrifuging the sonicated cultures at 4°C (12,400 × g for 60 min). Recombinant proteins were purified from the clear soluble lysates using a Ni²⁺ affinity column and gradient elution in a fast protein liquid chromatography (FPLC) system utilizing 20 mM HEPES, pH 7.0, containing 100 mM NaCl, 500 mM imidazole, and 5% glycerol. The purity of the recombinant proteins was determined by SDS-PAGE. Pure fractions were pooled, concentrated, and buffer exchanged with 20 mM HEPES buffer, pH 7.0, containing 100 mM NaCl. Protein concentrations were determined using an Epoch microvolume spectrophotometer system (BioTek, USA) at 280 nm. The fidelity of protein translation was confirmed using intact protein mass spectrometry essentially as previously described (68).

When LIC vectors were used, individual constructs were transformed into the chemically competent BL21 and Rosetta (DE3) cells. An expression trial was conducted in 10 ml LB medium containing ampicillin (50 µg ml⁻¹) (BL21) or ampicillin (50 µg ml⁻¹) and chloramphenicol (33 µg ml⁻¹) (Rosetta) to identify constructs with successful soluble expression of the recombinant proteins. Briefly, cultures were grown at 37°C (200 rpm) until A₆₀₀ was equal to 0.6 before they were induced by IPTG at a final concentration of 0.2 mM. Cultures were then grown overnight at 16°C (200 rpm) before aliquots of 1 ml were taken out and centrifuged at 4°C in a bench-top centrifuge for 5 min (15,000 × g). Supernatants were discarded, and cell pellets were resuspended in 700 µl lysis buffer containing 20 mM HEPES, pH 7.0, 500 mM NaCl, 40 mM imidazole, and 5% glycerol. Cells were sonicated, and the clear supernatants were separated from pellets by centrifugation in a bench-top cooling centrifuge for 15 min (15,000 × g). Clear supernatants were mixed with 4× SDS-PAGE loading dye and boiled for 10 min before 10 to 15 µl were analyzed by SDS-PAGE. Pellets were resuspended in 150 to 300 µl of the same lysis buffer and mixed with 4× SDS-PAGE loading dye before 10 µl was analyzed on SDS-PAGE.

For the large-scale production of MBP-Fn3-X181 in the Rosetta strain, the same protocol described for the production of X181, Fn3-X181, sfGFP-Fn3-X181, Fn3-X181-sfGFP, and X181-sfGFP was employed with the exception of the antibiotic ampicillin (50 µg ml⁻¹), which was used instead of kanamycin.

Carbohydrate sources.

Tamarind seed galactoxyloglucan, konjac glucomannan, barley β-glucan, beechwood xylan, and carob galactomannan were purchased from Megazyme (Bray, Ireland). Hydroxyethylcellulose was purchased from Amresco (Solon, OH, USA). Guar galactomannan was purchased from Sigma-Aldrich (St. Louis, MO, USA). Glc₄-based xyloglucan oligosaccharides were prepared from tamarind seed galactoxyloglucan using the endo-xyloglucanase CjGH5E (32), essentially as previously described (69).

Cellulose-binding capacity.

Qualitative analysis of cellulose-binding capacity was performed as previously described (31). Briefly, 100 µg of the recombinant protein His₆-Fn3-X181-sfGFP was mixed with 10 mg Avicel type PH-101 in a 200-µl reaction volume containing 50 mM phosphate buffer (pH 7.0). The mixture was gently agitated while sitting on ice for 1 h. The supernatant was removed by centrifugation for 5 min (12,000 × g) and mixed with SDS-PAGE loading dye, and 6 µl was analyzed by SDS-PAGE. Avicel pellets were washed twice with 250 µl of 50 mM phosphate buffer (pH 7.0) and then resuspended in 200 µl of the same buffer. Forty microliters of 6× SDS-PAGE loading dye was added, and then bound proteins were released by boiling for 10 min before 6 µl was analyzed by SDS-PAGE.

Affinity gel electrophoresis.

The methodology for the native affinity gel electrophoresis was adapted from a previously described protocol (70). Briefly, 7 µg of the native recombinant His₆-Fn3-X181-sfGFP was loaded on native 10% (wt/vol) polyacrylamide gels containing 0.1% final concentration of the tested polysaccharides barley β-glucan, hydroxyethylcellulose, beechwood xylan, konjac glucomannan, tamarind seed galactoxyloglucan, carob galactomannan, and guar galactomannan. Electrophoresis was conducted at room temperature for approximately 10 h at 90 V due to the slow migration of the recombinant protein. Protein bands were visualized by both fluorescence (Bio-Rad gel documentation system with an excitation filter of 450 nm and an emission filter of 510 nm) and Coomassie blue staining.

Activity assays.

One microgram of His₆-Fn3-X181-sfGFP was added to 2 mg ml⁻¹ tamarind galacto-xyloglucan (XyG) in a 200-µl reaction volume containing 50 mM phosphate buffer, pH 7.0, and 2 mM CaCl₂. The reaction mixture was incubated at 40°C for 30 min before the bicinchoninic acid (BCA) assay (71) was used to detect the possible formation of new reducing ends.

Isothermal titration calorimetry.

Isothermal titration calorimetry was performed using a MicroCal VP-ITC calorimeter essentially as previously described (64, 65, 72). Purified His₆-Fn3-X181-sfGFP was dialyzed against 20 mM HEPES buffer, pH 7.0, containing 100  mM NaCl and 2 mM CaCl₂. Tamarind galactoxyloglucan (2.5 mg ml⁻¹) and the corresponding xyloglucan oligosaccharides (10 mM) were dissolved independently in the same buffer. The recombinant protein (35 to 40 µM) was placed in the sample cell, while the substrate was loaded in the injecting syringe. After the sample cell temperature equilibrated to 25°C, a first injection of 2 µl of the ligand, followed by 25 subsequent injections of 10 µl each, was performed while stirring at 280 rpm. The resulting heat of the reaction was recorded, and the data were analyzed using the Origin software program.