Characterization of a bifunctional cellulase and its structural gene. The cell gene of Bacillus sp. D04 has exo- and endoglucanase activity.

Bacillus sp. D04 secreted a bifunctional cellulase that had a molecular weight of 35,000. This cellulase degraded Cm-cellulose, cellotetraose, cellopentaose, p-nitrophenyl-β-D-cellobioside, and avicel PH101. Based on the high performance liquid chromatography analysis of the degradation products, this cellulase randomly cleaved internal β-1,4-glycosidic bonds in cellotetraose and cellopentaose as an endoglucanase. It also hydrolyzed the aglycosidic bond in p-nitrophenyl-β-D-cellobioside and cleaved avicel to cellobiose as an exoglucanase. Cellobiose competitively inhibited the p-nitrophenyl-β-D-cellobioside degrading activity but not Cm-cellulose degrading activity. Ten mMp-chloromercuribenzoate inhibited p-nitrophenyl-β-D-cellobioside degrading activity completely, but Cm-cellulose degrading activity incompletely. Cm-cellulose increased p-nitrophenyl-β-D-cellobioside degrading activity, and vice versa, whereas methylumbelliferyl-β-D-cellobiose strongly inhibited p-nitrophenyl-β-D-cellobioside degrading activity. The cellulase gene (cel gene), 1461 base pairs, of Bacillus sp. D04 was cloned. The nucleotide sequence of the cel gene was highly homologous to those of Bacillus subtilis DLG and B. subtilis BSE616. The cel gene was overexpressed in Escherichia coli, and its product was purified. The substrate specificity and substrate competition pattern of the purified recombinant cellulase were the same as those of the purified cellulase from Bacillus sp. D04. These results suggest that a single polypeptide cellulase had both endo- and exoglucanase activities and each activity exists in a separate site.

Cellulosic substrates hydrolyzed by only one type of cellulase are catagorized as follows. Acid-swollen cellulose, Cm-cellulose, cellulose azure, and trinitrophenyl Cm-cellulose are hydrolyzed by endoglucanases (1). MUC 1 (4) and pNPC (5) are used as substrates for the determination of exoglucanase activity, and MUG (4) and pNPG (5) are cleaved by ␤-glucosidases. Filter paper and avicel are efficiently hydrolyzed by the synergistic action of endo-and exoglucanases, but not by either one alone (6). Some organisms (for example, Trichoderma sp.) (6 -9, 11) produce all three types of cellulases and efficiently degrade cellulose by their synergistic effect. A cellulolytic hydrolase with a considerable level of endo-, exoglucanase, and xylanase activity has been described (3,(12)(13)(14). For example, Saul reported a cellulase gene (cel B) of Caldocellum saccharolyticum with a Cm-cellulose-degrading domain in the C-terminal region and an MUC degrading domain in the N-terminal region (14). A polysaccharide hydrolase of the rumen fungus Neocallimatrix patriciarum has a multifunctional catalytic domain with high endoglucanase, cellobiohydrolase, and xylanase activities (12,13).
Extensive recent studies on proteins (such as cellulase, protease, and amylase) secreted by Bacillus species (15) have shown that the following Bacillus species produce cellulases: Bacillus cereus (16), Bacillus licheniformis (17), Bacillus subtilis (18), and Bacillus polymyxa (19). Because these strains did not produce all three types of cellulase, they did not extensively hydrolyze crystalline cellulose. We have investigated another strain of this species, Bacillus sp. D04, having the ability to degrade crystalline cellulose. We have determined that the cellulase of Bacillus sp. D04 differed from that of other Bacillus species in several respects. In particular it has both endo-and exoglucanase activity. It degraded Cm-cellulose, cellotetraose, and cellopentaose as an endoglucanase and cleaved aglycosidic bonds in pNPC as an exoglucanase. It also cleaved avicel to cellobiose. Substrate competition assays showed that the cellulase of Bacillus sp. D04 had separate sites for endo-and exoglucanase activity. * This work was supported in part by grants from the Korea Ministry of Education, Korea Institute of Energy and Resources, and the Research Center for Molecular Microbiology, Seoul National University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U27084 for cel gene cellulase.
§ To whom correspondence should be addressed.

