HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 52, 50245-50254, December 27, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/52/50245    most recent M204433200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by McLean, B. W. Articles by Haynes, C. A. Search for Related Content PubMed PubMed Citation Articles by McLean, B. W. Articles by Haynes, C. A. Social Bookmarking                      What's this?

Carbohydrate-binding Modules Recognize Fine Substructures of Cellulose^*

Bradley W. McLean^§¶, Alisdair B. Boraston^§¶**, Darren Brouwer, Nooshafarin Sanaie^§§, Colin A. Fyfe, R. Antony J. Warren^§, Douglas G. Kilburn^§¶, and Charles A. Haynes^¶§§¶¶

From the  Protein Engineering Network of Centres of Excellence, PENCE Inc., National Business Centre, Edmonton, Alberta T6G 2S2, Canada, ^§ Department of Microbiology and Immunology, ^¶ The Biotechnology Laboratory, the  Department of Chemistry, and the ^§§ Department of Chemical and Biological Engineering, University of British Columbia, Vancouver, British Columbia V6T 1Z4, Canada

Received for publication, May 7, 2002, and in revised form, August 19, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Competition isotherms are used to identify the set of cellulose substructures to which cellulose binding modules (CBMs) from families 2a, 3, 4, 9, and 17 bind. The experiments are based on coupling a unique fluorescent tag to each CBM in a manner that does not alter the natural binding properties of the CBM and therefore allows the surface and solution concentrations of each CBM to be monitored as a function of time and composition. Adsorption and surface exchange of like or competing CBMs are monitored using a range of cellulose preparations varying in both crystallinity and provenance. CBMs from families 2a, 3, 4, 9, and 17 are shown to recognize different physical forms of prepared cellulose. The demonstration of the very fine binding specificity of cellulose-specific CBMs implies that the polysaccharide targets of CBMs extend down to the resolution of cellulose microstructures.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Glycoside hydrolases are modular enzymes, containing one or more catalytic modules and often at least one carbohydrate binding module (CBM).¹ The basic function of CBMs is thought to be to localize the enzyme on the substrate to enhance the efficiency of degradation (1, 2). Nature has designed a number of different CBMs for this purpose, and structures and at least partial binding specificities have been reported for several CBMs in their isolated states (3-7). There are currently 29 families of CBMs, 13 of which contain members that bind cellulose. CBMs from different families are known to have different properties. CBMs within a family, although probably having similar folds, can also have quite different binding specificities (6, 8, 9).

Ascribing a more precise role to members of each CBM family that binds cellulose remains difficult due to incomplete characterization of their biological functions, our limited understanding of binding mechanisms, and our incomplete knowledge of the natural sorbent. In plant tissues, CBMs are thought to bind to regions within the complex cellulose/hemicellulose architecture. Although the exact structure of cellulose in its natural states is unknown, biophysical studies of prepared cellulose samples have identified three prevalent structural classes (10): crystalline cellulose, amorphous cellulose, and para-crystalline cellulose. If present, fine differences in structure within these three classes have not been identified. Hemicellulose is defined by an amorphous architecture that can contain mannans, xylans, galactans, and glucans of mixed linkages. These backbones can be highly decorated with sugar side chains.

In addition to increasing the local concentration of the enzyme on the substrate, CBMs may provide other functions. For example, CBMs from family 2a, which bind crystalline cellulose preparations with micromolar dissociation constants, are thought to disrupt the crystalline architecture and thereby increase substrate accessibility (11, 12). There is also evidence that families of CBMs that specifically bind cellulose could be involved in targeting enzymes to distinct regions of this complex substrate. In contrast to CBMs from family 2a, the family 4 CBM, CfCel9B-CBM4-1 from endoglucanase 9B of Cellulomonas fimi, binds amorphous cellulose and water-soluble cello-oligosaccharides but shows no affinity for crystalline cellulose. However, this type of comparative analysis has not yet been applied to all families of CBMs known to bind cellulose.

