The Fibronectin Type 3-Like Repeat from the Clostridium thermocellum Cellobiohydrolase CbhA Promotes Hydrolysis of Cellulose by Modifying Its Surface

Abstract

Fibronectin type 3 homology domains (Fn3) as found in the cellobiohydrolase CbhA of Clostridium thermocellum are common among bacterial extracellular glycohydrolases. The function of these domains is not clear. CbhA is modular and composed of an N-terminal family IV carbohydrate-binding domain (CBDIV), an immunoglobulin-like domain, a family 9 glycosyl hydrolase catalytic domain (Gh9), two Fn3-like domains (Fn3_1,2), a family III carbohydrate-binding domain (CBDIII), and a dockerin domain. Efficiency of cellulose hydrolysis by truncated forms of CbhA increased in the following order: Gh9 (lowest efficiency), Gh9-Fn3_1,2 (more efficient), and Gh9-Fn3_1,2-CBDIII (greatest efficiency). Thermostability of the above constructs decreased in the following order: Gh9 (most stable), Gh9-Fn3_1,2, and then Gh9-Fn3_1,2-CBDIII (least stable). Mixing of Orpinomyces endoglucanase CelE with Fn3_1,2, or Fn3_1,2-CBDIII increased efficiency of hydrolysis of acid-swollen cellulose (ASC) and filter paper. Scanning electron microscopic studies of filter paper treated with Fn3_1,2, Fn3_1,2-CBDIII, or CBDIII showed that the surface of the cellulose fibers had been loosened up and crenellated by Fn3_1,2 and Fn3_1,2-CBDIII and to a lesser extent by CBDIII. X-ray diffraction analysis did not reveal changes in the crystallinity of the filter paper. CBDIII bound to ASC and filter paper with capacities of 2.45 and 0.73 μmoles g⁻¹ and relative affinities (K_r) of 1.12 and 2.13 liters g⁻¹, respectively. Fn3_1,2 bound weakly to both celluloses. Fn3_1,2-CBD bound to ASC and filter paper with capacities of 3.22 and 0.81 μmoles g⁻¹ and K_rs of 1.14 and 1.98 liters g⁻¹, respectively. Fn3_1,2 and CBDIII contained 2 and 1 mol of calcium per mol, respectively. The results suggest that Fn3_1,2 aids the hydrolysis of cellulose by modifying its surface. This effect is enhanced by the presence of CBDIII, which increases the concentration of Fn3_1,2 on the cellulose surface.

Many proteins have modular structures. The fibronectin type III domain (Fn3) is a common constituent of animal proteins. Its main functions are to mediate protein-protein interactions and to act as a linker “to get the required biological function in the right space” (5). Fn3 domains have been found in intracellular, extracellular, and membrane-spanning proteins (22). Bacterial Fn3 sequences have been found only in extracellular glycohydrolases. In contrast to carbohydrate-binding domains (CBDs) with their tendency to be combined with a particular family catalytic domain (33), the Fn3 modules are randomly distributed, often in multiple copies among bacterial glycohydrolases belonging to different families (polygalacturonidases, cellulases, chitinases, pullulanases, amylases, etc.). It has been postulated that these domains serve as long linkers mediating proper interaction between catalytic and substrate-binding modules (22). However, the G+C content of Fn3-encoding segments from gram-positive bacteria correspond in most cases with that of the host genome (22). This suggests that at least some of these Fn3-encoding fragments have been in place long enough to acquire the character of the host organism and to perform a biological role other than the simple linking of two domains.

The Clostridium thermocellum cellulosome, a multiprotein complex that is specialized in hydrolysis of insoluble polysaccharides, is composed of a cellulosome-integrating protein, CipA, and catalytic subunits displaying activities of endo- and exoglucanases, xylanases, acetyl-xylan esterases, feruloyl esterases, mannanases, chitinases, and lichenases (1, 3, 8, 20, 31). All cellulosomal components have modular structures. Catalytic subunits contain at least a catalytic domain and a dockerin domain. Incorporation of the subunits into the complex occurs via interactions between their dockerin domains and cohesins of CipA. CipA binds the whole complex to insoluble substrates and anchors it to the cell surface (2). The subunits often contain additional domains like second catalytic, Ig- and Fn3-homologous domains, and CBDs (1, 31). Elucidation of biological functions of these domains might provide information of the cellulosome′s structural integrity and high activity.

One of the cellulosomal cellobiohydrolases, CbhA, is a multidomain protein composed of an N-terminal family IV carbohydrate-binding domain (CBDIV), an immunoglobulin-like domain, a family 9 glycosyl hydrolases catalytic domain (Gh9), two Fn3-like repeats (Fn3_1,2), a family III carbohydrate-binding domain (CBDIII), and a dockerin domain (38). In the present work we report the ability of Fn3_1,2 to change the surface of cellulose and promote its hydrolysis by the catalytic domain.