Purification of Cellulase
Bacillus sp. D04 was cultured in 2 liters of medium (containing M9 minimal salts, 0.5% glucose and 0.5% avicel) at 45°C for 13 h. After the medium was centrifuged at 11,000 ϫ g for 10 min, the supernatant was concentrated by ultrafiltration (10,000 nominal molecular weight cutoff membrane was used). Ten mM potassium phosphate buffer (pH 5.8) was added to the concentrated sample, and the sample was reconcentrated to 100 ml. This sample was passed through Cm-Sepharose CL-6B (7 ϫ 50 mm) equilibrated with 10 mM potassium phosphate buffer (pH 5.8) at a flow rate of 15 ml/h. The sample flow-through was loaded directly onto hydroxylapatite (7 ϫ 30 mm) equilibrated with 10 mM potassium phosphate buffer (pH 5.8) and eluted with a 30-ml 10 -250 mM potassium phosphate salt gradient at a flow rate of 10 ml/h. The concentration of protein was measured with Bradford solution (Bio-Rad).

Cellulase Enzyme Assay
Cm-cellulose and Avicel Degrading Activity Assay-The Cm-cellulase assay consisted of 800 l of 1% Cm-cellulose in 10 mM potassium phosphate buffer (pH 5.8) and 200 l of diluted enzyme solution, incubated at 45°C for 20 min. Avicel degrading activity was measured as follows: 500 l of 10 mg/ml avicel in 10 mM potassium phosphate buffer (pH 5.8) was mixed with 500 l of suitably diluted enzyme and then incubated for 72 h at 45°C in a shaking incubator. The remaining avicel was removed by centrifugation, and the amount of reducing sugar was detected with 3,5-dinitrosalicylic reagent. One unit of Cm-cellulose and avicel degrading activity was defined as the amount of enzyme required for producing 1 mol of glucose/min.
pNPC Degrading Activity Assay-The reaction mixture, consisting of 800 l of pNPC at 1 mg/ml in 50 mM sodium acetate buffer (pH 5.8) and 200 l of suitably diluted enzyme, was incubated at 45°C for 20 min. The p-nitrophenol released from pNPC was detected at 420 nm after adding 1 ml of 2% sodium carbonate. One unit of enzyme activity was defined as the amount of enzyme required for producing 1 mol of p-nitrophenol/min.
MUC Degrading Activity Assay-The reaction mixture, consisting of 800 l of 1 mg/ml MUC in 50 mM sodium acetate buffer (pH 5.8) and 200 l of suitably diluted enzyme, was incubated at 45°C for 20 min. The reaction was stopped by adding 3.5 ml of 0.5 M glycine/NaOH buffer (pH 10.4). Fluorescence measurements were made on a Tasco FP-777 spectrofluorometer at 20°C, with an excitation wave length of 365 nm and detection at 450 nm. One unit of enzyme activity was defined as the amount of enzyme required for producing 1 mol of 4-methylumbelliferone/min (9).

High Performance Liquid Chromatography Analysis of Degradation Products
The reaction mixture, consisting of 80 l of oligosaccharides released from cellulosic substrates, 20 l of 10% trichloroacetic acid, and 100 l of 0.3% (w/v) ethanolic solution of dansyl hydrazine, was heated at 80°C for 10 min, and then cooled (20 -22). Samples were dried, dissolved in 78% acetonitrile solution, and analyzed with -Bondapak NH 2 column (3.9 ϫ 300 mm, Waters) (23). The dansyl hydrazone of oligosaccharides were detected at 254 nm. The 78% acetonitrile solution was used as an elution solvent, and the flow rate was 1.5 ml/min.