This paper reports the recognition of different physical forms of cellulose by particular CBMs from families 2a, 3, 4, 9, and 17. Competition isotherms are used 1) to identify the set of cellulose substructures to which each of these CBMs binds, 2) to evaluate possible correlations between CBM binding specificity and catalytic activity of the cognate enzyme(s), and 3) to identify combinations of CBMs that may result in enzyme competition(s) for the same sorbent binding sites. The competition isotherm experiments we report involve coupling a unique fluorescent tag to each CBM in a manner that does not alter the natural binding properties of the CBM. This allowed the surface and solution concentrations of each CBM to be monitored as a function of time and composition. The technique therefore provides a method for observing surface exchange of like or competing CBMs. These studies were performed using a range of cellulose preparations, varying in both provenance and crystallinity. For each preparation, cross-polarization magic angle-spinning (CP/MAS) solid state NMR was used to determine the relative degree of crystallinity and the presence and relative distribution of other cellulose substructures. The reproducibility and relative simplicity of these model-substrate preparations allowed for straightforward interpretation of our competition isotherm data.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Avicel PH-101 was obtained from FMC International (County Cork, Ireland). Bacterial microcrystalline cellulose (BMCC) and phosphoric acid-swollen cellulose (PASC, from Avicel) were prepared as described previously (13-15). Oregon Green 514 carboxylic acid succinimidyl ester (OG) and 6-([7-amino-4-methylcoumarin-3-acetyl]amino)hexanoic acid (AX) were obtained from Molecular Probes, Inc. (Eugene, OR). CBMs were produced in Escherichia coli from appropriate gene fragments and purified as described previously (4-7, 16, 17). The family to which a CBM belongs is indicated by a number, and the subfamily is shown by a letter (e.g. 2a). If a protein contains multiple CBMs from the same family, they are numbered consecutively from the N terminus of the protein (e.g. 4-1). The CBMs used were CfXyn10A-CBM2a from xylanase 10A of C. fimi; CfCel6A-CBM2a from endoglucanase 6A of C. fimi; CcCbpA-CBM3 from cellulose-binding protein A of Clostridium cellulovorans; CfCel9B-CBM4-1 from endoglucanase 9B of C. fimi; CcCel5A-CBM17 from endoglucanase 5A of C. cellulovorans; and TmXyn10A-CBM9-2 from xylanase 10A of Thermotoga maritima. For brevity, these CBMs will hereafter be referred to as CBM2a, Cel6A-CBM2a, CBM3, CBM4-1, CBM17, and CBM9-2, respectively. Their molar extinction coefficients (M¹), calculated from the aromatic acid contents (18), were 27,625 (CBM2a), 36,105 (Cel6A-CBM2a), 24,300 (CBM3), 21,370 (CBM4-1), 31,310 (CBM17), and 43,430 (CBM9-2).

CBMs were labeled with OG or AX (absorbance maxima 506 and 352 nm; emission maxima 537 and 447 nm, respectively) according to the supplier's instructions. OG and AX react with primary amines. The number of primary amines in the CBMs ranged from one in CBM4-1 to 16 in CBM9-2. The conditions for labeling were adjusted so that each CBM, except for CBM3, was labeled to a similar extent. This ranged from 0.4 mol of OG/mol of CBM2a to 1.2 mol of AX/mol of CBM2a. The poor solubility of CBM3 precluded labeling. Except for CBM9-2, the binding of the CBMs to cellulose was unaffected by the conjugated dyes. For each CBM except CBM9-2, the binding isotherm for the AX- or OG-labeled CBM measured by the depletion method and 280-nm absorbance readings was indistinguishable from that measured for the respective unlabeled CBM, indicating that the tag had no measurable effect on binding properties. Details concerning measurement of isotherms by the depletion method are provided in Ref. 16. CBM9-2-AX did not bind to cellulose and therefore was not used in our studies. The extent of labeling was calculated from the extinction coefficients (86,000 M¹ for OG and 18,500 M¹ for AX). Preparations of labeled CBMs were diluted with the corresponding unlabeled CBM to give final concentrations appropriate for the sensitivity of the fluorimeter. The concentrations of the final solutions were determined by reference to a standard curve.

Binding Experiments-- The affinities of CBMs for insoluble cellulose were determined by regressing binding isotherm data to a modified Langmuir-type binding model as described previously (16). Competition between CBMs for binding to cellulose was determined by obtaining an adsorption isotherm for a CBM-OG in the presence of a constant excess of a CBM-AX or vice versa, using microcentrifuge tubes blocked with bovine serum albumin to minimize nonspecific adsorption of the CBMs to the tubes. Exchange between CBM2a-OG in solution and CBM2a-AX bound to cellulose was determined as follows. CBM2a-AX (final concentration ~20 µM) was incubated with and without 1 mg of BMCC (total volume 1 ml) for 3 h at 4 °C. After centrifugation at 13,000 rpm, 900 µl of the supernatant solution were removed from each tube. The amount of CBM2a-AX bound to the cellulose was calculated from the difference in fluorescence of the samples with and without cellulose and reference to a standard curve. Previously, we showed that CBM2a remains irreversibly bound to cellulose in the absence of protein in the supernatant phase (19). Each BMCC pellet was therefore washed three times with 900 µl of 50 mM potassium phosphate buffer (pH 7.0) by centrifugation. Cellulose pellets were resuspended in three different concentrations of CBM2a-OG (~2, ~4, and ~15 µM) and incubated for 3 h. After centrifugation, the concentrations of CBM2a-OG and CBM2a-AX in the supernatants were determined by fluorescence. The concentrations of CBM2a-AX and CBM2a-OG adsorbed to the cellulose were calculated from the differences in fluorescence between the starting and final solutions.

CP/MAS Solid State NMR-- Solid state NMR spectra were obtained with a Bruker DSX-400 spectrometer, operated at frequencies of 400.13 MHz for ¹H and 100.61 MHz for ¹³C and using a 4-mm triple-tuned probe with zirconia rotors and Kel-F caps. Chemical shifts were referenced to tetramethylsilane using adamantane as a secondary reference for ¹³C spectra. The ¹H  ¹³C CP Hartmann-Hahn match condition was achieved using adamantane. Unless specified otherwise, all spectra were collected at room temperature on wet samples (estimated to be 40-60% water content) at a spinning rate of 5 kHz. Cellulose spectra were the sum of 40,000 scans, collected using a standard CP pulse sequence: a 4.5-µs proton 90° pulse, a 1000-µs contact pulse, and a 2-s delay between repetitions. No line broadening or resolution enhancement was applied during processing of the data.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

¹³C-CP/MAS Spectra of Cellulose-- The relative crystallinities, determined by x-ray diffraction, of Valonia cellulose (from Valonia ventricosa), BMCC, and Avicel are 1.0, 0.76, and 0.50, respectively (13, 20). The crystallinity of PASC is unknown, but it is expected to be less than that of Avicel because of the exposure to acid during its preparation. ¹³C CP/MAS NMR spectroscopy confirmed the relative crystallinities of BMCC and Avicel, and allowed an estimation of that of PASC.