MATERIALS AND METHODS

Bacterial strains, culture conditions, and plasmids.

C. thermocellum JW20 was used as a source of genomic DNA. The bacterium was grown anaerobically under a nitrogen atmosphere at 60°C in a prereduced medium with 1% (wt/vol) cellobiose as described earlier (14, 15). E. coli BL21(DE3)pLys (Stratagene Cloning Systems, La Jolla, Calif.) was used as the cloning host for the T7 RNA polymerase expression vector pET-21b(+) (Novagen, Madison, Wisc.). It was grown in Luria-Bertani medium supplemented with ampicillin (100 μg/ml).

Isolation of genomic DNA.

Isolation of genomic DNA from C. thermocellum was done by the method of Marmur (24) with the modifications reported earlier (16).

Primer design, PCR, and cloning.

Flanking primers containing restriction sites were designed according to the DNA sequence of cbhA (38) (Fig. 1 and Table 1) and synthesized with an Applied Biosystems DNA synthesizer. DNA fragments were amplified by PCR using the primers in combination and with purified genomic DNA from C. thermocellum as a template. PCRs were done on a 480 thermal cycler (Perkin-Elmer, Norwalk, Conn.). The reactions were carried out with Taq polymerase (New England Biolabs, Beverly, Mass.). The annealing temperature was 54°C, and the extension time depended on the length of the fragment. PCR products were separated by 1% agarose gel electrophoresis and extracted from the gel using the Geneclean II Kit (Bio 101, La Jolla, Calif.). The extracted DNA fragments were digested with restriction enzymes and ligated into the pET-21b(+) vector linearized with the same enzymes. The ligation products were used to transform BL21(DE3)pLys competent cells. Each construct was verified by both restriction analysis and DNA sequencing.

FIG. 1.

Domain structure of truncated forms of CbhA used in the present study. All variants were cloned into pET-21b. Arrows indicate the direction of PCR amplification. Numbers refer to primers from Table 1. DD, dockerin domain. linker sequences are designated by heavy black lines.

TABLE 1.

Oligonucleotides used for cloning and site-directed mutagenesis^a

No. and primer	Sequence	Restriction site	Locationb	Direction
1 Gh9F	5′-CTAGCTAGCTATATACTTCCGCAGCCTGAT-3′	NheI	1157-1177	Forward
2 Gh9R	5′-TTTTCCTTTTGCGGCCGCCTCGTCAAGATAAGCTGTAAC-3′	NotI	2860-2877	Reverse
3 Fn3F	5′-CTAGCTAGCACAGACAGTGAAACCGATAAG-3′	NheI	2887-2907	Forward
4 Fn3R	5′-TTTTCCTTTTGCGGCCGCTACAAATATTTTTACGGAATC-3′	NotI	3415-3435	Reverse
5 CBDIIIF	5′-CTAGCTAGCGATGTAAAAGTACAGTATTTG-3′	NheI	3448-3468	Forward
6 CBDIIIR	5′-TTTTCCTTTTGCGGCCGCACCGCCCGGCGGCGTTCCCCA-3′	NotI	3862-3882	Reverse
7 Fn32F	5′-CTAGCTAGCCCTACTGTAAAACTTACT-3′	NheI	3277-3315	Forward
8 D1107AF	5′-TTTGTTATAAGATATGCAGCAAACTCCTTCCATGATCAG-3′		3730-3786	Forward
9 D1107AR	5′-CTGATCATGGAAGGAGTTTGCTGCATATCTTATAACAAA-3′		3730-3786	Reverse
10 D1116AF	5′-TTCCATGATCAGTCGAACGCATATTCGTTCGATCCAACT-3′		3766-3804	Forward
11 D1116AR	5′-AGTTGGATCGAACGAATATGCGTTCGACTGATCATGGAA-3′		3766-3804	Reverse
12 D953AF	5′-AGGGTTGATTTCCTTGTTGCAGGTGAAGTAATCGGTTCA-3′		3277-3315	Forward
13 D953AR	5′-TGAACCGATTACTTCACCTGCGTCAACAAGGAAATCAAC-3′		3277-3315	Reverse

^aRestriction sites are underlined. Mutated nucleotides are given in bold.

^bAmino acid residues numbering is according to that of CbhA (38).

Site-directed mutagenesis.

pET-21b(+) containing DNA fragments encoding the Fn3_1,2 or CBDIII domains served as the template for PCRs. Amino acid residues of interest were individually changed to alanine using the oligonucleotide primers listed in Table 1. PCRs with mutagenesis primers were carried out using the QuickChange site-directed mutagenesis kit (Stratagene). PCR products were used to transform BL21(DE3)pLys competent cells. Plasmid DNA in each case was isolated and sequenced. Mutants that possessed the correct nucleotide changes were used for further study.