Cloning and Determination of Nucleotide Sequence of the Cellulase Gene
Chromosomal DNA from Bacillus sp. D04 was partially digested with Sau3AI producing 3-5-kb DNA fragments. These were isolated by density gradient centrifugation in 10 to 40% sucrose using an SW 28 rotor at 20,000 rpm for 20 h at 20°C. Fragments were ligated with the dephosphorylated BamHI site of pBluescript KS(ϩ) and then transferred into an Escherichia coli DH5␣ strain. Transformants with Cmcellulose degrading activity were screened on an L-agar plates containing ampicillin at 100 g/ml, 0.5% Cm-cellulose, and trypan blue at 0.1 mg/ml. To screen MUC degrading activity, transformants were transferred onto L-agar plates containing ampicillin at 100 g/ml and MUC at 50 g/ml. For the determination of the nucleotide sequence of the cel gene, serial deletion of the gene was done by using Erase-a-Base kit (Promega). The sequence of cellulase gene was determined from both strands by the dideoxynucleotide chain termination method using a Sequenase kit (U.S. Biochemical Corp.).

PCR Amplification of the cel Gene and Construction of pCO1
The cel gene in pBluescript KS(ϩ) was amplified by PCR with 5Ј-CATATGAAACGGTCAATCTCT-3Ј (ATG in the NdeI site (CATATG) was a start codon of the cel gene) and M13 reverse primer. Amplification was done by 30 cycles of PCR at standard reaction conditions: reaction volume, 50 l; reaction composition, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl 2 , 50 M dNTP, 2 fmol of template, 10 pmol primer, and 2 units of Vent DNA polymerase; cycle profile, 1 min at 95°C, 1 min at 50°C, 1.5 min at 72°C. The PCR products were purified and digested with HindIII. These fragments were ligated with pBluescript KS(ϩ) that had been digested with SmaI and HindIII. This recombinant DNA was named pCO1.

Construction of pCO 2 and Transfer into E. coli BL21(DE) pLysS
The pCO1 vector was digested with EcoRV and then partially digested with NdeI to generate the 1.5-kb cel gene. The overexpression vector, pET-3a-d (Novagen, Inc.), was digested with BamHI, and then end-filling was done with a Klenow fragment. This overexpression vector was digested with NdeI and then ligated with the 1.5-kb cel gene. This recombinant plasmid was named as pCO 2 . The pCO 2 vector was Overexpression of the Cellulase Gene E. coli BL21(DE)pLysS with pCO 2 was cultured in Luria-Bertani medium containing ampicillin at 50 g/ml and chloramphenicol at 30 g/ml for 12 h and then transferred into TBGM9 medium (tryptone at 10 mg/ml, NaCl at 5 mg/ml, M9 salt, 0.4% (w/v) glucose, and 1 mM MgSO 4 ) containing ampicillin at 50 g/ml. To obtain high levels of transcription, these cells were grown to mid-log phase, IPTG was added to a final concentration of 0.5 mM, and growth continued for 3 h at 30°C. The reaction mixture containing 10 l of 10 mM potassium phosphate (pH 5.8), 40 l of the purified cellulase (0.1 mg/ml), and 30 l of 10 mg/ml cellotetraose were incubated for 0 (a), 60 (b), and 120 (c) min at 45°C. Panel B, HPLC analysis of products released from cellopentaose. 40 l of 10 mg/ml cellopentaose was used as a substrate instead of cellotetraose. These samples were incubated for 0 (a), 60 (b), and 120 (c) min at 45°C. Panel C, HPLC analysis of products released from avicel. 800 l of 1% (w/v) avicel solution and 800 l of the purified cellulase (0.1 mg/ml) were mixed and incubated for 72 h at 45°C. (a) was a control that did not contain cellulase, whereas (b) contained purified cellulase. These reaction products were modified with dansyl hydrazine as described under "Experimental Procedures." The absorbance was measured at 254 nm. The numbers 1-4 represent the dansyl hydrazones of: 1, cellobiose; 2, cellotriose; 3, cellotetraose; and 4, cellopentaose.