The features of the signals at 81-93 ppm arising from C-4 of the anhydroglucose monomers in cellulose are indicative of cellulose structure (10, 21, 22). Highly ordered C-4 atoms located in the internal crystalline regions of the microfibril give rise to a sharp downfield signal at ~89 ppm; C-4 atoms that are at the cellulose surface, or within significantly less ordered, amorphous structures give rise to the broader upfield signal at 84 ppm (22). Based on the relative areas of the signals at 89 and 84 ppm, a large proportion of the glucose monomers in BMCC are internal and highly ordered, indicative of significant crystallinity (Fig. 1A). The proportion of ordered monomers is smaller in Avicel (Fig. 1B) and very much smaller in PASC (Fig. 1C), in accordance with their relative crystallinities.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   13C-CP/MAS spectra of insoluble cellulose preparations: BMCC (A), Avicel (B), and PASC (C). The horizontal bars above the spectrum of Avicel indicate the spectral ranges of the corresponding carbon atoms of the glucose monomer unit of cellulose. The spectral ranges indicated also apply to the BMCC and PASC spectra.

Binding of Individual CBMs to Cellulose-- Measured equilibrium constants and Langmuir saturation constants for binding of individual CBMs to BMCC and to PASC are shown in Table I. All of the CBMs bound to PASC, but only CBM2a, Cel6A-CBM2a, and CBM3 bound to BMCC. The measured affinities were consistent with previously reported values (14, 16, 23). For CBMs binding only to PASC, binding capacities ranged from ca. 6 to 24 µmol·g¹. However, the capacities of BMCC and PASC for CBM2a, Cel6A-CBM2a, and CBM3 were consistent and similar (Table I).

A Simple Langmuir-type Model for Binding Competition-- Consider two binding modules, CBM-X and CBM-Y, that have affinities K_aX and K_aY, respectively, for the same sites on a surface for which the total number of binding sites is [N]_o. A simple Langmuir-type theory can then be defined that shows that the concentration of CBM-X bound to the sorbent surface [B]_X will depend on [B]_Y, the concentration of bound CBM-Y, according to the following relation,

(Eq. 1)

where [F]_X is the equilibrium concentration of CBM-X in the solution phase. Equation 1 differs from the classic Langmuir equation for adsorption of a single solute from solution by the term (1  [B]_Y/[N]_o), which serves to reduce the initial slope of the isotherm for CBM-X with increasing total concentration of CBM-Y present in the system. In the presence of CBM-Y, CBM-X will therefore appear to bind with a lower affinity. Under binding conditions where the total concentration of CBM-X is vanishingly small, Equation 1 reduces to the following,

(Eq. 2)

where

(Eq. 3)

where ([B]_Y)* is the concentration of CBM-Y, in the absence of CBM-X, bound to sites shared by both sorbates. Equation 3 defines a useful assay for determining whether two CBMs compete for the same binding sites on the sorbent surface. In a competition-binding experiment, increasing concentrations of CBM-X are equilibrated with cellulose in the presence of a fixed total concentration of CBM-Y ([CBM-Y]_o), and the bound concentrations, [B]_X and [B]_Y, of the two CBMs are measured. If some or all of the binding sites for CBM-X are shared by CBM-Y, values of K_aX(app) regressed from the initial slope of the binding isotherm for CBM-X will be less than K_aX. However, if CBM-X and CBM-Y exclusively bind to different sites on the surface, then ([B]_Y)* is 0 at all values of [CBM-Y]_o, and K_aX(app) will everywhere be equal to K_aX. The ratio K_aX(app)/K_aX at saturating concentrations of CBM-Y should therefore provide a sensitive measure of binding-site competition between CBM-X and CBM-Y.

Equation 1 also provides a means for determining the extent of binding-site overlap. For a system in which CBM-X and CBM-Y compete for all available binding sites on the sorbent surface, Fig. 2 shows model predictions relating the surface concentration of CBM-Y, _Y (= [B]_Y/[N]_o), to the surface concentration of CBM-X, _x (= [B]_X/[N]_o), at various initial surface loadings of CBM-Y (i.e. different ([B]_Y)* values. At surface saturation, _X and _Y are linearly related with a slope of 1. Curvature in the plot is observed when ([B]_Y)* is not sufficient to completely saturate the sorbent surface.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Competition binding model predictions (Equation 1) for different initial surface loadings ([B]Y)* of CBM-Y. Fractional initial surface loadings are indicated on each curve. Results are independent of the values of KaX, KaY, and [N]o.