Sequence analysis.

The Genetic Computer Group (version 10; University of Wisconsin Biotechnology Center, Madison, Wisc.) on the VAX/VMX system on the BioScience Computing Resource at the University of Georgia was used to analyze sequence data. Multiple alignments were generated with the MEME, PILEUP, and BLASTP programs.

Protein purification.

All proteins that originated from CbhA were six-His-tagged at the C terminus. They were purified from BL21(DE3)pLys cultures (1 liter) harboring pET-21b(+) with the DNA fragment of interest. Harvest was 5 h after induction with 1 mM isopropyl-β-d-thiogalactopyranoside. All purification steps were done at 4°C, except for the fast-performance liquid chromatography, which was run at room temperature. After collection, the cells were washed with 20 mM Na-phosphate buffer, pH 7.5, containing 0.5 M NaCl, and disintegrated with a French press. Cell debris was removed by centrifugation. Clear supernatant was applied onto a Ni-nitrilotriacetic acid agarose (Qiagen Inc., Valencia, Calif.) column equilibrated with the 20 mM Na-phosphate, pH 7.5, containing 0.5 M NaCl. The column was washed with 20 mM Na-phosphate-0.5 M NaCl, pH 6.0. Proteins were eluted by a gradient of 0 to 0.5 M imidazole, pH 6.0. Fractions containing recombinant proteins were combined, concentrated by precipitation with ammonium sulfate, and dialyzed against 20 mM Tris-HCl buffer, pH 7.5, containing 0.1 M NaCl. Dialyzed proteins were further purified by gel exclusion chromatography on a TSK 3000SW column (TosoHaas, Montgomeryville, Pa.). Eluted samples were concentrated using Centricon 10 concentrators (Amicon, Inc., Beverly, Mass.) and stored at 4°C. All proteins were purified close to homogeneity as ascertained by sodium dodecyl sulfate gel electrophoresis. Endoglucanase CelE from the anaerobic fungus Orpinomyces PC-2 (6) was a gift from Aureozyme, Inc.

Substrates used.

Acid-swollen cellulose (ASC) was prepared by treatment of Avicel PH-105 (FMS Corp., Philadelphia, Pa.) with phosphoric acid (19). Whatman No. 3 filter paper and para-nitrophenol (PNP) cello-oligosaccharides were from Sigma (St. Louis, Mo.). Cello-oligosaccharides were from Seikagaku (Tokyo, Japan).

Enzyme assays.

The activities of CbhA, its truncated variants containing the catalytic domain, and CelE were assayed at 60 and 50°C, respectively, in a 50 mM sodium citrate buffer, pH 6.0. The activity was monitored by measuring the release of reducing sugars or PNP (14, 15). The reducing sugars were determined with dinitrosalicylic acid reagent (25). Activity was expressed as micromoles of product (cellobiose equivalent or PNP) released per minute per milligram of enzyme. Efficiency of cellulose hydrolysis was determined with ASC and filter paper as substrates. Incubations were in 1 ml of 50 mM sodium citrate, pH 6.0, containing 50 mg of cellulose with 1 μmol of protein for 10 h at 60°C in the case of CbhA variants and at 50°C in the case of CelE.

High-performance liquid chromatography.

Sugars released from cello-oligosaccharides, carboxymethyl cellulose, and cellulose were analyzed with a Hewlett-Packard 1100 series high-performance liquid chromatographer equipped with an autoinjector and a 1047 RI detector using a Bio-Rad Aminex PHX-42A carbohydrate column. Water was used as a mobile phase at a flow rate of 0.6 ml min⁻¹.

Thermostability assay.

Proteins were incubated in 50 mM Tris-HCl buffer, pH 7.5, at 60°C. At time intervals aliquots were taken and centrifuged, and the supernatant solution was assayed for the residual activity. Half-life times were calculated using the equation for a one-phase exponential decay in the GraphPad Prism program.

Cellulose-binding assay.

Adsorption assays were done at a room temperature in 1.5-ml microcentrifuge tubes. Proteins were mixed with cellulose (1 g/liter) in 50 mM sodium citrate buffer (pH 6.0) to a final volume of 0.5 ml. Tube contents were continuously mixed by rotation. After equilibration for 2 h, cellulose and bound protein were removed by centrifugation at 10,000 × g for 10 min. Centrifugation was repeated twice to ensure the removal of all of the cellulose. The amount of unbound protein was determined. The amount of bound protein was calculated from the difference between the initial amount of protein and that of the unbound protein. Data presented are from five replicates. The data were analyzed with the computer program GraphPad Prism using the one-side binding hyperbola equation that describes the binding of a ligand to a receptor that follows the law of mass action. The fit converged for all sets of data. The algorithm minimized the sum of the squares of the actual distances of the points from the curve. To determine the relative equilibrium association constant Kr, which can be used to compare the affinities of various related ligands for a given preparation of cellulose, we used the method of Gilkes et al. (9).