Purification of Overexpressed Cellulase
The overexpressed recombinant cellulase from E. coli BL21(DE)pLysS was partially purified by ammonium sulfate fractionation as previously described (24). This partially purified cellulase was dialyzed in 20 mM sodium acetate buffer (pH 4.8) and then loaded onto Cm-Sepharose CL-6B (15 ϫ 50 mm) equilibrated with dialysis buffer. The proteins were eluted with a 80-ml 20 -500 mM sodium acetate (pH 4.8) salt gradient at a flow rate of 20 ml/h.

Activity Staining of Cellulase
The protein sample was mixed with protein loading dye and then incubated at 68°C for 1 h. These samples were loaded onto a 10% polyacrylamide gel containing 0.1% Cm-cellulose, then subjected to electrophoresis. After SDS-PAGE, one of the gels was stained with Coomassie Blue R250. Another was soaked and gently shaken in 50 mM phosphate buffer (pH 6.8) containing 25% isopropanol for 30 min. It was transferred to 50 mM phosphate buffer (pH 6.8) and shaken for 30 min. The buffer was removed, and the gel was incubated for 20 min at 37°C. This gel was stained with 1% Congo Red solution for 5 min and destained with 1 M NaCl/NaOH solution.

RESULTS
Purification of Cellulase-Because Bacillus sp. D04 secreted cellulase into medium, concentrated medium was used as starting material for enzyme purification. Many other proteins were removed by passage through the Cm-Sepharose CL-6B (Fig.  1A). The sample that eluted at 180 mM potassium phosphate from hydroxylapatite had both Cm-cellulose and pNPC degrading activities (Fig. 2). SDS-PAGE of this sample revealed only a 35,000-Da single polypeptide (Fig. 1A). The molecular weight of the native form of this cellulase, determined by gel permeation chromatography (Superose 12, Pharmacia Biotech Inc.), was also about 35,000. Activity staining showed that this purified protein had Cm-cellulose degrading activity (Fig. 1B). The steps in purification of this protein are given in Table I.
Substrate Specificity of the Purified Cellulase-The activity of the purified cellulase was assayed with various cellulosic substrates. This cellulase degraded Cm-cellulose, pNPC, MUC, and avicel PH101 (Table II). However, the specific activity toward avicel was much lower than that of the soluble substrates. Neither MUG nor pNPG was hydrolyzed (Table II).
HPLC Analysis of Oligosaccharides from Cellulosic Substrates-HPLC analysis showed that a single peak was detected at 280 nm as a reaction product on hydrolysis of pNPC by this cellulase (data not shown). The retention time of it was the same as that of p-nitrophenol. Therefore we identified it as p-nitrophenol. This means that enzyme cleaved only aglycosidic bond in pNPC, producing cellobiose and p-nitrophenol. The enzyme cleaved cellulosic substrates (such as cellotetraose and cellopentaose) to glucose, cellobiose, and other oligosaccharides. Since these compounds are not detected at any wave length, we modified them with dansyl hydrazine because sugar dansyl hydrazones could be detected at 254 nm. On the basis of HPLC analysis, the purified cellulase cleaved cellotetraose to cellobiose and cellotriose (Fig. 3A). It also produced cellobiose, cellotriose, and cellotetraose from cellopentaose (Fig. 3B).
These results indicate that the purified cellulase randomly cleaved internal ␤-1,4-glycosidic bonds in these cellulosic substrates as an endoglucanase. Based on the above result, it would seem that the smallest substrate recognized by the endoglucanase of Bacillus sp. D04 is a cellotetraose. Both endoand exoglucanase activities were detected by using cellotetraose and cellopentaose as substrates. But pNPC is not a substrate for endoglucanase of Bacillus sp. D04 because it is shorter than cellotetraose. Therefore only exoglucanase activity was detected by using pNPC as a substrate. The enzyme also produced cellobiose from avicel as an exoglucanase (Fig.  3C).