In contrast, a plot of _X versus _Y is predicted to have 0 slope if CBM-X and CBM-Y exclusively bind to different sites, since ([B]_Y)* is equal to 0 under these conditions. Slopes between 0 and 1 will be observed when CBM-X shares only a fraction of its total available binding sites with CBM-Y, thereby providing a measure of the extent of binding site overlap.

The model was tested using CBM2a labeled with two different fluorescent probes, AX and OG. Increasing concentrations of CBM2a-OG were equilibrated with BMCC in the presence of a fixed concentration (11 or 15 µM) of CBM2a-AX. Fig. 3A shows that the presence of 11 µM CBM2a-AX greatly reduces the initial slope of the binding isotherm for CBM2a-OG. K_{aCBM2a-OG(app)}/K_aCBM2a-OG values of 0.24 ± 0.05 and 0.07 ± 0.03 were regressed from initial slope data for [CBM2a-AX]_o values of 11 and 15 µM (data not shown), respectively, indicating significant binding site overlap between CBM2a-OG and CBM2a-AX. As expected, an increase in [B]_CBM2a-OG results in a concomitant decrease in [B]_CBM2a-AX (Fig. 3A). The presence of free CBM2a-OX in the absence bound protein is indicative of the presence of a small factors that contribute to the absorbance reading not specific to CBM2a. As a result, all isotherm data were regressed using a modified form of the Langmuir isotherm model that corrects for these nonspecific effects. This effect and the regression process are described in detail by McLean et al. (16).

View larger version (12K): [in this window] [in a new window]   Fig. 3.   CBM2a binding competition results. A, adsorption isotherm for CBM2a-OG binding to PASC at 4 °C in the absence (filled circles) and presence (unfilled circles) of 11 µM CBM2a-AX (~85% of saturation); the fractional surface coverage of CBM2a-AX as a function of free CBM2a-OG (filled triangles) is also plotted. B, fractional surface coverage of CBM2a-AX ([B]CBM2a-AX/[N]o = CBM2a-AX) as a function of the fractional surface coverage of CBM2a-OG ([B]CBM2a-OG/[N]o = CBM2a-OG).

The fraction of bound CBM2a-AX (_CBM2a-AX) approaches 0 (Fig. 3B) as the surface concentration of CBM2a-OG approaches saturation (_CBM2a-OG  1). The regressed slopes for the linear regions of the two [B]_CBM2a-AX/[N]_o versus [B]_CBM2a-OG/[N]_o plots were 1.0 ± 0.1 (11 µM 2a-AX) and 0.8 ± 0.1 (15 µM CBM2a-AX), indicating competition between CBM2a-AX and CBM2a-OG at all available binding sites.

Surface Exchange Experiments-- Jervis et al. (19) and others (24, 25) have reported that CBM2a binds irreversibly to cellulose. Their conclusions are based on the observation that after long equilibration times the removal of CBM2a from the solution phase does not result in appreciable desorption of bound CBM2a. However, no studies have been reported that test the ability of bound CBM2a to exchange with free CBM2a when it is present in solution. The differentially labeled preparations of CBM2a were used to test the exchange of bound with unbound CBM2a. Equilibration of a total concentration of 18.2 µM CBM2a-AX with BMCC at 4 °C gave concentrations of free and bound CBM2a-AX of 4.2 µM and 14.1 µmol·g¹ (the saturation value for CBM2a-AX on BMCC), respectively. Washing the sample three times with buffer removed less than 0.1% (i.e. <1 nmol·g¹) of the bound CBM2a-AX, in accordance with the previous reports of irreversible binding. However, when exposed to different concentrations (1.4, 3.2, and 14.0 µM) of free CBM2a-OG, bound CBM exchanged with the free CBM (Table II). The total concentration of CBM2a bound to BMCC (i.e. [B]_CBM2a-AX plus [B]_CBM2a-OG) remained constant at ~14 µmol·g¹, regardless of the concentration of CBM2a-OG added, but the proportion of CBM2a-OG in the bound material depended on the concentration added, an indication of free exchange. This was important for the binding competition experiments because it meant that the results would not be skewed by the "irreversibility" of binding of some CBMs, rather that they would reflect the relative affinities of the CBMs competing for the same binding sites.

View this table: [in this window] [in a new window]   Table II Exchange of CBM2a bound to BMCC at 4 °C in 50 mM potassium phosphate, pH 7.0  Note: [CBM2a]Total = [CBM2a-OG]o + [CBM2a-AX]o. Exchange based on equilibration of various loadings of CBM-2a-OG into the following system at 4 °C containing 1 mg of BMCC: [CBM2a-AX]o = 18.2 µM, [F]CBM2a-AX = 4.2 µM, and [B]CBM2a-AX = 14.1 µmol g1. A total concentration of 18.2 µM CBM2a-AX with BMCC at 4 °C to give concentrations of free and bound CBM2a-AX of 4.2 µM and 14.1 µmol · g1 (the saturation value for CBM2a-AX on BMCC), respectively.