Metal content.

The metal content of the proteins was determined by plasma emission spectrophotometry (Jarrell-Ash 965 ICP). Each data point was a mean of three replicates.

Protein analysis.

During purification, protein concentrations were determined with the Coomassie protein assay reagent (Pierce, Rockford, Ill). In other experiments, protein concentrations were determined on the basis of A₂₈₀ values. Each sample was extensively dialyzed against buffer and then centrifuged at 20,000 × g for 60 min. The absorbance was read against dialysis buffer (used as a blank). Molar absorption coefficients (calculated from aromatic amino acid [tryptophan and tyrosine] content) were (M⁻¹ cm⁻¹) 161,340 for Gh9; 177,270 for Gh9-Fn3_1,2; 197,100 for Gh9-Fn3_1,2-CBDIII; 34,480 for Fn3_1,2-CBDIII; 14,650 for Fn3_1,2; 10,810 for Fn3₂; and 19,830 for CBDIII.

Preparation of calcium-free polypeptides.

Calcium in polypeptides was removed by 48 h of dialysis at 4°C against 20 mM EDTA, pH 8.0, followed by dialysis against three changes of calcium-free 20 mM sodium phosphate buffer, pH 6.0. To remove traces of multivalent cations, the buffer was stirred overnight with 5% (wt/vol) Chelex-100 (Bio-Rad). All glassware, plasticware, and quartz cells were soaked overnight in 4 M HCl and then thoroughly rinsed with deionized water.

CD spectra.

Circular dichroism (CD) measurements were carried out at 25°C on a Jasco J-710 spectropolarimeter with a quartz cell with a 1.0-mm path length. The cell temperature was controlled to within ±0.1°C by circulating water via a Neslab R-111 water bath through a cell jacket. The results were expressed as mean residue ellipticity (MRE), which is defined as MRE = (100 MRE_obs)/lc, where MRE_obs is the observed MRE in degrees, c is the residue concentration (moles per liter), and l is the light path length in centimeters. The spectra obtained were averages of five scans. The spectra were smoothed via an internal algorithm in the Jasco software package, J-710 for Windows. Secondary structure was estimated by utilizing the MRE value at 222 nm. Taking peptide length into account, percent Helix = 100[MRE]/39,500(1 − 2.57/n), where n is the number of residues. All protein samples were in 20 mM sodium phosphate buffer (pH 6.0) at a concentration of 10 μM.

Observation by electron microscopy.

Samples of filter paper (100 mg) were incubated alone or with 10 μmol of either Fn3_1,2, CBDIII, or Fn3_1,2-CBDIII in 0.5 ml of 25 mM sodium citrate buffer-5 mM calcium chloride (pH 6.0) containing 0.05% sodium azide for 20 h at room temperature and constant rotation. Then, the proteins were removed by treatment with proteinase K followed by an intensive wash with distilled water. The residual cellulose samples were dried under vacuum, secured onto scanning electron microscope (SEM) aluminium stubs using sticky tabs, and coated with chromium in an Edward′s Auto 306 vacuum evaporator for 10 s at 3 × 10⁻⁷ millitorr. The samples were viewed with a LEO 982 field-emission scanning electron microscope. The micrographs were taken at a constant voltage of 2.0 kV.

X-ray diffraction analysis.

Sections (4 cm² each) of filter paper were treated with 10 μmol of either Fn3_1,2, Fn3_1,2-CBDIII, or CBDIII under the same conditions as for SEM. The control sample was treated similarly but with no protein added. The samples were mounted onto glass slides for X-ray diffraction analysis. X-ray diffraction spectra were collected with CuK_α using an X-ray diffractometer (Philips Analytical Inc., Natick, Mass.) equipped with a focusing monochromator and operated at 35 KV and 20 mA. Samples were scanned from diffraction angles of 3 to 32^o 2θ at a rate of 2^o 2θ/min.

RESULTS

Analysis of hydrolysis products.

