Directed Mutagenesis of Specific Active Site Residues onFibrobacter succinogenes1,3–1,4-β-d-Glucanase Significantly Affects Catalysis and Enzyme Structural Stability*

The functional and structural significance of amino acid residues Met39, Glu56, Asp58, Glu60, and Gly63 ofFibrobacter succinogenes1,3–1,4-β-d-glucanase was explored by the approach of site-directed mutagenesis, initial rate kinetics, fluorescence spectroscopy, and CD spectrometry. Glu56, Asp58, Glu60, and Gly63 residues are conserved among known primary sequences of the bacterial and fungal enzymes. Kinetic analyses revealed that 240-, 540-, 570-, and 880-fold decreases in k cat were observed for the E56D, E60D, D58N, and D58E mutant enzymes, respectively, with a similar substrate affinity relative to the wild type enzyme. In contrast, no detectable enzymatic activity was observed for the E56A, E56Q, D58A, E60A, and E60Q mutants. These results indicated that the carboxyl side chain at positions 56 and 60 is mandatory for enzyme catalysis. M39F, unlike the other mutants, exhibited a 5-fold increase inK m value. Lower thermostability was found with the G63A mutant when compared with wild type or other mutant forms ofF. succinogenes 1,3–1,4-β-d-glucanase. Denatured wild type and mutant enzymes were, however, recoverable as active enzymes when 8 m urea was employed as the denaturant. Structural modeling and kinetic studies suggest that Glu56, Asp58, and Glu60 residues apparently play important role(s) in the catalysis of F. succinogenes 1,3–1,4-β-d-glucanase.

With respect to the kinetic properties of bacterial 1,3-1,4-␤-D-glucanases, these enzymes reside with the retaining glycosidase activity leading to the hydrolysis of glycosidic bonds with a net retention of the anomeric configuration during ␤-glucan hydrolysis (9). The 1,3-1,4-␤-D-glucanase enzyme has been proposed to follow a general acid-base catalytic mechanism in which specific amino acid residues acting as a general acid or a nucleophile are required for completing a catalytic reaction of the enzyme (10). A number of methodologies have been previously applied to the characterization of catalytic mechanism(s) of Bacillus 1,3-1,4-␤-D-glucanases, including site-directed mutagenesis, affinity labeling, and x-ray crystallography analysis. The Glu 134 and Glu 138 amino acid residues within the Bacillus licheniformis enzyme have been identified as the nucleophile and the catalytic acid/base, respectively (11,12). The Glu 105 and Glu 109 residues of Bacillus macerans and H(A16-M)1,3-1,4-␤-D-glucanases have also been shown to confer a catalytic function similar to that observed for Glu 134 and Glu 138 of the B. licheniformis enzyme (4,12,13).
A key ruminal bacterial enzyme producer, Fibrobacter succinogenes, plays a major role in plant fiber degradation in the rumen of major livestock species. Several of the enzymes related to the degradation of cellulose or hemicellulose polymers of plant cell walls from this organism have been isolated and partially characterized (14). A F. succinogenes 1,3-1,4-␤-D-glucanase was isolated and investigated by Erfle and co-workers (15,16). This enzyme consists of a protein sequence with a circular permutation in which two highly conserved catalytic domains (namely A and B) of the enzyme are in a reverse orientation as compared with that of other 1,3-1,4-␤-D-glucanases (6,8,16). A 5 times repeated segment, PXSSSS, was only observed in the C-terminal, nonhomologous region of the amino acid sequence of the Fibrobacter enzyme relative to the bacilli or other bacterial and fungal 1,3-1,4-␤-D-glucanases. Nevertheless, alignment of the amino acid sequences of the F. succinogenes enzyme with other 1,3-1,4-␤-D-glucanases suggests that a number of amino acid residues in the highly conserved region may play important roles in catalysis of the enzyme (see Fig. 1). In an attempt to identify the possible involvement of specific amino acid residues in the catalytic activity of F. succinogenes 1,3-1,4-␤-D-glucanase, we evaluated the potential functional significance of the Met 39 , Glu 56 , Asp 58 , Glu 60 , and Gly 63 residues of the enzyme, using a combination of various approaches including site-directed mutagenesis, fluorescence spectroscopy, circular dichroism spectrometry, kinetics, and structural modeling. Several lines of evidence were obtained, providing some useful information about the structural and functional relationship of F. succinogenes 1,3-1,4-␤-D-glucanase. The most significant findings of this investigation are: 1) substitutions of Glu 56 , Asp 58 , and Glu 60 with either alanine or glutamine can completely abolish enzymatic activity; 2) replacement of Gly 63 with alanine can greatly reduce thermostability of the enzyme; and 3) Met 39 is essential for the 1,3-1,4-␤-D-glucanase of F. succinogenes to maintain the most effective catalytic efficiency for the enzyme. This study provides new information that may be used to improve this ruminal enzyme for industrial utilization as a feed or food-processing aid, using either rational design or DNA shuffling approaches.