CBM2a in Competition with CBM4-1, CBM17, and CBM9-2-- CBM4-1 shows no measurable binding affinity for BMCC but binds PASC strongly at 4 °C with a K_a of 0.25 (± 0.02) × 10⁶ M¹ (Table I). The binding of CBM2a to PASC is not significantly affected by the presence of CBM4-1 (Fig. 4A). Alone, CBM2a binds to PASC with an affinity of 1.2 ± 0.1 × 10⁶ M¹. When CBM4-1 is present at a total concentration of 19 µM (an amount that is ~95% of CBM4-1 saturation), the apparent K_a for the binding of CBM2a to PASC is reduced only slightly to 1.0 ± 0.1 × 10⁶ M¹. The amount of bound CBM4-1 decreased less than 10% as CBM2a approached saturation, indicating that the vast majority of bound CBM4-1 is unaffected by the presence of CBM2a. This conclusion is further supported by the near 0 slope (m = 0.07) of the _CBM4-1 versus _CBM2a plot (Fig. 4B).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Binding competition results for CBM2a and CBM4-1. A, adsorption isotherm for CBM2a-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 14.8 µM CBM4-1-AX (~80% of saturation); the fractional surface coverage of CBM4-1-AX as a function of free CBM2a-OG filled triangles is also plotted. B, fractional surface coverage of CBM4-1-AX as a function of the fractional surface coverage of CBM2a-OG. C, adsorption isotherm for CBM4-1-AX binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 18 µM CBM2a (~65% of saturation); the fractional surface coverage of CBM2a-OG as a function of free CBM4-1-AX filled triangles is also plotted. D, fractional surface coverage of CBM2a-OG as a function of the fractional surface coverage of CBM4-1-AX.

The reverse experiment (i.e. the influence of CBM2a on the binding affinity of CBM4-1) was also performed to verify data consistency. Fig. 4, C and D, shows that the binding of CBM4 -1 to PASC was not perturbed by the presence of CBM2a. K_aCBM4-1(app)/K_aCBM4-1 = 1 and the plot of _CBM2a versus _CBM4-1 shows 0 slope when the system contains CBM2a at a total concentration of 18 µM. Similarly, the binding of CBM17 and CBM9-2 are not perturbed by the presence of CBM2a. The affinity of CBM17 decreased a statistically insignificant amount from 0.76 ± 0.08 × 10⁵ to 0.67 ± 0.1 × 10⁶ M¹ in the presence of 18.8 µM CBM2a (Fig. 5). At high concentrations of CBM17 (near saturation), only about 10% of the bound CBM2a was displaced.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Binding competition results for CBM2a and CBM17. A, adsorption isotherm for CBM17-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 18.8 µM CBM2a-AX (~85% of saturation); the fractional surface coverage of CBM2a-AX as a function of free CBM17-OG filled triangles is also plotted. B, fractional surface coverage of CBM2a-AX as a function of the fractional surface coverage of CBM17-OG.

CBM9-2 binds specifically to the reducing ends of sugars (4, 6); its binding site accommodates only two sugar rings, such as cellobiose. Alone, it binds to PASC with an affinity of 0.63 ± 0.04 × 10⁶ M¹ at a sorbant capacity of 8.4 ± 0.1 µmol·g¹ (Table I). Fig. 6A shows that the presence of 8 µM CBM9-2 does not affect the binding affinity of CBM2a; in the presence of near saturating levels of CBM9-2, the apparent affinity of CBM2a-OG decreased slightly from 2.3 × 10⁶ to 1.7 × 10⁶ M¹ (Fig. 6B).

View larger version (11K): [in this window] [in a new window]   Fig. 6.   Binding competition results for CBM2a and CBM9-2. A, adsorption isotherm for CBM2a-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 8 µM CBM9-2. B, adsorption isotherm for CBM2a-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 18.8 µM CBM9-2. These two competition experiments were performed using two independent preparations of CBM2a-OG, resulting in slight differences in affinity and capacity.

CBM2a in Competition with CBM3 and Cel6A-CBM2a-- CBM3, the binding module from C. cellulovorans cellulose-binding protein A, binds to insoluble cellulose (26, 27). The affinity of CBM3 for PASC and BMCC is very similar to that of CBM2a (Table I). With CBM3 present as a competitor (not labeled) at a total concentration of 15 µM (~72% of saturation), the apparent affinity of CBM2a for PASC is significantly decreased to 0.17 ± 0.02 × 10⁶ M¹, approximately one-third of the affinity of CBM2a binding to PASC alone. This indicates that there are sites shared by both CBM3 and CBM2a. In this experiment, the affinity of the OG-labeled CBM2a was slightly lower than that measured in other experiments. This is probably due to incomplete separation of free Oregon Green from the CBM2a-OG conjugates when preparing the labeled CBM2a-OG. However, the free fluorescent probe does not contribute to the reduction in apparent affinity of CBM2a observed in the presence of CBM3.