Gh9 from C. thermocellum CbhA and CelE from Orpinomyces sp. strain PC-2 hydrolyzed ASC, Avicel, and filter paper, producing cellobiose and cellohexaose, respectively, as main hydrolysis products. Attachment of Fn3_1,2 or Fn3_1,2-CBDIII to Gh9 as well as mixing of Fn3_1,2, CBDIII, or Fn3_1,2-CBDIII with either Gh9 or CelE did not affect the type of reducing sugars produced. To study the mode of action of Gh9, cello-oligosaccharides and PNP-cello-oligosaccharides were used as substrates, and the hydrolysis products were determined. Gh9 hydrolyzed substrates from the nonreducing end as it produced glucose and cellobiose from cellotriose, cellobiose from cellotetraose, glucose and cellobiose from cellopentaose, cellobiose from PNP-cellobioside, cellobiose and glucose from PNP-cellotrioside, cellobiose from PNP-cellotetraoside, and glucose and cellobiose from PNP-cellopentaoside. In our previous publication we reported that the catalytic domain of another C. thermocellum cellobiohydrolase, CelK, which displayed more than 90% identity to Gh9 of CbhA, also acted on nonreducing ends of cellulosic substrates (14).

Hydrolysis of ASC and filter paper.

The truncated forms of CbhA, Gh9, Gh9-Fn3_1,2, and Gh9-Fn3_1,2-CBDIII, differed in their hydrolytic activities. Gh9-Fn3_1,2 and Gh9-Fn3_1,2-CBDIII produced four and seven times more reducing sugars from acid-swollen cellulose and two and three times more reducing sugars from filter paper, respectively, than Gh9 alone (Table 2). Mixing of Gh9 with CBDIII did not affect the hydrolysis rate of the celluloses. Mixing of Gh9 with Fn3_1,2, or Fn3_1,2-CBDIII resulted in only a slight increase in reducing sugars production, showing that the positive effect of Fn3_1,2 on hydrolysis was higher when this domain was directly attached to the catalytic domain than when it was simply mixed with it. Mixing of Orpinomyces endoglucanase CelE with CBDIII, Fn3_1,2, or Fn3_1,2-CBDIII resulted in 21, 42, and 86% increases in production of reducing sugars from acid-swollen cellulose, and 15, 27, and 60% filter paper, respectively. No sugars were released upon treatment of filter paper with CBDIII, Fn3_1,2, or Fn3_1,2-CBDIII (Table 2).

TABLE 2.

Hydrolysis^a of ASC and filter paper by truncated forms of CbhA and combinations of Gh9 and CelE with CBDIII, Fn3_1,2, and Fn3_1,2-CBDIII

Protein(s) used	Amt of reducing sugars produced (μg of glucose/ml) with:
	ASC	Filter paper
Gh9	32.1 ± 1.25	10.4 ± 1.03
Gh9-Fn31,2	128.3 ± 6.90	19.7 ± 2.23
Gh9-Fn31,2-CBDIII	245.5 ± 11.51	26.7 ± 2.38
Gh9 + CBDIII	30.5 ± 2.89	9.1 ± 1.41
Gh9 + Fn31,2	34.1 ± 1.85	10.3 ± 1.56
Gh9 + Fn31,2-CBDIII	37.5 ± 3.67	11.3 ± 1.31
Fn31,2	0.2 ± 0.01	NDb
CBDIII	0.1 ± 0.01	ND
Fn31,2-CBDIII	0.2 ± 0.02	ND
CelE	15.4 ± 1.63	2.5 ± 0.20
CelE + CBDIII	18.6 ± 1.44	2.9 ± 0.12
CelE + Fn31,2	21.9 ± 1.55	3.2 ± 0.21
CelE + Fn31,2-CBDIII	28.6 ± 1.52	4.0 ± 0.25

^aFifty milligrams of acid-swollen cellulose or filter paper was incubated with 1 μmol of protein as indicated in 1 ml of 50 mM sodium-citrate buffer (pH 6.0) for 10 h at 60°C for Gh9 variants and at 50°C for CelE variants.

^bND, not detected.

Thermostability of proteins.

Half-life times measured by determination of catalytic activity of Gh9, Gh9-Fn3_1,2, and Gh9-Fn3_1,2-CBDIII at 60°C were 116, 97, and 88.5 h, respectively. The results show that specific interactions between the CbhA domains affect the final properties of these polypeptides. Previously we found that combination of an N-terminal CBDIV and a Gh9 of CelK was more stable than these domains alone (14, 15, 17). Since CbhA and CelK are highly homologous proteins (15), we suggest that the presence of CBDIV at the N terminus of CbhA also stabilizes this protein. In contrast, when Fn3_1,2 and CBDIII were attached to the C terminus of Gh9 of CbhA, the combined proteins had reduced stability in comparison to Gh9 alone. This underlines the importance of the location of domains in modular polypeptides.

Morphological changes of cellulose fibers (Fig. 2).

FIG. 2.

Scanning electron micrographs of cellulose fibers untreated (row A) and treated with CBDIII (row B), Fn3_1,2 (row C), and Fn3_1,2-CBDIII (row D) at different magnifications (columns 1 to 3). The control fibers have a smooth and even surface. Surfaces of the fibers treated with Fn3_1,2 or Fn3_1,2-CBDIII have been significantly eroded. CBDIII alone caused minor irregularities of cellulose surface.