Differential Inhibition of Cellulase Activity with Various Inhibitors-The Cm-cellulose and pNPC degrading activities of  this cellulase were differentially inhibited by several inhibitors. pCMB at 10 mM completely inhibited the pNPC degrading activity but inhibited Cm-cellulose degrading activity by 67%.
In 40 mM cellobiose, Cm-cellulose degrading activity was not inhibited, but the pNPC degrading activity was inhibited by 57.8% (Table III). Cellobiose changed the K m but not the V max of pNPC degrading activity as a competitive inhibitor (Fig. 4A), whereas both the V max and K m of pNPC degrading activity were changed by pCMB as a mixed-type inhibitor (Fig. 4B). Both Cm-cellulose and pNPC degrading activity required Ca 2ϩ , but were strongly inhibited by Zn 2ϩ . Mg 2ϩ slightly increased pNPC degrading activity and weakly inhibited Cm-cellulose degrading activity (Table III). Substrate Competition-To investigate whether a purified cellulase contains each endo-and exoglucanase active site, we performed substrate competition assays. At a high ratio of Cm-cellulose (0.5%, w/v) to pNPC (0.005%, w/v), pNPC degrading activity was not inhibited, but was increased by Cm-cellulose (Fig. 5A). Cm-cellulose degrading activity was not inhibited, but was slightly increased in the presence of various concentrations of pNPC (Fig. 5B). But 60% of the pNPC degrading activity was inhibited by MUC even at a low ratio of MUC (0.01%, w/v) to pNPC (0.005%, w/v) (Fig. 5C).
Cloning and Nucleotide Sequence of the cel Gene-L-agar plates containing Cm-cellulose and trypan blue were used to clone the gene for Cm-cellulose degrading activity from the genomic library in pBluescript KS(ϩ), which was described under "Experimental Procedures." Cm-cellulose was stained with trypan blue, but the hydrolyzed Cm-cellulose was not. As a result, a halo formed around the colony with the Cm-cellulose degrading activity. Eleven colonies having Cm-cellulose degrading activity were obtained (data not shown). To determine the MUC degrading activity of these colonies, they were transferred onto an L-agar plate containing MUC. The colony with exoglucanase activity cleaved MUC to cellobiose and methylumbelliferone which emitted fluorescence when it was exposed to UV light. All colonies with Cm-cellulose degrading activity emitted fluorescence under the UV light after incubating for 12 h at 37°C on L-agar plates containing MUC (data not shown). Therefore, 11 colonies had a cellulase gene with both Cmcellulose and MUC degrading activities. The nucleotide sequence of this gene (Fig. 6) showed one open reading frame of 1461 base pairs was a possible gene encoding the cellulase. A potential promoter (Ϫ35 (TAGACAAT) and Ϫ10 (TACAAT)) and the Shine-Dalgarno sequence (ribosomal binding site) were identified in the upstream region. Based on the nucleotide sequence homology with other cellulase genes, the cellulase gene of Bacillus sp. D04 has a high homology with those of B. subtilis DLG (26) and B. subtilis BSE616 (27) (Fig. 6).
Characteristics of Recombinant Cellulase-The purified recombinant cellulase degraded Cm-cellulose, pNPC, MUC, and  (Table II), but proteins extracted from E. coli DH5␣ strain did not degrade these cellulosic substrates. Cm-cellulose slightly increased pNPC degrading activity, and vice versa (Fig.  5, D and E). The pNPC degrading activity was strongly inhibited by MUC (Fig. 5F).