CBM2a and Cel6A-CBM2a share 52% amino acid sequence identity with 67% amino acid sequence similarity, including three surface tryptophan residues implicated in the binding of family 2a CBMs to insoluble cellulose (14, 23, 28, 29). These two family 2a binding modules bind similarly to the insoluble cellulose preparations PASC and BMCC (Table I) with an affinity for PASC of ~1 × 10⁶ M¹ and an affinity for BMCC of ~3 × 10⁶ M¹. In competition experiments, K_a(app)/K_a = 0.33 for binding of CBM2a-OG to PASC in the presence of 15.8 µM Cel6A-CBM2a-AX (Fig. 7, A and B); similarly, K_a(app)/K_a = 0.28 for binding of CBM2a-OG to BMCC in the presence of 12.5 µM Cel6A-CBM2a-AX (Fig. 7C). This indicates that the majority of Cel6A-CBM2a bound to the cellulose surfaces directly competes with CBM2a. As bound CBM2a approaches saturation, the fraction of bound Cel6A-CBM2a is decreased to ~10% (Fig. 7D).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Binding competition results for CBM2a and Cel6A-CBM2a. A, adsorption isotherm for CBM2a-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 15.8 µM Cel6A-CBM2a-AX; the fractional surface coverage of Cel6A-CBM2a-AX as a function of free CBM2a-OG filled triangles is also plotted. B, fractional surface coverage of Cel6A-CBM2a-AX as a function of the fractional surface coverage of CBM2a-OG. C, adsorption isotherm for CBM2a-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 12.5 µM Cel6A-CBM2a-AX; the fractional surface coverage of Cel6A-CBM2a-AX as a function of free CBM2a-OG filled triangles is also plotted. D, fractional surface coverage of Cel6A-CBM2a-AX as a function of the fractional surface coverage of CBM2a-OG.

CBM4-1 in Competition with CBM17-- CBM4-1 and CBM17 have similar substrate specificities. Both bind amorphous cellulose, as well as cello-oligosaccharides greater than 5 or 6 glucose units in length (7, 17, 30, 31). CBM17, however, binds to PASC with ~4-fold higher affinity and capacity than does CBM4-1 (Table I).

Fig. 8A shows that the initial slope of the binding isotherm for CBM17-OG to PASC in the presence of 15 µM CBM4-1-AX is approximately half of that observed when no CBM4-1 was present, giving a K_a(app)/K_a value of 0.38 (Table III). The corresponding _CBM4-1 versus _CBM17 plot (Fig. 8B) shows that approximately one-third of the sites bound by CBM4-1 are shared by CBM17. In the complementary experiment, the apparent affinity of CBM4-1-AX was also reduced in the presence of a constant amount of CBM17-OG (Fig. 8C, Table III). However, in this case, only a small fraction of the CBM17-OG molecules bound were displaced as the CBM4-1 approached saturation; a majority of the sites that are bound by CBM17-OG are not recognized by CBM4-1-AX. This difference arises from the fact that the capacity of PASC for CBM17 is 4 times that for CBM4-1, so that a population of shared sites that represents one-third of the total number of binding sites for CBM4-1 represents only one-twelfth of the total binding sites for CBM17. Nevertheless, our results indicate that the amorphous regions of PASC are heterogeneous, composed of sites recognized by only CBM4-1, CBM17, or both CBMs.

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Binding competition results for CBM17 and CBM4-1. A, adsorption isotherm for CBM17-OG binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 15 µM CBM4-1-AX; the fractional surface coverage of CBM4-1-AX as a function of free CBM17-OG filled triangles is also plotted. B, fractional surface coverage of CBM4-1-AX as a function of the fractional surface coverage of CBM17-OG. C, adsorption isotherm for CBM4-1-AX binding to PASC at 4 °C in the absence (filled circles) and presence (open circles) of 15 µM CBM17-OG; the fractional surface coverage of CBM17-OG as a function of free CBM4-1-AX filled triangles is also plotted. D, fractional surface coverage of CBM17-OG as a function of the fractional surface coverage of CBM4-1-AX.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Binding Reversibility-- Our binding isotherm data for CBM2a on BMCC agree with previous studies (14, 16, 23) indicating that the surface concentration of CBM2a does not decrease following its dilution or removal from the solution phase. Thus, CBM2a binding to BMCC appears irreversible in a strict thermodynamic sense; dilution of free CBM2a does not lead to a concomitant reduction in bound CBM2a necessary for equilibrium to be reestablished. This suggests that there is a kinetic barrier to desorption under solution conditions where the process is thermodynamically favored. Kinetic barriers to desorption are not uncommon, particularly when the sorbate is a macromolecule and therefore adsorbs by making multiple energetically favorable contacts with the sorbent surface (32). Ion exchange and reversed phase chromatography resins are important examples of the utility of this phenomenon.

Gilkes et al. (14) have shown that removal of CBM2a from xylanase 10A of C. fimi reduces the hydrolytic activity of the enzyme on BMCC, indicating that the presence of the CBM enhances enzyme function. It is, however, difficult to rationalize this enhancement in activity given the irreversible nature of the CBM2a binding reaction. Enzyme activity is unlikely to be enhanced by the addition of a binding module that irreversibly anchors the enzyme at a specific position on the insoluble substrate surface, since proximal cleavage sites accessible to the enzyme's active site will be quickly exhausted, effectively rendering the bound enzyme inactive. A partial resolution to this apparent paradox was provided by Jervis et al. (19), who used fluorescence recovery after photobleaching experiments to show that CBM2a bound irreversibly to crystalline cellulose (sheets of V. ventricosa) diffuses on the cellulose surface with a surface diffusivity of 4.1 ± 0.5 × 10¹¹ m² s¹ at 20 °C. Thus, binding of CBM2a includes a mechanism for bound enzyme macromolecules to sequentially access multiple sites on the sorbent surface through surface diffusion.