Filter paper is about 45% crystalline cellulose. Changes in the surface morphology of native cellulose and cellulose treated with CBDIII, Fn3_1,2, and Fn3_1,2-CBDIII were investigated by scanning electron microscopy (Fig. 2). Surfaces of control fibers were relatively smooth and polished with few natural irregularities (Fig. 2A). Distinct surface changes were evident after treatment. Fibers treated with CBDIII looked slightly loosened up (Fig. 2B). Surfaces of fibers treated with Fn3_1,2 were exfoliated, showing underlying microfibrils (Fig. 2C). Fibers treated with Fn3_1,2-CBDIII had irregular, roughed, and crenellated surfaces (Fig. 2D). The results are in agreement with the earlier observations that a Cellulomonas fimi CBD erodes the surfaces of cotton fibers (7).

Characterization of filter paper by X-ray diffraction analysis.

The X-ray diffractograms of the Fn3_1,2-, Fn3_1,2-CBDIII-, and CBDIII-treated and control filter papers indicated the presence of cellulose I structure with diffraction maxima at approximately 14, 16, and 22.5^o 2θ (Fig. 3). Peak intensities for the control sample and three samples treated with the polypeptides were similar. The peaks at 22.5^o 2θ were relatively sharp for all samples, suggesting that treatment had minimal effect on crystallinity of the cellulose. Absence of any effect of CBDIII alone on cellulose was in line with SEM data. The lack of observable changes in diffraction intensity and crystallinity by treatment with Fn3_1,2 or Fn3_1,2-CBDIII and the significant morphological changes as revealed by SEM suggest that these polypeptides acted only at the surface of cellulose and that surface alterations were masked by diffraction from the larger mass of unaffected cellulose in the interior of the fibers.

FIG. 3.

X-ray diffractograms of filter paper samples treated with Fn3_1,2 (1), Fn3_1,2-CBDIII (2), or CBDIII (3) and untreated control sample (4). The diffraction intensities were similar for all treated samples and the control.

Calcium content.

One mole of CBDIII or Fn3₂ contained one mole of calcium, respectively. One mole of Fn3_1,2 repeat contained two moles of calcium, suggesting the presence of one calcium ion per each domain. Dialysis of CBDIII, Fn3_1,2, or Fn3₂ against 20 mM EGTA overnight resulted in almost total loss of calcium (Table 3). The calcium content of proteins could be restored by dialysis against 10 mM calcium chloride (data not shown).

TABLE 3.

Content of calcium in the recombinant polypeptides

Varianta	Amt (mol) of calcium per mol of protein
	Native protein	EGTA-treated protein
Fn31,2	1.94	0.03
Apo-Fn31,2	0.07
Fn32	1.03	0.02
Apo-Fn32	0.25
Fn31,2-CBDIII	2.87	0.02
CBDIII	0.94	0.03
Apo-CBDIII	0.05
D953AFn3	0.09
D1107ACBDIII	1.95
D1116ACBDIII	0.01

^aApo, protein without calcium.

Homology studies.

The amino acid sequence of CbhA Fn3₂ is 31% homologous with that of Fn3₁. A search of protein sequences with the databases SWISS-PROT, pir2, and Gen Pept New revealed that Fn3₂ has homologues with several modules found in glycohydrolases. It displayed sequence identities of 48 and 36%, respectively, to two modules of Vibrio furnissii chitinase (18) (accession no. T30199); 42 and 33%, respectively, to two modules of V. cholerae chitodextrinase (10) (accession no. D82428); 37, 34, 34, and 28%, respectively, to four modules of V. cholerae chitinase VC1952 (10) (accession no. C82135); 30 and 36%, respectively, to V. cholerae chitinase VC1073 (10) (accession no. D82246); 42% to Thermomonospora fusca exocellulase Cel48A (12) (accession no. T44496); 37% to Pseudoalteromonas sp. chitinase A (32) (accession no. T30933); 37% to Streptomyces coelicolor cellulase (Seeger et al., direct submission) (accession no. T35238); 42% to Deinococcus radiodurans hypothetical protein (37) (accession no. B75320); and 26% to Streptomyces plicatus chitinase X (28) (accession no. P11220) (Fig. 4). The fact that Fn3₂ is less homologous to Fn3₁ than to Fn3 domains from distant genera suggests that these domains have been acquired from different organisms due to horizontal gene transfer rather than having arisen by gene duplication.

FIG. 4.

Alignment of Fn3-like domains. Abbreviations (accession numbers are given in parentheses): Chi_Vibrfu, Vibrio furnissii chitinase (T30199); Chitd_Vibrcho, Vibrio cholerae chitodextrinase (D82428); Chi_Vibrcho, V. cholerae chitinase VC1952 (C82135); Chit_Vibrcho, V. cholerae chitinase VC1073 (D82246); CbhA_Clotm, C. thermocellum cellobiohydrolase CbhA (X80993); Cel_Thermfu, T. fusca exocellulase Cel148A (T44496); Chi_Pseudal, Pseudoalteromonas sp. strain Chitinase (T30933); Cel_Streptco, S. coelicolor cellulase (T35238); Hypot_Deinra, D. radiodurans hypothetical protein (B75320); Chit_Strepli, S. plicatus chitinase (P11220). Accession numbers are from SWISS-PROT, pir2, and Gen Pept New databases.

Alignment of amino acid sequences of 12 CBDIIIs and 7 Fn3 domains did not show any homology except for a short R-X₁-X₂-X₃-VDG motif. The triplet VDG was well conserved in both domains (Fig. 5). Noticeably, position X₁ was represented by tyrosine in all CBDIIIs and by valine in all Fn3s; X₂ was represented by aromatic residues (tyrosine or tryptophan) in CBDIIIs and by aspartic or glutamic amino acids in Fn3s; X₃ was not conserved. Asp₉₅₃ in the VDG motif of Fn3₂ corresponded to Asp₇₇ of CBDIIIa of the scaffoldin subunit CipC (CBD_CipC) of Clostridium cellulolyticum that has been shown to chelate calcium ions (30), suggesting that the short motif found in both CBDIIIs and Fn3s is important for the formation of a calcium-binding center. Alignment of CBDIII of CbhA with other CBDIIIs (data not shown) revealed two highly conserved aspartic residues. These residues in CBDIII of CbhA are Asp₁₁₀₇ and Asp₁₁₁₆. Asp₁₁₁₆ corresponded to Asp₁₅₈ of CBD_CipC that binds calcium (30).

FIG. 5.

Alignment of the R-X₁-X₂-X₃-VDG motif of Fn3-like domains and CBDIIIs. Abbreviation of CBDIIIs (accession numbers are given in parentheses): CelN_Erwca and CelVI_Erwca, Erwinia caratovora cellulases N (S54744) and VI (S54744); Xyl_Caldsp, Caldocellum sp. strain xylanase (T31085); Cel_Anaert, Anaerocellum thermophilum cellulase (T31337); ManCel_Calsa, Caldocellum saccharolyticum β-mannanase (A48954); CelZ_Closr and CelY_Closr, Clostridium stercorarium cellulases Z (S12021) and Y (Z69359); CipA_Clotm, C. thermocellum CipA (S33527); CipC_Cloce, C. cellulolyticum CipC (PC6006); CelI_Clotm, C. thermocellum CelI (A47704); CelA_Bacla, Bacllus lautus CelA (B41897); Cel_Bacsu, B. subtilis cellulase (G69593). Abbreviations for Fn3-like domains are the same as in Fig. 4.

Mutagenesis.

To elucidate the role of calcium, Asp₉₅₃ of Fn3₁,₂ and Asp₁₁₀₇ and Asp₁₁₁₆ of CBDIII were individually mutated to alanines, giving D953A_Fn3, D1107A_CBDIII, and D1116A_CBDIII. Mutant D953A_Fn3 formed inclusion bodies and contained only traces of calcium. The presence of calcium in Fn3_1,2 and its absence in mutant D953A_Fn3 suggests that an aspartic acid residue from the VDG motif is indeed involved in chelating calcium in the Fn3 domains and CBDIIIs. D1107A_CBDIII contained 1 mol of calcium, while D1116A_CBDIII contained only traces of the ion. Thus, of the two aspartic acid residues conserved in CBDIII, only Asp₁₁₁₆ is needed in calcium binding.

Binding to cellulose.

Fn3_1,2 weakly bound ASC with a capacity of 0.94 μmol/g and a K_r of 0.26 liters/g and filter paper with a capacity of 0.32 μmol/g and a K_r of 0.11 liters/g (Table 4). CBDIII bound to ASC with a capacity of 2.45 μmol/g and a K_r of 1.12 liters/g and filter paper with a capacity of 0.81 μmol/g and a K_r of 1.96 liters/g. Elimination of calcium did not affect binding of Fn3_1,2 or of CBDIII to cellulose.

TABLE 4.

Parameters for binding of Fn3_1,2, CBDIII, and Fn3_1,2-CBDIII to ASC and filter paper

Proteina	ASC		Filter paper
	PCb	Krc	PC	Kr
Fn31,2	0.94	0.26	0.32	0.11
Apo-Fn31,2	0.97	0.38	0.25	0.09
CBDIII	2.45	1.12	0.73	2.13
Apo-CBDIII	2.27	1.06	0.81	1.96
Fn31,2-CBDIII	3.22	1.14	0.81	1.98

^aApo, protein without calcium.