EXPERIMENTAL PROCEDURES
Subcloning of F. succinogenes 1,3-1,4-␤-D-Glucanase Gene-The fulllength cDNA of F. succinogenes 1,3-1,4-␤-D-glucanase (Fs␤-glucanase) 1 in a pIJ10 plasmid was amplified and introduced with NcoI and EcoRI restriction enzyme recognition sites at the 5Ј and 3Ј ends, respectively, by using a PCR-based method. The two primers designed for the NcoI and EcoRI sites were 5Ј-TCACCACCATGGTTAGCGCAAAG-3Ј and 5Ј-GCCACGAATTCTGTTCAAAGTTCAC-3Ј, respectively. The PCR reaction was performed with a thermocycling program as follows: 94°C for 5 min, 55°C for 1 min, and 72°C for 1 min for 1 cycle; 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min for 30 cycles; and 94°C for 1.5 min, 55°C for 1.5 min, and 72°C for 10 min for 1 cycle. The resulting amplified DNA fragments were digested with NcoI and EcoRI, purified, and ligated onto the pET26b(ϩ) vector, which was predigested with NcoI and EcoRI. The recombinant gene sequence for Fs␤-glucanase, designated "pFsNcE," was confirmed by DNA sequencing using the chain termination method (17). In this DNA construct, a pelB leading peptide at the N terminus plus 19 extra amino acid residues including a His 6 tag at the C terminus to facilitate protein purification were included. The recombinant plasmid encoding for the wild type enzyme was then transformed into Escherichia coli BL21(DE3) host.
Mutant DNA was generated with a thermocycling program of 2 min at 95°C and 16 cycles of 30 s at 95°C, 60 s at 55°C, and 12 min at 68°C on a Hybaid TouchDown thermal cycler. The PCR-generated products were digested with 10 units of DpnI at 37°C for 1 h, prior to their use for transformation into XL-1 Blue cells. Mutations were confirmed by fluorescent dideoxy chain termination DNA sequencing using T 7 promoter and T 7 terminator primers. The mutagenesis plasmid was then transformed into BL21(DE3) host cells for the overexpression of mutant enzyme. Protein Production and Cellular Localization of Fs␤-glucanase in E. coli-Optimal protein production conditions and cellular localization of Fs␤-glucanase in E. coli cells were investigated. 5 ml of pregrown culture of the BL21(DE3) bacterial strain carrying pET26b(ϩ) containing the Fs␤-glucanase gene was added to 500 ml of fresh LB broth containing 30 g/ml kanamycin. The culture was shaken vigorously at 33°C until the A 600 nm reached 0.4 -0.6. Addition of 1 mM isopropyl-␤-D-thiogalactopyranoside (IPTG) to the culture then was performed, and the culture was further incubated for 12-24 h at 33°C. Small amounts of culture aliquots were collected with a constant time interval after the IPTG induction. The culture medium and cell extract prepared from the collected cell pellet at different time periods of IPTG induction were then employed for enzymatic activity assay, SDS-polyacrylamide gel electrophoresis (PAGE) according to Laemmli (19), and zymogram analysis.
Purification of Wild Type and Mutant Fs␤-glucanase-The wild type or mutant forms of Fs␤-glucanase produced in the above-described procedure were further purified. Approximately 80 -85% of the total Fs␤-glucanase expressed in E. coli host cells was secreted into the culture medium. The extracellular secreted enzyme was collected by centrifugation at 8,000 ϫ g for 15 min at 4°C and concentrated 10 times by volume using a Pellicon Cassette concentrator (Millipore, Bedford, MA) with a 10,000 M r cut-off membrane. The concentrated supernatant was then dialyzed against 50 mM Tris-HCl buffer, pH 7.8 (buffer A), and loaded onto a Q-Sepharose FF (Amersham Pharmacia Biotech) column pre-equilibrated with the same buffer. 1,3-1,4-␤-D-Glucanase proteins, either wild type or mutants, were collected from eluants of the column using a 0 -1 M NaCl salt gradient in buffer A. A second nickel-nitrilotriacetic acid affinity column equilibrated with 50 mM sodium phosphate, pH 8.0, 0.3 M NaCl, and 10 mM imidazole buffer (buffer B) was then employed for further purification of the enzymes. From a 10 -300 mM imidazole gradient eluant, a homogeneous enzyme preparation was obtained, as verified by SDS-PAGE. Protein concentration was quantified as described by Bradford (20) with bovine serum albumin as the standard.
Zymogram Analysis-A zymogram was used to measure the enzymatic activity of the wild type and mutant forms of Fs␤-glucanase, which was performed essentially according to a reported method (21) with minor modifications. A 12% SDS-polyacrylamide gel containing lichenan (1 mg/ml) and protein samples in sample buffer (19) pretreated at 90°C for 10 min was prepared for zymogram analysis. After electrophoresis, the gel was rinsed twice with 20% isopropyl alcohol in 50 mM sodium citrate buffer, pH 6.0, for 20 min to remove SDS, and then equilibrated in 50 mM sodium citrate buffer for 20 min. Before staining with Congo red solution (0.5 mg/ml), the gel was preincubated at 40°C for 10 min. The protein bands with 1,3-1,4-␤-D-glucanase activity were then visualized using the Congo red staining.
N-terminal Amino Acid Sequencing-Protein samples for N-terminal sequence determination were resolved on a 12% SDS-polyacrylamide gel followed by electrophoretic transfer onto a polyvinylidene difluoride membrane, using a Mini-Trans-Blot cell system (Bio-Rad). Transferred protein bands on the membrane were visualized using the 0.1% Amido Black staining and then excised with a clean sharp razor blade. Nterminal amino acid sequencing was carried out on an Applied Biosystems model 492 gas phase sequencer equipped with an automated on-line phenylthiohydantoin analyzer.
Kinetic Studies-The enzymatic activity of 1,3-1,4-␤-D-glucanase was measured by determining the rate of reducing sugar production from the hydrolysis of substrate (lichenan). The reducing sugar was measured and quantified by the method of Miller (22) with glucose as the standard. A standard enzyme activity assay was performed in a 0.3-ml reaction mixture containing 50 mM sodium citrate buffer (pH 6.0) and 2.7-8 mg/ml lichenan by starting the reaction with the enzyme addition. After incubation at 40°C for 10 min, the reaction was terminated by the addition of a salicylic acid solution (22). One unit of enzyme activity was defined as the amount of enzyme required for releasing 1 mol of reducing sugar (glucose equivalent). The specific activity is expressed as mol of glucose/min/mg of protein. Various amounts of the purified enzymes (0.24 -82.7 g/ml) were used in each kinetic assay reaction, depending on the enzymatic activity of the enzyme. Kinetic data were analyzed using the computer program ENZFITTER (Biosoft) and using enzyme kinetics. Circular Dichroism Spectrometry-CD studies on the wild type and mutant forms of F. succinogenes glucanase were carried out in a Jasco J715 CD spectrometer and a 1-mm cell at 25°C. Spectra were collected from 200 to 260 nm in 1.3-nm increments, and each spectrum was blank collected and smoothed by using the software package provided with the instrument.
Fluorescence Spectroscopy-The fluorescence emission spectra of the wild type and mutant forms of F. succinogenes 1,3-1,4-␤-D-glucanase were taken on an Amico-Bowman series 2 spectrometer (Spectronic Instruments, Inc.) at 25°C with 1 ϫ 1-cm cuvette. Excitation spectra were taken at 282 nm, and emission spectra were recorded at 302-440 nm, with a 4-nm slit for both spectra. Protein samples were treated with 4 -8 M urea or without any pretreatment (native form) in 50 mM phosphate pH 7.0 buffer before the spectra were recorded. The urea-denatured samples were followed with a renaturation procedure by dialysis against 50 mM phosphate pH 7.0 buffer at 4°C for 24 h, and the fluorescence emission spectra of the protein samples were then taken at the same parameters. The protein concentration for the wild type and mutant forms of the enzyme was 30 g/ml for each measurement.

Expression and Cellular Localization of the Recombinant
Fs␤-glucanase in E. coli Cultures-The conditions for expression of Fs␤-glucanase in engineered E. coli cells were optimized in this study. The wild type and mutant forms of the enzyme were effectively expressed and secreted into the LB medium as a soluble protein when host cells were grown at 33°C with IPTG induction. Production of the enzyme in the whole E. coli culture was detected 2 h after IPTG induction and reached a plateau of maximum activity 8 -16 h postinduction (data not shown). It was also found that enzymatic activity was detected primarily in the cell-associated fraction (cytosolic form) during the early stage (2-8 h) of induction, and upon prolonged IPTG induction for up to 16 h, enzymatic activity was mostly detected in the conditioned culture medium (extracellular form). The pelB leader sequence in the plasmid DNA construct was apparently fully functioning and facilitated the effective secretion of Fs␤-glucanase into the culture medium. Approximately 60% of the extracellular secretion form of the proteins was found as Fs␤-glucanase after IPTG induction for 16 -20 h, as determined using SDS-PAGE and zymogram analyses (data not shown). The results from SDS-PAGE and zymogram analyses are also in good agreement with data from enzymatic activity assays (data not shown), suggesting that the E. coli expressed enzyme can be collected as either a cytosolic or an extracellular form from the host bacterial culture.
Purification and Biochemical Characterization of Wild Type and Mutant Forms of Fs␤-glucanase-Homogeneous preparations of various recombinant enzymes were obtained by frac-tionation with a Q-Sepharose cation exchange column and followed by a separation with nickel-nitrilotriacetic acid affinity column, as described under "Experimental Procedures." The first 25-amino acid sequence at the N terminus of the purified Fs␤-glucanase was determined to be MVSAKDFSGA-ELYTLEEVQYGKFEA, which represents a matured form of the Fs␤-glucanase enzyme without the presence of a pelB leader peptide at the N terminus. The wild type enzyme has a molecular mass of 37,669 Da, as determined by mass spectrometry. The estimated isoelectric point of the recombinant enzyme is pH 6.7, as analyzed by a Genetics Computer Group, Inc. (Madison, WI) computer program. The three-dimensional structure of this enzyme, to our knowledge, has not been solved so far. Therefore, for the current study the target amino acid residues for mutation were chosen based on the evaluation and comparison of the amino acid sequence of the 1,3-1,4-␤-D-glucanases isolated from different organisms and on the prediction of their possible roles in catalysis. Fig. 1 shows that several of the amino acid residues of Fs␤-glucanase, including Glu 56 , Asp 58 , Glu 60 , and Gly 63 , are all conserved in the compared amino acid sequences. Methionine 39 is the only nonconservative amino acid residue observed in Fs␤-glucanase; in other words, the equivalent residues to position 39 of Fs␤-glucanase among other compared bacterial or fungal enzymes are all hydrophobic residues, including phenylalanine, isoleucine, and leucine. The expression conditions and purification protocol for the 11 mutant enzymes, namely M39F, E56A, E56D, E56Q, D58A, D58E, D58N, E60A, E60D, E60Q, and G63A, were similar to that of the wild type enzyme. Purity of the wild type and mutant enzymes was evaluated by SDS-PAGE analysis (Fig.  2). The wild type and the 11 mutant enzymes exhibited identical mobility and are present as greater than 96% homogeneity when using electrophoresis as a criterion. Zymogram analysis revealed that the mutant enzymes showed a similar or reduced level of enzymatic activity as compared with the wild type enzyme (data not shown). Similar protein expression profiles and yield levels were obtained from the culture supernatants collected for the wild type and the 11 mutant forms of Fs␤glucanase, as judged by SDS-PAGE analysis, and protein concentrations were determined by Bradford assay.
Fluorescence and Secondary Structure Analysis-The emission fluorescence spectra of the native, urea-denatured, and denatured-renatured wild type and mutant forms of Fs␤-glucanase were analyzed. Fig. 3 shows the superimposed emission spectra of both native wild type and G63A enzymes, with a maximum emission peaked at 336 nm. The emission spectra of 8 M urea-denatured wild type and mutant enzymes were batho- FIG. 1. Amino acid sequence alignment of the putative catalytic domain of 1,3-1,4-␤-D-glucanases (Lic) and 1,3 (7); Fs-Lic, F. succinogenes (16). The alignment was optimized by introducing gaps, denoted by a dot, and residues that are highly conserved in all sequences are shaded. Numbers on the right are the residue numbers of the last amino acid in each line. Asterisks denote the candidate residues for mutation in this study (R. Borriss and M. Krah, unpublished data; EMBL Data Base accession number O52754.sp_ bacteria). chromic (red) shift and also superimposed with a maximum emission at 350 nm and a slight shoulder peak at 376 nm (Fig.  3). Fluorescence spectra similar to those of the native enzyme were also observed for the wild type and mutant enzymes pretreated with 8 M urea and followed with a renaturation protocol (via dialysis against phosphate buffer for 24 h). In this case, the maximum emission spectra for the wild type and G63A proteins have been found to shift back to 336 nm, and the spectra were very similar to those of native enzymes. These phenomena on emission fluorescence spectra of native, denatured, and denatured-renaturated forms of the protein for the wild type and G63A enzymes were also obtained for the E56Q mutant, which is an enzymatically inactive Fs␤-glucanase.
CD spectrometry was employed for analyzing the performance of the native and urea-denatured then renatured wild type or mutant enzymes. All of the tested native forms of the wild type and the 11 mutant enzymes were shown to exhibit a maximum CD absorbance at 225 nm (data not shown). The renatured forms of the wild type, E56Q, and G63A glucanases showed CD spectra between 200 and 260 nm similar to their correspondent native forms of the proteins (data not shown). These results suggest that the single amino acid substitution of wild type Fs␤-glucanase in this study did not cause a global conformational change or aberrant folding of the enzyme. Folded structures were observed for the individual mutant enzymes as evidenced by analysis of the fluorescent and CD spectra, relative to those of the wild type enzyme. Therefore, the observed differences in the kinetic properties between the mutant and wild type forms of the Fs␤-glucanase enzyme apparently are not due to the disruption of the structural integrity of the enzyme.
Enzymatic activity assays were further carried out for the native and renatured form of the wild type and mutant enzymes. After renaturation, the wild type and mutant enzymes showed a recovery of ϳ85% activity relative to their original native forms of the enzyme. The results from the fluorescence, CD, and activity assays suggest that a great majority of the denatured Fs␤-glucanases, either as wild type or mutants, was able to effectively reassociate to their correspondent native forms of the protein. This study hence concludes that a reversible denaturation capability was observed for Fs␤-glucanase after the treatment with a high concentration of urea.
Kinetic Analysis of Wild Type and Mutant Forms of Fs␤glucanase-Experiments on kinetic studies were mainly performed by using lichenan as the substrate in standard enzyme assays, as described under "Experimental Procedures." The specific activity of the recombinant, wild type Fs␤-glucanase expressed in E. coli cells is 1388 Ϯ 82 units/mg, estimated with lichenan at 40°C in 50 mM sodium citrate buffer, pH 6.0. This value is slightly higher than that (960 units/mg) reported for the native enzyme (15). The affinity for lichenan (K m ), the turnover number (k cat ), and the catalysis efficiency (k cat /K m ) of the wild type enzyme were 1.91 mg/ml, 871 s Ϫ1 , and 456 s Ϫ1 (mg/ml) Ϫ1 , respectively.
In an attempt to evaluate the effect of specific amino acid substitutions on enzyme activity or other functions, the kinetic properties of the mutant enzymes were characterized. A dramatic change in the turnover rate was found for the mutations on amino acid residues Glu 56 , Asp 58 , and Glu 60 . For the E56A, E56Q, D58A, E60A, and E60Q mutants, we showed that no enzymatic activity was detectable for these variant protein forms. 2.5-, 241-, 570-, 880-, 540-, and 2.8-fold decreases in k cat were observed for M39F, E56D, D58N, D58E, E60D, and G63A, respectively, relative to the k cat value of the wild type enzyme. These results suggest that amino acid residues Glu 56 , Asp 58 , and Glu 60 may play an important role(s) for enzymatic activity and may be directly involved in the catalysis of Fs␤-glucanase. Small changes in K m values for lichenan were observed for the E56D, D58E, D58N, E60D, and G63A mutant enzymes; however, a 5.3-fold higher K m for lichenan was found for M39F as compared with the wild type enzyme. Comparison of the specificity constants, k cat /K m , is shown in Table II. The specificity constants decreased 4.5-and 14-fold in G63A and M39F, respectively. However, more significant changes (ϳ210 -970-fold decrease) were observed in the single mutations of Glu 56 , Asp 58 , and Glu 60 , relative to the wild type enzyme. These results indicate that the three acidic amino acid residues in Fs␤-glucanase may play important roles in the catalysis of the enzyme. We have also examined the substrate specificity of the wild type and mutant forms of Fs␤-glucanase, by using lichenan, barley ␤-glucan, larminarine, carboxymethyl cellulose, and xylan as test substrates in enzymatic activity assays. No detectable binding affinity was observed for the wild type or mutant Fs␤-glucanase enzymes when larminarine, carboxy- methyl cellulose, or xylan was used as the substrate.
Various buffers at different pH values were employed for evaluating the optimum pH and pH tolerance of the Fs␤-glucanase enzyme. Mutant and wild type enzymes exhibited similar pH response profiles between pH 4 and pH 10 with the pH optimum value being between 6 and 8 (conferring ϳ100% enzyme activity). At pH 5.0 and pH 10.0, the wild type and mutant enzymes were shown at only ϳ20 and 40% of their optimal activities, respectively (data not shown). The optimum temperatures for the wild type and D58N Fs␤-glucanase were observed at 50°C; however, for the M39F, E56D, D58E, and E60D mutants the optimum temperatures were found to have shifted to 40°C as compared with the wild type enzyme (data not shown). The G63A mutant shows an optimum temperature at 30°C, which is 20°C lower than that of the wild type enzyme.
Temperature Sensitivity of Wild Type and Mutant Fs␤-glucanases-The temperature sensitivity of wild type and mutant Fs␤-glucanases was investigated to further characterize the effect of introduced mutations. Replacement of amino acid residues in Met 39 , Glu 56 , Asp 58 , and Glu 60 did not cause significant changes in thermostability, as evaluated with a temperature range between 30 and 90°C. The wild type and mutant enzymes with specific mutations of acidic amino acid residues were shown to exhibit similar enzymatic activity between 20 and 45°C, but a dramatic loss of enzymatic activity was observed at temperatures higher than 50°C. In contrast, the G63A variant exhibited a greatly impaired thermostability, i.e. only 26% of the original activity was observed when the enzyme was treated at 35°C, and less than 10% of original activity was obtained when treated at a temperature higher than 40°C (Fig. 4).
Homologous Modeling-F. succinogenes 1,3-1,4-␤-D-glucanase is the only native, circular permuted protein reported so far in the family 16 endoglucanase in which the primary amino acid sequences of the enzyme family are arranged in the order of domain A followed by domain B. However, a reverse orientation of the two domains (B to A) was observed for Fs␤glucanase. The three-dimensional structure of a de novo circularly permuted variant, cpA16M-59, was reported by Hahn et al. (23). This cpA16M-59 1,3-1,4-␤-D-glucanase variant was generated by PCR mutagenesis on the gene encoding an H(A16-M) hybrid enzyme with a sequence organization corresponding to the F. succinogenes protein. In this study, a structural model of Fs␤-glucanase was built by Modeller 4 (24), using the homologous mimic enzyme cpA16M-59 with Protein Data Bank entry 1cpm (23) as a template. The entire structural model was briefly energy-minimized, and the final model has 96% of the nonglycyl and nonpropyl residues falling in the most favored or in the additionally allowed regions in the Ram-achandran plot, as analyzed by the PROCHECK program (25). The modeled Fs␤-glucanase structure consists mainly of two antiparallel ␤-sheets with seven and eight strands, respectively (Fig. 5), arranged atop each other to form a compact, sandwichlike structure. The overall ␤-sheet model structure of the Fs␤glucanase is similar to that of cpA16M-59 with only minor changes in the surface loop regions. The two ␤-sheets are bent to give rise to a concave and a convex side of the molecule. The Glu 56 , Asp 58 , and Glu 60 amino acid residues are located at the cleft on the concave side of the protein molecule (Fig. 5), and this cleft was likely the substrate binding side as previously suggested based on the structure analysis of a protein-inhibitor complex of the Bacillus 1,3-1,4-␤-D-glucanase with epoxyalkyl ␤-oligoglucosides (4). DISCUSSION Bacterial 1,3-1,4-␤-D-glucanases belong to the category of retaining enzymes and are classified into family 16 endoglucanases (3). More than 50% protein sequence homology was found among the bacterial enzymes, including those from Clostridium and Bacilli and the ruminal fungus Orpinomyces (5). In comparison, only ϳ30% homology was found for the F. succinogenes 1,3-1,4-␤-D-glucanase with respect to the other bacterial enzymes. In addition, a naturally occurring circular permutation structure was only found for the F. succinogenes 1,3-1,4-␤-D-glucanase. Although the biochemical properties of the native F. succinogenes enzyme (15) and its encoding cDNA sequence have been characterized (16), further studies on the structure-function relationship were not reported. In this study, a number of conserved amino acid residues, as revealed by sequence comparison studies, were investigated for their potential involvement in the catalysis of the enzyme.
Studies on the three-dimensional structure of a covalent protein-inhibitor complex of H(A16-M) with 3,4-epoxybutyl-␤-D-cellobioside (4) have shown that the inhibitor binds to the Glu 105 amino acid residue, which corresponds to the Glu 56 residue of the Fs␤-D-glucanase enzyme, based on the amino acid sequence alignment of these two proteins. Site-directed mutagenesis and chemical modification of the glutamic acid residues, e.g. Glu 105 and Glu 109 in H(A16-M), Glu 105 in Bacillus amyloliquefaciens (26), Glu 103 in B. macerans (27), and Glu 134 and Glu 138 in B. licheniformis (11), suggested that the glutamic  4. Temperature sensitivity of wild type and mutant forms of F. succinogenes 1,3-1,4-␤-D-glucanase. The purified wild type (q), M39F (E), E58Q (f), and G63A (Ⅺ) enzymes at a enzyme concentration of 0.007-1.24 mg/ml were incubated for 10 min at 35,40,45,50,55,60,65,70,75, and 80°C, respectively, in 50 mM citrate buffer, pH 6.0. Enzyme activity was assayed by the method of Miller (22), as described under "Experimental Procedures," immediately after incubation and is expressed as a percentage of the relative activity at 40°C. The protein concentration was 0.24 -82.7 g/ml in each assay. Each assay was performed either in duplicate or in triplicate. acid residues are crucial to catalysis of 1,3-1,4-␤-D-glucanase. In this study, possible functions of three acidic amino acid residues, including Glu 56 , Asp 58 , and Glu 60 were investigated. These three residues are located at the cleft of the concave side of the Fs␤-glucanase protein molecule, as shown in the structure model in Fig. 6. Dramatic decreases in the turnover rate (k cat ) of Fs␤-glucanase and minor changes in lichenan affinity relative to that of wild type enzyme were found for the single mutations of Glu 56 , Asp 58 , and Glu 60 (Table II). These results suggest that the structure of these mutant forms of Fs␤-glucanase remain undisturbed, thus allowing the substrates to bind to the enzyme. The results from the fluorescence and circular dichroism spectrometric studies also show evidence that the mutant enzymes can retain a folded protein structure similar to that of the wild type enzyme without undergoing significant global structural changes (Fig. 3). However, severe disruption of the enzymatic activity of Fs␤-glucanase demonstrates that these three acidic residues may play important roles in the catalytic mechanism.
Structural modeling of the mutant enzymes was also undertaken to gain more insight into the effect of mutation of Glu 56 , Asp 58 , and Glu 60 . In this model, the carboxyl group (O⑀-1 atom) of Glu 56 , corresponding to Glu 105 of cpA16M-59, is hydrogen bonded to the carboxyl group (O␦-1) of Asp 58 (Fig. 6). Isofunctional replacement of Glu 56 , Glu 60 , and Asp 58 with aspartate or glutamate, as in mutants E56D, D58E, and E60D, yield low but detectable activities. This indicates that a similar hydrogenbonding network may be still maintained in the mutant Fs␤glucanases with only minor local structural rearrangements in the side chains, which do not change the substrate binding affinity in the mutant enzymes. Isosteric replacement of the Glu 56 and Glu 60 with glutamine rendered the enzyme inactivated, although very low level residual enzymatic activity was still detectable (e.g. 0.2-0.4% activity relative to the wild type enzyme; Table II) when the glutamic acid residues were substituted with aspartate. It has been suggested that the ionizable carboxylic acid groups in the catalytic site, acting as a general acid for proton transfer or as a nucleophile in their anionic form, were found to be essential for the catalysis of Bacillus 1,3-1,4-␤-D-glucanase. These observations are in good agreement with the previous studies on Glu 134 and Glu 138 of B. licheniformis (11,12) and on Glu 105 and Glu 109 of B. macerans enzymes (26,27), using site-directed mutagenesis and chemical rescue approaches. The residue equivalent to Glu 56 was proposed to act as a catalytic nucleophile, and the Glu 60 counterpart was proposed to function as a general acid in the Bacillus 1,3-1,4-␤-D-glucanases. Moreover, a comparison between the activity of E56Q (not detectable), D58N (0.2%), and E60Q (not detectable) further shows that Glu 56 and Glu 60 may be directly involved in catalysis and play important roles in the enzymatic reaction. Detectable residual activity in D58N demonstrates that the carboxyl group is not mandatory for the enzymatic reaction. It is speculated here that Asp 58 may play a structural role in stabilizing the active site structure as well as an important electrostatic role in affecting the pK a of the nucleophilic residue of Glu 56 , as was previously suggested for the equivalent residue of Asp 58 in B. licheniformis glucanase (Asp 136 ) (11). The enzymatic activities of E56A, D58A, and E60A were not detected, suggesting that the alanine side chain is neither capable of making hydrogen bonds with the neighboring residues nor capable of serving in an electrostatic role or acting as a general acid-base in the hydrolysis reaction of Fs␤-glucanase, which are important for maintaining the catalytic function of the enzyme.
In the present Fs␤-glucanase structure model, Met 39 is lo- cated at a position pointing away from the active site and is buried in the hydrophobic core of the enzyme. Based on the amino acid sequence comparison of the bacterial and fungal 1,3-1,4-␤-D-glucanases (Fig. 1), the corresponding position to the amino acid residue Met 39 of Fs␤-glucanase is phenylalanine or other hydrophobic residues (e.g. Ile or Leu) in the compared bacterial or fungal enzymes. In this study, we intended to mutate Met 39 of Fs␤-glucanase to phenylalanine, and that resulted in 2-, 5-, and 14-fold decreases in k cat , affinity for lichenan, and k cat /K m , respectively, as compared with that of the wild type enzyme. These results indicate that the methionine at position 39 is essential for the Fs␤-glucanase to maintain the most effective catalytic efficiency for the enzyme. Moreover, this mutation apparently has resulted in a more heat-sensitive Fs␤-glucanase; only ϳ30% of original activity was detected after a heat treatment at 45°C for 10 min. With the same heat treatment, ϳ80% of original activity was maintained for the wild type and the D58N mutant enzymes. However, this mutation is tolerated in the folding of the entire structure, and this interpretation also supported the results of fluorescence spectrometry assays (data not shown).
Slight decreases in specific activity and substrate binding affinity were observed for the replacement of Gly 63 with alanine; however, significant reduction of the enzyme stability was also observed for this mutant. After incubation at 35°C for 10 min, the G63A mutant only retains ϳ25% of its original enzymatic activity. Therefore, the conserved Gly 63 residue among various bacterial and fungal 1,3-1,4-␤-D-glucanases ( Fig. 1) is first reported in this study as an important residue for enzyme stability. This residue is located at the end of a ␤-strand, which contains three active site residues, Glu 56 , Asp 58 , and Glu 60 . It has been shown that the thermal stability of the Fis protein (28) and the Arc repressor (29) was largely reduced when glycines located in turn regions were mutated to alanines. Therefore, it is possible that the Gly 63 plays a similar role in 1,3-1,4-␤-D-glucanases in that it stabilizes the folding of the entire protein. We found that only ϳ5% residual activity was detected in the mutant enzyme treated with 40°C for 10 min, whereas the wild type was still fully active after the same treatment.
In summary, this report first identified several key amino acid residues that are involved in the catalysis of F. succinogenes 1,3-1,4-␤-D-glucanase. It is proposed here that Glu 56 and Glu 60 may function as general acid/base residues based on the detailed comparison in kinetics. We also demonstrated that Gly 63 is important for the stability of the enzyme. Our study has provided useful information on the structure-function relationship of the naturally occurring, circular permutated protein structure of a bacterial 1,3-1,4-␤-D-glucanase isolated from rumen F. succinogenes. Further studies of the three-dimensional structure of the enzyme using the x-ray crystallography approach are in progress.