However, the binding mechanism for CBM2a must also include a means for removing bound enzyme that has become inactive due to proteolysis or other mechanisms, since surface build-up of catalytically inactive protein would also lead to a reduced enzyme activity. Our results indicate that following removal of CBM2a-AX from the solution phase, exposure of bound CBM2a-AX to free CBM2a-OG results in exchange of the bound and free forms of CBM2a. This explains how active enzyme can be replenished on (and inactive enzyme removed from) the sorbent surface. It also illustrates the complexity of the binding reaction for CBM2a on crystalline cellulose. Although the net desorption reaction,

(Eq. 4)

does not proceed under thermodynamically favorable conditions, the sorbate exchange reaction,

(Eq. 5)

can proceed, but it results in no net reduction in surface concentration. In Equations 4 and 5, P, S, and P' represent the protein molecule (P) initially adsorbed to the sorbent surface (S) and the protein molecule (P') initially free in solution. To our knowledge, no current theories for reversible or irreversible adsorption of protein predict surface exchange for systems where adsorption appears macroscopically (i.e. thermodynamically) irreversible, suggesting that we have much to learn about the complex nature of the CBM2a binding reaction.

Identification of New Cellulose Substructures-- To date, cellulose-binding CBMs have been grouped into two general classes based on their ability to bind crystalline cellulose. CBMs that bind crystalline cellulose, represented in this work by CBM2a, Cel6A-CBM2a, and CBM3, contain a planar binding site that includes two or more aromatic amino acids known through mutagenesis studies to be essential to binding (16, 23, 33). It is thought that these planar binding sites complement the ordered surfaces of crystalline cellulose (23, 33, 34). In contrast, CBMs that bind amorphous cellulose but cannot bind crystalline cellulose, including CBM17 and CBM4-1, contain binding pockets or grooves designed to accommodate single cellulose chains (7, 17, 30, 31, 35).

Our competition isotherm data indicate that although it can bind predominantly amorphous preparations of cellulose such as PASC, CBM2a does not share surface binding sites with CBMs that show specificity for cello-oligosaccharides or amorphous cellulose (e.g. CBM4-1 and CBM17); it also does not share binding sites with CBM9-2, which binds to the reducing ends of cellulose polymers. The small amount of competition observed between bound CBM2a and these other CBMs presumably occurs only at boundaries between different cellulose chain organizations (e.g. crystalline and amorphous microstructures or reducing ends) at the surface of the sorbent sample. Given the demonstrated of CBM2a for crystalline cellulose, it is possible that binding to PASC is restricted to small microcrystalline surfaces within in the sample. However, binding capacities, [N]_o, for CBM2a do not correlate with the measured bulk crystallinity of the cellulose samples. [N]_o values for CBM2a are 14.3, 2.6, and 11.7 µmol/g cellulose for binding to PASC, Avicel, and BMCC, respectively. Thus, binding capacity is highest for the cellulose sample (PASC) shown to have the lowest crystallinity, as demonstrated by the disappearance of the sharp downfield C-4 peak at ~89 ppm in the ¹³C-CP/MAS spectrum of PASC compared with either BMCC or Avicel (Fig. 1). Since bulk crystallinity (which is estimated by CP/MAS and measured by x-ray diffraction) does not provide a direct measure of crystalline surface suitable for CBM2a binding, it is possible that PASC contains a large number of high surface area/volume microcrystalline regions while exhibiting relatively low bulk crystallinity. This would produce sufficient surface area to allow high capacity binding of CBMs (i.e. CBM2a, Cel6A-CBM2a, and CBM3) that specifically bind crystalline cellulose. This theory, which our data support, leads to the conclusion that CBMs can be truly specific for crystalline or amorphous cellulose. However, our results cannot fully discount the possibility that CBM2a (as well as Cel6A-CBM2a and CBM3) binds to unique noncrystalline substructures of PASC that show no affinity for CBM4-1, CBM9-2, or CBM17.

The competition isotherm data reported in this study also indicate that the binding specificities of CBMs known to bind amorphous forms of cellulose are much finer than previously thought. The presence of unique binding sites for CBM17 and for CBM4-1 on PASC suggests that the structure of cellulose chains in amorphous regions of insoluble cellulose do not adopt random conformations. Rather, the cellulose chains adopt specific structures that are discriminated by CBM17 and CBM4-1. The nature of these structures is unclear. However, it is known that the carbohydrate-binding site of CBM4-1 is a cleft in which a 5-mer stretch of the cellulose chain binds edge-on.² In contrast, the binding site of CBM17 is more surface-exposed, and the cellulose chain is bound with the glucopyranoside ring faces of one side of the chain facing the bottom of the binding site (i.e. turned 90° about the long axis relative to the orientation of cellulose bound to CBM4-1) (5). These differing binding site architectures are probably responsible for the ability of the CBMs to distinguish fine structure in amorphous regions of cellulose. As a result, the unique binding characteristics of CBMs may provide a means of identifying fine substructures in cellulose samples.