^bMaximum amount of protein bound to 1 g of cellulose (μmoles g⁻¹).

^cRelative equilibrium association constant (liter g⁻¹).

Circular dichroism spectra.

The far-UV CD spectra of Fn3_1,2, CBDIII, and Fn3_1,2-CBDIII contained a single strong negative band with a maximum at approximately 215 nm, suggesting the presence of almost exclusively β-sheets in the secondary structure of the domains (Fig. 6). Such structure is characteristic for all known CBDs and Fn3 domains. CD spectra of native and calcium-free CBDIII were identical, indicating an absence of structural changes upon elimination of the ion. Elimination of calcium from Fn3_1,2 resulted in some decrease of β-sheet content. Probably, the presence of calcium stabilizes folding of Fn3_1,2. It is possible that appearance of the D953A_Fn3 mutant in inclusion bodies was a result of calcium loss followed by exposure of the hydrophobic core of the domain (5, 23). Interaction between domains stabilized Fn3_1,2-CBDIII even in the absence of calcium, as CD spectra of holo and apo forms of Fn3_1,2-CBDIII were similar (Fig. 6).

FIG. 6.

CD spectra of native and calcium-free Fn3_1,2 (A), CBDIII (B), and Fn3_1,2-CBDIII (C). The proteins were in 20 mM sodium phosphate buffer (pH 7.0) at a concentration of 10 μM. Measurements were performed with a J710 spectrometer (Jasco).

DISCUSSION

Fn3₁ and Fn3₂ of CbhA display a low level of homology to human fibronectin. However, they have been identified as Fn3-like domains on the basis of hydrophobic cluster analysis showing an 80% score, secondary structure prediction, and enhanced content of valine and hydroxylated amino acids (12, 38). The role of Fn3 modules in bacteria has not been studied. The fact that Fn3 domains have been found in only one class of bacterial enzymes, glycosyl hydrolases, which act against insoluble substrates, suggests that these domains participate in substrate utilization. This suggestion is in line with the observation that absence of one or both Fn3-like domains from Bacillus circulans chitinase A1 separating catalytic and binding modules did not affect binding to chitin but altered hydrolytic activity of the enzyme on colloidal chitin (36). Several three-dimensional structures of Fn3 modules (21, 23) and module pairs (4, 11) from animal sources have been resolved. They have a common molecular topology: seven antiparallel β-strands are arranged in two sheets (A-B-E and C-C-F-G), enclosing a core of highly conserved hydrophobic residues (5, 12, 27). This fold is similar to the fold of ChiN domains found only at the N terminus of family 18 chitinases (12, 26) and to the fold of CBDs (30, 35). The crystal structure of Serratia marcescens chitinase A composed of a ChiN module and a catalytic domain showed that the ChiN domain is in an ideal orientation to guide the ”loose end“ of the chitin chain towards the catalytic groove where the terminal sugars can be cleaved (27). This observation suggests that the ChiN domain acts as so-called helping CBD (30), as has been observed for T. fusca endo/exocellulase E4. The crystal structure of a truncated variant of this enzyme composed of an N-terminal family 9 catalytic domain and family IIIc helping CBD with no ability to bind cellulose revealed close interaction between the two domains where the role of the CBDIIIc module was to direct substrate to catalytic site (29). There are some other common features in CBDs and Fn3 domains. It has been demonstrated that isolated CBD from C. fimi CenC disrupted the surfaces of cellulose fibers (7). This unhydrolytic activity is similar to the activity of Fn3_1,2 reported here.

The presence of one calcium-binding center is characteristic for many CBDs (13, 33, 34, 35) Each Fn3₁ and Fn3₂ also binds 1 mol of calcium. Alignment of the deduced amino acid sequences of CBDs and Fn3-like domains revealed no homology except for one short conserved motif. It included an aspartic acid residue chelating calcium in C. cellulolyticum scaffoldin CBDIIIa (30) and in Fn3_1,2.

The present study is the first demonstration of the effect of the Fn3_1,2 module on cellulose. Probably, the loosening up of the cellulose surface involves exfoliation, separation of cellulose chains, and exposure of additional sites of cellulose for hydrolysis by the catalytic domain. Fn3_1,2 affects hydrolysis more significantly when it is covalently attached to the catalytic domain rather than simply mixed with it. Similar to the ChiN domain (27) and CBDIIIc (29), Fn3_1,2 probably directs single cellulose chains into the catalytic center. CBDIII located at the C terminus of Fn3_1,2 mediates tighter interaction between Gh9-Fn3_1,2 and substrate, thus further enhancing surface modification and hydrolysis of cellulose. This idea is supported by the observation that the majority of Fn3-like domains in glycosyl hydrolases are specifically located between catalytic and carbohydrate-binding modules.