DISCUSSION
The following results suggest that the purified 35,000-Da cellulase secreted by Bacillus sp. D04 has both endo-and exoglucanase activity. The endoglucanase of Clostridium themocellum, Cellulomona fimi, and other Bacillus species hydrolyze Cm-cellulose, swollen cellulose, cellotetraose, and cellopentaose, but not pNPC (1). The exoglucanase of Ruminococcus flavafaciens FD-1 (2) and Aspergillus fumigatis (4) hydrolyzed pNPC, MUC, and filter paper, but not Cm-cellulose. However, the cellulase of Bacillus sp. D04 hydrolyzed Cmcellulose, pNPC, and MUC (Table II). Moreover, this cellulase cleaved only the aglycosidic bond in pNPC as does an exoglucanase of Trichoderma viride and Sporotrichum pulveralentum (5), and randomly cleaved internal ␤-1,4-glycosidic bonds in cellotetraose and cellopentaose (Fig. 3, A and B) as an endoglucanase. These results imply that cellulase of Bacillus sp. D04 has both endo-and exoglucanase activity. The presence of both activities in the purified cellulase is confirmed by the fact that this cellulase also degraded crystalline cellulose (Fig. 3C), even though the hydrolysis efficiency of avicel was less than that of soluble cellulosic substrates. Probably, this was due to the low affinity of the purified cellulase against a crystalline cellulose.
To determine whether the active site of endo-and exoglucanase are separately existed, we studied differential effects of compounds that specifically inhibited one type of cellulase activity. The cellobiose competitively inhibited pNPC degrading activity, but did not inhibit Cm-cellulose degrading activity (Table III). However, since the K i of cellobiose was 35.4 mM (Fig. 4A), cellobiose was not a strong inhibitor in pNPC degrading activity. pCMB, a thiol protease inhibitor, inhibited pNPC degrading activity completely and Cm-cellulose degrading activity incompletely (Table III). Therefore endo-and exoglucanase activities were differently inhibited by cellobiose and pCMB. Xue et al. (13) showed that the polysaccharide hydrolase from N. patriciarum has a multifunctional catalytic domain that contains endoglucanase, cellobiohydrolase, and xylanase activities. On the basis of the substrate competition assays of this enzyme, Cm-cellulose and xylan strongly inhibited hydrolysis of MUC (13). Thus, they clearly demonstrated that only one active site has three types of enzyme activities. But the substrate competition pattern of the cellulase of Bacillus sp. D04 was different from those of N. patriciarum. At a high ratio of Cm-cellulose to pNPC or vice versa, one substrate did not inhibit hydrolysis of the another substrate (Fig. 5, A  and B). But as MUC and pNPC were common substrates for exoglucanase, MUC strongly inhibited pNPC degrading activity even if the ratio MUC (0.01%, w/v) to pNPC (0.005%, w/v) was low (Fig. 5C). Thus, above results imply that the purified cellulase has separate sites of endo-and exoglucanase activity.
In order to rule out the possibility that enzymatic activity of either the endo-or the exoglucanase in the purified cellulase from Bacillus sp. D04 is due to a minor contaminating protein, we overexpressed the cel gene from a pET family vector in E. coli and compared its characteristics to those of the purified cellulase from Bacillus sp. D04. We deduced amino acid sequence from the cel gene. The 29 amino acids (from Met (1) to Ala (28), Fig. 6) in the N terminus was a typical ␤-glucanase signal peptide of Bacillus species (28). As an estimated molecular weight based upon amino acid composition of the cel gene was about 55,000. E. coli BL21(DE) pLysS with this gene produced 55,000-Da protein with Cm-cellulose degrading activity. But a 35,000-Da protein with this activity was also detected, which is the molecular mass of the cellulase purified from Bacillus sp. D04. These results indicate that the cellulase was produced as a precursor form from the cel gene and then processed (such as elimination of signal peptide, etc.) to its mature form. The purified 35,000-Da protein with cellulase activity was used as a recombinant cellulase. The substrate specificity and competition pattern of recombinant cellulase were the same as those of a purified cellulase from Bacillus sp. D04. These results clearly eliminate the possibility that the purified cellulase from Bacillus sp. D04 might have a minor contaminating protein involved in catalyzing either the endoor the exoglucanase activity. Therefore a single polypeptide cellulase of Bacillus sp. D04 has both two kinds of activity. To localize each endo-and exoglucanase activity site in the cellulase, we are attempting to develop mutant which has only one type of glucanase activity. The arrowheads point to zones in which Cm-cellulose was degraded by cellulase.