Table IV provides a new classification system for cellulose based on CBM binding specificities. Two classes of crystalline cellulose are defined, along with three distinct classes of amorphous cellulose. Crystalline cellulose is defined as cellulose that binds CBM2a, Cel6A-CBM2a, and CBM3 but does not bind CBM4-1, CBM9-2, or CBM17. Creagh et al. (24) report results for binding of CBM2a on BMCC that indicate the existence of two classes of binding sites on the crystalline cellulose surface. The first, which we designate class I crystalline cellulose, binds CBM2a relatively strongly, with an affinity and capacity similar to that reported here (Table I). The second, class II crystalline cellulose, binds CBM2a more weakly. Similar results were reported by Jervis (19) for binding of CBM2a on V. ventricosa cellulose, indicating that the results are independent of the origin of the crystalline cellulose sample. While not directly identified through binding studies, the potential for additional classes of crystalline cellulose is suggested by Carrard et al. (36), who demonstrated that the cellulolytic activity of the catalytic module of Clostridium thermocellum Cel9A depended on whether it was fused to CBM1, CBM2a, or CBM3. The structural differences in the various classes of crystalline-cellulose binding sites are unclear. They could be related to the relative proportions of crystalline cellulose allomorphs, including cellulose I, I, and II.

View this table: [in this window] [in a new window]   Table IV Classification of surface substructures of cellulose based on binding specificities of CBMs Relative binding affinity of each CBM is indicated by +, no affinity by , and affinity that was not experimentally determined by ND.

Our results also indicate the existence of at least three classes of amorphous cellulose based on CBM binding specificities. Competition isotherm data indicate that CBM9-2 binds to unique sites within the amorphous cellulose architecture, consistent with its structure that shows binding is specific to the reducing ends of sugars (4, 6). Class I binding sites on amorphous cellulose are therefore defined as those that bind CBM9-2 but do not bind CBM2a, CBM4-1, or CBM17.

Only a small fraction of bound CBM17-OG is displaced as binding of CBM4-1 approaches saturation. Class II amorphous cellulose is therefore defined as that which binds CBM17 but does not bind CBM2a, CBM4-1, or CBM9-2. Similarly, class III amorphous cellulose binds CBM4-1 but does not bind CBM2a, CBM9-2, or CBM17. Class IV amorphous cellulose binds both CBM4-1 and CBM17 but none of the other CBMs tested. Competition isotherm studies using a broad set of CBMs covering all cellulose-binding families therefore appear to provide a useful method for classifying cellulose samples. This will be particularly valuable if the presence of certain classes of cellulose within a sample can be correlated with desirable paper and textile properties such as fiber adhesiveness and surface area. Further competition experiments using an expanded library of crystalline cellulose specific CBMs and cellulose with different proportions of cellulose allomorphs may reveal further details of the complex surface structure of cellulose.

Biological Implications-- Of the 28 CBM families, examples have been found that bind to celluloses, chitins, xylans, mannans, starches, and a variety of glucans with diverse linkages. More often than not, the specificity of the binding module mirrors the specificity of the catalytic module. In all of the exceptions, it can be argued that the polysaccharide targets of the CBM and catalytic module are in close association (e.g. cellulose and xylan in plant cell walls), and, thus, a similar purpose is served. The demonstration of the very fine binding specificity of cellulose-specific CBMs implies that this targeting function extends down to the resolution of cellulose microstructures. In one case, this was demonstrated to have significant effects on the hydrolysis of cellulose when using CBMs that target crystalline regions (36). This has not yet been considered for noncrystalline microstructures, such as those targeted by CBM17 and CBM4-1, but a similar effect is possible and should be explored. Two fundamental questions should be addressed based on the observations presented here. 1) Does the fine specificity of cellulose specific CBMs reflect the fine specificity of the catalytic modules? 2) Is the targeting important in ensuring complete substrate coverage and, thus, efficient degradation of cellulose by a cellulolytic system?

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Present address: Dept. of Pathology, University of British Columbia, BC Research Institute for Children's and Women's Health, 950 West 28th Ave., Vancouver, British Columbia V52 4H4, Canada.

^** Present address: Dept. of Chemistry and The York Structural Biology Laboratory, University of York, Heslington, York YO10 5DD, United Kingdom.

^¶¶ To whom correspondence should be addressed: Biotechnology Laboratory, 237-6174 University Blvd., University of British Columbia, Vancouver, British Columbia, Canada V6T 1Z3. Tel.: 604-822-5136; Fax: 604-822-2114; E-mail: israels@chml.ubc.ca.

Published, JBC Papers in Press, August 20, 2002, DOI 10.1074/jbc.M204433200

² A. B. Boraston, D. Nurizzo, V. Notenboom, V. Ducros, D. R. Rose, D. G. Kilburn, and G. J. Davies (2002) J. Mol. Biol. 319, 1143-1156.

	ABBREVIATIONS

The abbreviations used are: CBM, cellulose binding module; CP/MAS, cross-polarization magic angle-spinning; BMCC, bacterial microcrystalline cellulose; PASC, phosphoric acid-swollen cellulose; AX, 6-([7-amino-4-methylcoumarin-3-acetyl]amino)hexanoic acid; OG, Oregon Green 514 carboxylic acid succinimidyl ester.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles: