Characterization of a Neopullulanase and an cx-glucosidase from Bacteroides thetaiotaomicron 95-1

Transcription

1 JOURNAL OF BACTERIOLOGY, May 1991, p Vol. 173, No /91/ $2./ Copyright 1991, American Society for Microbiology Characterization of a Neopullulanase and an cx-glucosidase from Bacteroides thetaiotaomicron 95-1 KAREN A. SMITH* AND ABIGAIL A. SALYERS Department of Microbiology, University of Illinois, Urbana, Illinois 6181 Received 2 August 199/Accepted 19 February 1991 Previously, we constructed a gene disruption in the pullulanase I gene of Bacteroides thetaiotaomicron 5482A. This mutant, designated B. thetaiotaomicron 95-1, had a lower level of pullulanase specific activity than did wild-type B. thetaiotaomicron but still exhibited a substantial amount of pullulanase activity. Characterization of the remaining pullulanase activity present in B. thetaiotaomicron 95-1 has identified an a(1->4)-d-glucosidic bond cleaving pullulanase which has been tentatively designated a neopullulanase. The neopullulanase (pullulanase II) is a 7-kDa soluble protein which cleaves a(1-4)-d-glucosidic bonds in pullulan to produce panose. The neopullulanase also cleaved a(1-4) bonds in amylose and in oligosaccharides of maltotriose through maltoheptaose in chain length. An a-glucosidase from B. thetaiotaomicron 95-1 was characterized. The a-glucosidase was partially purified to a preparation containing three proteins of 8, 57, and 5 kda. Puliulan and amylose were not hydrolyzed by the a-glucosidase. a(1->4)-d-glucosidic oligosaccharides from maltose to maltoheptaose were hydrolyzed to glucose by the a-glucosidase. The a-glucosidase also hydrolyzed a(1-6)- linked oligosaccharides such as panose (the product of the pullulanase II action on pullulan) and isomaltotriose. Bacteroides thetaiotaomicron, a gram-negative obligate anaerobe that is found in the human colon in high numbers (6), can ferment a wide variety of polysaccharides (16). Among these polysaccharides are the starch amylose and the starchlike polysaccharide pullulan (Fig. 1). Previously, we reported the cloning and characterization of pullulanase from B. thetaiotaomicron. The cloned pullulanase (puilulanase I) was active as a monomer of approximately 77 kda and cleaved ac(1-6)-d-glucosidic linkages of pullulan to produce maltotriose (18). Directed insertional mutagenesis was used to construct a pullulanase I-minus mutant of B. thetaiotaomicron so that the physiological significance of the cloned pullulanase could be investigated. Subsequent examination of B. thetaiotaomicron 95-1 (the pullulanase I-minus mutant) revealed that although the cloned pullulanase comprised approximately 3 to 5% of the total pullulanase activity in B. thetaiotaomicron, this enzyme was not essential for growth of B. thetaiotaomicron on pullulan. In fact, the 95-1 mutant was able to grow on pullulan at a rate similar to that of wild-type B. thetaiotaomicron. Therefore, there must be a second pullulan-degrading enzyme in B. thetaiotaomicron. Four types of pullulan-hydrolyzing enzymes have been described: (i) a glucoamylase (EC ) (11, 19) which hydrolyzes pullulan from the nonreducing ends to produce glucose, (ii) a pullulanase (EC ) (1) from Klebsiella pneumoniae which hydrolyzes a(1-*6)-d-glucosidic linkages of puilulan to produce maltotriose, (iii) an isopullulanase (EC ) (14) from Aspergillus niger which hydrolyzes a(1-*4)-d-glucosidic linkages of pullulan to produce isopanose, and (iv) a neopullulanase from Bacillus stearothermophilus which hydrolyzes a(1-*4)-d-glucosidic linkages in pullulan to produce panose (7). The most common pullulanase is the Klebsiella pullulanase type, which hydrolyzes the a(1-*6)-glucosidic linkages of pullulan to produce maltotriose. The B. thetaiotaomicron pullulanase I was this type of pullulanase. In this report we describe the purification and * Corresponding author characterization of a second pullulanase (II), tentatively designated a neopullulanase, from B. thetaiotaomicron. In addition, we describe another enzyme, an a-glucosidase, which may be involved in pullulan and starch breakdown by B. thetaiotaomicron. MATERIALS AND METHODS Bacterial strains and media. B. thetaiotaomicron 5482A (ATCC 29148) was obtained from the culture collection of Virginia Polytechnic Institute Anaerobe Laboratory, Blacksburg. B. thetaiotaomicron 95-1, a derivative of B. thetaiotaomicron 5482A, has been described previously (18). It contains an insertion in the a(1--6)-glucosidic linkage-cleaving pullulanase I gene. B. thetaiotaomicron 5482A and 95-1 were grown in defined medium similar to the basal medium described by Varel and Bryant (2) except that.1 M potassium phosphate buffer (ph 7.2) was used in place of carbonate buffer. The final concentration of carbohydrate in the medium was.5%. The atmosphere was 8% nitrogen- 2% carbon dioxide. To B. thetaiotaomicron 95-1 minimal medium broth, erythromycin at 1,ug/ml was added. Pullulan was obtained from Sigma Chemical Co., St. Louis, Mo. Enzyme assays. (i) Pullulanase and amylase assays. Pullulanase and amylase activities were measured by determining the rate of increase in reducing sugar concentration when pullulan or amylose was incubated with enzyme at 37 C. A pullulan concentration of 4 mg/ml was found to be saturating. However, amylose is fairly insoluble, and a saturated solution contains approximately 2 mg of amylose per ml. The reaction mixture (2. ml) contained 1.8 ml of pullulan (4 mg/ml) or amylose (2 mg/ml) in 2 mm potassium phosphate buffer (ph 6.5) and.2 ml of the appropriately diluted enzyme. Samples of.25 ml were removed at intervals during the 3-min incubation and heated for 5 min at 1 C. The increase in reducing sugar was measured by the method of Dygert et al. (2). One unit of enzyme was defined as 1,ug of reducing sugar equivalent (glucose) produced per min in 2 mm potassium phosphate buffer (ph 6.5) at 37 C. For column fractions, the reaction mixture (.7 ml) contained

2 VOL. 173, 1991 n i Pullulan JG G-G-J Amylose [G-G-G-G-G-G] Maltotnose G-G-G Panose G Isomaltotriose G G-G Maltose G-G Isomaltose G G FIG. 1. Structures of carbohydrates used in this study. G, Glucose; n, the a(1-4) bond cleaved by the Bacillus stearothermophilus neopullulanase and the B. thetaiotaomicron pullulanase II; i, the a(l1-4) bond cleaved by the A. niger isopullulanase; p, the a(1-6) bond cleaved by the K. pneumoniae pullulanase and the B. thetaiotaomicron pullulanase I. Horizontal lines indicate at(1-4)-dglucosidic bonds; vertical lines indicate a(1-6)-d-glucosidic bonds..35 ml of pullulan (4 mg/ml) or amylose (2 mg/ml),.3 ml of 2 mm potassium phosphate buffer (ph 6.5), and.5 ml of each column fraction. Samples of.25 ml were drawn at and 3 min and heated at 1C. The reducing sugar concentration was measured as described above. (ii) a-glucosidase assay. a-glucosidase activity was measured by determining the rate of hydrolysis of p-nitrophenyla-d-glucopyranoside at 37 C. The reaction mixture (.5 ml) contained.4 ml of 2 mm potassium phosphate buffer (ph 6.5),.5 ml of 2 mmp-nitrophenyl-a-d-glucoside, and.5 ml of the appropriately diluted enzyme. The increase in A45 was measured with a Gilford 25 recording spectrophotometer. A unit of enzyme activity was defined as 1,umol of p-nitrophenol liberated per min in 2 mm potassium phosphate buffer (ph 6.5) at 37C. The extinction coefficient for p-nitrophenol under these conditions was 1.3 x 14 M-1 cm-l. Column fractions were assayed for a-glucosidase activity by mixing in the wells of a microtiter plate 75 ptl of 2 mm p-nitrophenyl-a-d-glucopyranoside in potassium phosphate buffer (ph 6.5) and 75 RId of the column fraction. The microtiter plate was incubated at 37 C for 1 to 4 h. Wells were inspected visually for yellow color formation. Purification of a soluble pullulanase from B. thetaiotaomicron B. thetaiotaomicron 95-1 was grown in 2 liters of defined medium which contained pullulan as the sole carbon source and erythromycin (1,ug/ml) to maintain the chromosomal insertion. The cells were harvested at an optical density (65 nm) of.7 to 1. by centrifugation (1, x g, 2 min, 4C), washed twice, and resuspended in 2 to 4 ml of potassium phosphate buffer (ph 7.5). Resuspended cells were disrupted by sonication four times for 3 to 4 min at 5 to 6% maximum tip output. Disrupted cells were centrifuged (15, x g, 2 min, 4 C) twice to remove cell debris, and the supernatant fluid (cell extract) was used in the purification scheme described below. (i) Step 1: ultracentrifugation. Cell extract was centrifuged at 2, x g for 12 h at 4 C, and the supernatant was decanted. The pellet from this centrifugation was resuspended in 2 ml of 2 mm potassium phosphate buffer (ph 7.5) and centrifuged again (2, x g, 2.5 h, 4 C). The first ultracentrifugation step pelleted most of the pullulanase activity with the membranes. The subsequent membrane wash contained the largest portion of pullulanase activity. This wash was applied to the DEAE-Sephacel column. G ENZYMES OF B. THETAIOTAOMICRON %3 (ii) Step 2: DEAE-Sephacel chromatography. The second soluble protein fraction (membrane wash) was applied to a DEAE-Sephacel column (1.5 by 12 cm) which had been equilibrated with 2 mm potassium phosphate buffer (ph 7.5). The column was washed with buffer until no eluting protein was detected by A28. The column was eluted with a linear to.15 M NaCl gradient, and 1.9-ml fractions were collected. The fractions were assayed for puilulanase, amylase, and a-glucosidase activity as well as NaCl concentration and protein (A28). Fractions that contained pullulanase activity (fractions 28 through 4) but no a-glucosidase activity were pooled. (iii) a-cyclodextrin-sepharose 6B affinity column chromatography. The pooled pullulanase activity from the previous step was concentrated in Amicon Centriprep 3 concentrators (Amicon, Beverly, Mass.) in a clinical centrifuge at 4 C. The Amicon Centriprep 3 was also used to desalt and equilibrate the pooled fractions to 2 mm potassium phosphate (ph 6.5). The concentrated activity was applied to an a-cyclodextrin-sepharose 6B column (1 by 7 cm) which had been equilibrated with 2 mm potassium phosphate buffer (ph 6.5). The a-cyclodextrin column was prepared according to the manufacturer's directions, using 3,umol of a-cyclodextrin (Sigma) per g of epoxy-activated Sepharose 6B (Pharmacia). The column was washed extensively with buffer until no eluting protein was detected (A28). The column was eluted with a linear gradient of to.15 M NaCl, and 1.7-ml fractions were collected. The fractions were assayed for pullulanase, amylase, and a-glucosidase activity as well as NaCl concentration and protein (A28). Fractions 26 through 36, which contained the highest levels of pullulanase activity, were pooled, concentrated, and desalted in Amicon Centriprep 3 concentrators. The enzyme preparation was stored in 2% glycerol at -2 C. A portion of the preparation was used to determine the enzyme's substrate specificity and end products. Partial purification of the B. thetaiotaomicron 95-1 a-glucosidase. (i) Steps 1 and 2. Ultracentrifugation and DEAE- Sephacel anion-exchange chromatography were performed as described above for the pullulanase. The DEAE-Sephacel fractions (fractions 21 through 27) containing oa-glucosidase activity but no pullulanase activity were pooled and concentrated in Amicon Centriprep 3 concentrators. (ii) Step 3: Mono Q ion-exchange chromatography. The a-glucosidase concentrate was dialyzed against 4 liters of Tris-HCl (ph 7.) overnight at 4 C and then applied to an FPLC Mono Q column (Pharmacia, Uppsala, Sweden) which had been equilibrated with 2 mm Tris-HCl (ph 7.). Under these conditions, the a-glucosidase activity remained on the column but could be eluted with a gradient of 2 to 7% 2 mm Tris-HCl-.2 M NaCl. The flow rate was 1 ml/min, and.5-ml fractions were collected. Fractions were assayed for enzyme activity and protein concentration. Fractions containing high levels ofa-glucosidase activity were pooled, concentrated in Amicon Centricon 3 microconcentrators, and stored in 2% glycerol at -2 C. A portion of the pooled fractions was incubated with the various carbohydrates to determine the enzyme's breakdown products and substrate specificity. Estimation of B. thetaiotaomicron 95-1 pullulanase native molecular weight. The native molecular weight of the soluble pullulanase of B. thetaiotaomicron 95-1 was estimated by gel filtration chromatography on a Bio-Rad AO.5M column. The column was equilibrated and eluted with.25 M NaCl-2 mm potassium phosphate buffer (ph 6.5). Molecular mass

3 2964 SMITH AND SALYERS standards were alcohol dehydrogenase (15 kda), phosphorylase b (97 kda), bovine serum albumin (68 kda), carbonic anhydrase (29 kda), and lysozyme (14 kda). Gel electrophoresis. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- PAGE) on 1% polyacrylamide gels as described by Laemmli (8). Proteins were visualized by staining with Coomassie brilliant blue of FASTSTAIN (Zoion Research Allston, Mass.). Molecular mass standards were myosin (H chain; 2 kda), phosphorylase b (97 kda), bovine serum albumin (68 kda), ovalbumin (43 kda), carbonic anhydrase (29 kda), and lysozyme (14 kda). For isoelectric focusing (IEF), tube gels (9 mm by 13 cm) were used. The IEF gels contained 5% acrylamide, 12% glycerol,.5% Triton X-1, and 2% ampholytes (Serva, Garden City Park, N.Y.) (for the ph range 4 to 9, the ratio of ph 3 to 1/pH 5 to 8/pH 7 to 9 was 3:1:1 [vol/vol/vol]; for the ph range 4 to 7, the ratio of ph 4 to 6/pH 3 to 7/pH 3 to 1 was 2:1:1 [vol/vol/vol]). Electrophoresis was done as described by Righetti and Drysdale (12). After electrophoresis, gels were cut into.5-cm slices. Gel slices to be assayed for enzyme activity were eluted in.5 ml of 2 mm potassium phosphate buffer (ph 6.5). For determination of isoelectric point, a ph gradient gel was prepared in an identical manner except that slices were eluted with.5 ml of deionized water and then the ph of each slice was determined. Triton X-1 (.5%) was included in the IEF tube gels used for electrophoresis of the Triton X-1 membrane extract (see below). Protein determination. In most cases, protein concentration was measured by the method of Lowry et al. (1). For the final step in enzyme purification, protein concentrations were measured by the method of Whitaker and Granum (21). Solubilizatlon of membrane-bound enzyme activities from B. thetaiotaomicron membranes. The membrane pellet, which has been washed twice with 2 mm potassium phosphate buffer (ph 7.5) and twice with 2 mm potassium phosphate containing.25 M NaCl (ph 7.5), was suspended in 3 ml of 2 mm potassium phosphate buffer (ph 7.5). Then 1 M KCl, Triton X-1, and 1 M potassium phosphate buffer (ph 7.5) were added to make a final concentration of.15 M KCl-.1 M KPO4-1.5% Triton X-1 (5). This solution was agitated gently for 3 h at 4 C and centrifuged at 2, x g for 2.5 h at 4 C. The resulting Triton X-1 extract was collected, and any remaining membranes were suspended in 2 mm potassium phosphate buffer (ph 7.5). The enzyme activities contained in the membrane extract were then separated on an IEF tube gel (ph range 4 to 7), and gel slice eluants were used to digest the various carbohydrates. Analysis of carbohydrate breakdown products. The partially purified pullulanase, a-glucosidase, and fractions from the IEF gel of the Triton X-1 membrane extract were used to digest various carbohydrates (2 mg/ml in 2 mm KPO4 [ph 6.5]) at 37 C. Carbohydrates tested were pullulan, maltoheptaose, maltohexaose, maltopentaose, maltotetraose, maltotriose, panose, isomaltotriose, maltose, and maltitol (Sigma) (see Fig. 1 for structures). A portion (2,ul) of each carbohydrate was taken prior to incubation at 37 C, and a second portion (2 to 25 R.l) was taken after digestion for approximately 12 to 18 h. Each portion was spotted on Whatman 3MM chromatography paper and resolved by descending paper chromatography. Chromatograms were developed for 2 to 3 h in a solvent system of ethyl acetate-glacial acetic acid-h2 (3:1:1 [vol/vol/vol]), and the carbohydrates were visualized by spraying the chromatogram with a p-anisidine-phthalic acid mixture (15). Pullulanase and amylase activities and specific TABLE 1. activities during the stages of purification Sp act Total activity (U) Fraction (U/mg of protein)a Puilulanase Amylase Pullulanase Amylase Cell extract ,285 13,3 Soluble protein ,958 4,15 DEAE-Sephacel 838 1,389 1,89 3,13 a-cyclodextrin column a One unit of activity is defined as 1,ug of reducing sugar equivalent (glucose) produced per min at 37 C. RESULTS J. BACTERIOL. Characteristics of the B. thetaiotaomicron 95-1 pullulanase. As reported previously, the pullulanase I insertional mutant B. thetaiotaomicron 95-1 was able to grow on pullulan and retained 5 to 7%o of the wild-type level of pullulanase activity. The remaining activity was presumably due to one or more pullulanases distinct from the enzyme characterized previously. To facilitate characterization and purification of the remaining enzymes, we used extracts from B. thetaiotaomicron 95-1 to eliminate interference from the previously characterized pullulanase I. Approximately 8% of the total pullulanase activity was either soluble or eluted from the membranes with 2 mm potassium phosphate buffer (ph 7.5). The other 2%o of pullulanase activity remained associated with the membranes. The soluble pullulanase activity was loaded onto a DEAE-Sephacel column and, after washing, eluted with a linear gradient of. to.15 M NaCl in potassium phosphate buffer (ph 7.5). A single peak of pullulanase activity eluted at approximately.15 M NaCl. Also, a single peak of a-glucosidase activity eluted at approximately.12 M NaCl. The nonoverlapping portions of the pullulanase (fractions 28 to 4) and a-glucosidase (fractions 21 to 27) peaks were pooled. Approximately 64% of the pullulanase activity in the soluble protein fraction was recovered from the DEAE- Sephacel column (Table 1). The next step in the purification utilized an a-cyclodextrin-sepharose 6B column. a-cyclodextrin is an inhibitor of some types of pullulan-cleaving enzymes, and the use of o-cyclodextrin affinity column chromatography in purification of a(1-4)-glucosidic linkage-hydrolyzing pullulanases has been reported previously (3). The B. thetaiotaomicron 95-1 pullulanase activity was retained by the column. Since the a-glucosidase was not retained by the column, this step removed any remaining a-glucosidase activity from the pullulanase preparation. Elution of the a-cyclodextrin-sepharose column with a linear gradient of. to.15 M NaCl yielded a single peak of pullulanase activity which was pooled in fractions 26 through 36. About 11 to 13% of the total pullulanase activity contained in the soluble ceil protein was recovered after the affinity chromatography step. The pooled and concentrated fractions from the oa-cyclodextrin- Sepharose 6B column were analyzed by SDS-PAGE, and the most purified preparation (Fig. 2, lane b) contained only one protein band. This preparation of the 95-1 pullulanase was used to digest various carbohydrates to determine its substrate specificity and end products (Table 2). The previously characterized B. thetaiotaomicron pullulanase I produced maltotriose from pullulan but did not hydrolyze amylose. Thus, pullulanase I appears to be specific for a(1-6)-glucosidic linkages. However, the pullula-

4 VOL. 173, 1991 ENZYMES OF B. THETAIOTAOMICRON a b a b c d e - 2kDa - 97kDa i - 68kDa - 43kDa FIG. 2. Ten percent SDS-PAGE gel of the most purified preparations of the a(1-+6)-pullulanase (I) and the a(1-4)-pullulanase (II or neopullulanase). Lanes: a, pullulanase I (3 jig) purified from Escherichia coli containing pks3-17, the pullulanase I clone; b, pullulanase II (27,ug) purified from B. thetaiotaomicron Arrowheads indicate positions of protein standards. nase examined in this study (pullulanase II) degraded amylose as well as pullulan (Table 1). Also, the product of pullulan digestion was panose rather than maltotriose (Fig. 3). Although maltotriose and panose had similar migration distances on paper chromatograms, maltotriose yielded a yellow-brown product while panose produced a green-brown product when the chromatogram was developed with the p-anisidine-phthalic acid reagent. Furthermore, to confirm that the pullulanase II end product was panose rather than isopanose, the pullulanase II pullulan hydrolysate was digested with A. niger amyloglucosidase and then analyzed by paper chromatography. When amyloglucosidase hydrolyzes panose, glucose is the only product, but isomaltose and glucose would be the products produced from isopanose. The only product detected was glucose. Thus, the pullulanase II end product was panose rather than isopanose. In addition, pullulanase II hydrolyzed maltotriose [a(1-*4)] to maltose and glucose, whereas it did not hydrolyze isomaltotriose [ot(1-*6)]. These characteristics identified the purified pullulanase II as an enzyme which cleaved the a(1-*4)- rather than the a(1->6)-d-glucosidic bonds. Pullulanase II was able to cleave,b-cyclodextrin (a cyclic molecule of TABLE 2. Comparison of end products of carbohydrate digestion for pullulanase I, pullulanase II, and the a-glucosidase Substrate End products of digestion with:a Pullulanase I Pullulanase II a-glucosidase Pullulan G3 G' Maltoheptaose - G, G2 (G-G5) G Maltohexaose - G, G2 (G-G5) G Maltopentaose - G, G2 (G-G4) G Maltotetraose - G, G2 (G-G3) G Maltotriose - G, G2 G Panose - - G Isomaltotriose - - G Maltose - - G Maltitol - - Degrades amylose No Yes No a G, Glucose; G2, maltose; G3, maltotriose; G3', panose; G4, maltotetraose; G5, maltopentaose; -, the carbohydrate was not hydrolyzed by the enzyme. Minor products of digestion are given in parentheses. FIG. 3. Paper chromatogram of pullulan digested with pullulanase II. Lanes: a, pullulan digested with a most purified preparation of pullulanase II; b, undigested pullulan; c, pullulan (undigested), panose, and glucose; d, pullulan (undigested), maltotriose, and glucose; e, glucose. maltoheptaose), suggesting that the enzyme cleaves the carbohydrate molecule endolytically. IEF gels of the purified preparation of the pullulanase II indicated that the enzyme's pi was approximately 5.4. In addition, gel filtration on a Bio-Rad AO.5M column estimated its native molecular mass at approximately 7 kda. The denatured molecular mass from SDS-PAGE is approximately 7 kda (Fig. 2, lane b). Pullulanase I had been determined previously (18) to have a pl of 5.6, a native molecular mass of approximately 77 kda, and a denatured molecular mass of approximately 8 kda (Fig. 2, lane a). Characterization of the B. thetaiotaomicron 95-1 a-glucosidase. The a-glucosidase activity was separated from the pullulanase activity at the DEAE-Sephacel anion-exchange chromatography step. The B. thetaiotaomicron 95-1 soluble protein fraction which was loaded on the DEAE-Sephacel column contained approximately 24 to 27% of the total a-glucosidase activity contained in the crude cell extract. Subsequent membrane washes continued to wash ot-glucosidase activity off the membranes. Approximately 7% of the total a-glucosidase activity in the crude cell extract was recovered when the activities of the soluble fractions were summed. Approximately 14 to 19% of the ca-glucosidase activity contained in the crude cell extract was recovered after the DEAE-Sephacel anion-exchange chromatography step. The a-glucosidase activity fraction contained many. AI

5 2966 SMITH AND SALYERS J. BACTERIOL. TABLE 3. Substrate End products of digestion with Triton X-1- solubilized membrane fractions End products of digestion with:a Peak pi 4.4 Peak pi 5.4 Peak pl 5.8 Pullulan G3' G3 Trace G Maltohexaose G, G2 G, G2 G Maltopentaose G, G2, G3 G, G2 G Maltotetraose G, G2, G3 G, G2 G Maltotriose G, G2 G, G2 G Panose - - ND Isomaltotriose - Trace G2' G Maltose - - ND Degrades amylose Yes Yes No a G, Glucose; G2, maltose; G2', isomaltose; G3, maltotriose; G3, panose; -, the enzyme did not degrade the substrate; ND, not determined. proteins after the DEAE-Sephacel step. Chromatography on a Mono Q column produced an a-glucosidase preparation containing only three proteins, as determined by SDS- PAGE: 8, 57, and 5 kda. The pl of the ot-glucosidase was 5.7 Ṫhis preparation of the a-glucosidase was used to determine the enzyme's substrate specificity and end products (Table 3). The enzyme had little or no activity against puilulan, amylose, and,-cyclodextrin but hydrolyzed G2 through G7 to glucose. Also the enzyme degraded both panose and isomaltotriose to glucose. Thus, it appeared that the ca-glucosidase is able to hydrolyze both a(1-*4) and a(l-+6) linkages. Characterization of enzyme activities from B. thetaiotaomicron 95-1 membranes. As reported previously (18), 16 to 18% 1.I * ~ Amylase c 1. ) IV.83 Cu.65.4 i[i11,.2 of the pullulanase activity of B. thetaiotaomicron 5482A remained associated with the membranes even after.25 M NaCl washes. This was also true for the 95-1 mutant. Approximately 5% of the activity which remained with the membranes was released by Triton X-1. Electrophoresis of this membrane extract on IEF tube gels (ph range 4 to 7) revealed three activity peaks (Fig. 4). Two activity peaks (pl 4.4 and 5.4) hydrolyzed both pullulan and amylose but did not hydrolyze p-nitrophenyl-ca-d-glucopyranoside. The third peak (pl 5.8) had activity against the p-nitrophenyl-a-dglucopyranoside but no activity against pullulan and little activity against amylose. DISCUSSION Our results demonstrate that B. thetaiotaomicron has at least two soluble pullulanases. Pullulanase I cleaves oa(1-36)- D-glucosidic bonds and produces maltotriose from pullulan. In contrast, pullulanase II cleaves ot(1-->4)-d-glucosidic linkages and produces panose from pullulan. Furthermore, pullulanase II, unlike pullulanase I, hydrolyzed amylose as well as pullulan. There was pullulanase activity which remained associated with the B. thetaiotaomicron 95-1 membranes after repeated salt washes and had to be solubilized with Triton X-1. This membrane activity appeared to be due to two different enzymes. One had a pi value and substrate specificity virtually identical to those of pullulanase II and could be a membrane form of pullulanase II. The other enzyme differed from pullulanases I and II with respect to pi but had a substrate specificity spectrum similar to that of pullulanase II. There have been several reported cases in Bacteroides species of enzymes which cleave the same substrate or bond in a substrate but are genetically distinct from each other (4, ph X N _ CD I ooa N X) X X N o N- It CD cn M ) (D N a _ cm c C N a) cu co It NU CU C') C'i o st t 't It ) Ie t U) It) It) It ) ItO ED 6161 I I I IIII I i,, i i Pullulanase Iaa-Glucosidase E LI GEL Slice FIG. 4. Pullulanase, amylase, and a-glucosidase activities in a Triton X-1 extract of B. thetaiotaomicron 95-1 membranes that had been analyzed on an IEF tube gel (ph range 4 to 7).

6 VOL. 173, ). Thus, it is possible that the pi 4.4 enzyme peak is a third pullulanase. Since both of the Bacteroides pullulanases degraded pullulan to trisaccharides but did not degrade the trisaccharides, B. thetaiotaomicron presumably needs an enzyme which hydrolyzes these products to glucose. The a-glucosidase appears to be such an enzyme. The most purified preparation cleaved small a(1-*4) oligosaccharides (G7 through G2) to glucose but did not hydrolyze amylose (<5%). Thus, this enzyme appeared to hydrolyze only short-chain oligosaccharides. The a-glucosidase also cleaved panose and isomaltotriose to glucose, indicating that the enzyme was able to hydrolyze both a(1-*4)- and a(1->6)-d-glucosidic bonds. The ability of the a-glucosidase to hydrolyze a(1-*6)-dglucosidic bonds may be important in the breakdown of pullulan, since pullulanase I does not cleave a(1-*6) linkages in small oligosaccharides such as panose. Moreover, the fact that the a-glucosidase can hydrolyze maltoheptaose but cannot hydrolyze 3-cyclodextrin suggests that the enzyme acts in an exo rather than endo manner to degrade oligosaccharides. Although most of the a-glucosidase activity is soluble, some activity remained with the membranes. The pl 5.8 peak of the Triton X-1 membrane extract had the same substrate specificity spectrum as the soluble a-glucosidase and a similar pl. Again, this may be a membrane-bound form of the soluble enzyme. B. thetaiotaomicron grows on amylose as well as on pullulan. Thus far, we have not found an amylase that hydrolyzes amylose but not pullulan. However, pullulanase II degrades both pullulan and amylose. Thus, despite the absence of a specific amylase, pullulanase I, pullulanase II, and the a-glucosidase acting in concert should be able to hydrolyze either pullulan or amylose to glucose. The pullulanase II examined in this study had properties similar to those of three previously characterized types of puilulan-degrading enzymes: an isopullulanase from A. niger, an a-amylase of Thermoactinomyces vulgaris (13, 17), and a neopullulanase of Bacillus stearothermophilus (7). The Bacteroides pullulanase II, Thermoactinomyces a-amylase, and Bacillus neopullulanase produce panose from pullulan hydrolysis. The Bacillus neopullulanase also produces maltose and glucose in addition to panose (panose/maltose/ glucose ratio of 3:1:1). Furthermore, both the Bacteroides pullulanase II and the Bacillus neopullulanase appeared to act endolytically, since they degrade,b-cyclodextrin. The Thermoactinomyces a-amylase was not tested for the ability to hydrolyze P-cyclodextrin. Both the Thermoactinomyces oa-amylase and the Bacteroides pullulanase II hydrolyzed amylose. The fact that the Thermoactinomyces oa-amylase can hydrolyze pullulan differentiates it from most a-amylases found in nature, since most ax-amylases are unable to hydrolyze pullulan. However, the activity of the Thermoactinomyces a-amylase against pullulan was approximately one-fourth of its activity against amylose, whereas the activity of the Bacteroides pullulanase II against amylose was approximately equal to its activity against pullulan. The Bacillus neopullulanase had very little activity against starch (i.e., only small amounts of glucose and maltose were detected on paper chromatograms after starch was digested for 24 h with the neopullulanase). The Bacteroides pullulanase II resembles the A. niger isopullulanase in that it hydrolyzed only a(1-4)-d-glucosidic linkages but differs in that isopullulanase produced isopanose from pullulan rather than panose. The preference for only a(1-*4) linkages differentiates the B. thetaiotaomicron pullulanase II from the Bacillus neopullulanase and the ENZYMES OF B. THETAIOTAOMICRON Thermoactinomyces a-amylase, which hydrolyzed both a(1->4)- and a(1-+6)-d-glucosidic bonds. Thus, the Bacteroides pullulanase II appears not to be identical to any of the four types of previously reported pullulan-degrading enzymes, although it shares some properties with two types. Since the Bacteroides pullulanase II produces panose from pullulan, we have tentatively designated it a neopullulanase. ACKNOWLEDGMENTS This work was supported by Public Health Service grant RO1 Al from the National Institute of Allergy and Infectious Diseases. Karen Smith was supported by Cell and Molecular Biology training grant GM 7283 from the National Institutes of Health. REFERENCES 1. Bender, H., and K. Wallenfels Pullulanase (an amylopectin and glycogen debranching enzyme) from Aerobacter aerogenes. Methods Enzymol. 8: Dygert, S., L. H. Li, D. Florida, and J. A. Thoma Determination of reducing sugar with improved precision. Anal. Biochem. 13: Enevoldsen, B. S., L. Reimann, and N. L. Hansen Biospecific affinity chromatography of pullulanase. FEBS Lett. 79: Gherardini, F. C., M. Babcock, and A. A. Salyers Purification and characterization of two a-galactosidases associated with catabolism of guar gum and other a-galactosides by Bacteroides ovatus. J. Bacteriol. 161: Gherardini, F. C., and A. A. Salyers Characterization of an outer membrane mannanase from Bacteroides ovatus. J. Bacteriol. 169: Holdemann, L. V., I. J. Good, and W. E. Moore Human fecal flora variation in bacterial composition within individuals and possible side effects of emotional stress. Appl. Environ. Microbiol. 31: Kuriki, T., S. Okada, and T. Imanaka New type of pullulanase from Bacillus stearothermophilus and molecular cloning and expression of the gene in Bacillus subtilis. J. Bacteriol. 17: Laemmli, U. K Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227: Linn, S., T. Chan, L. Lipeski, and A. A. Salyers Isolation and characterization of two chondroitin lyases from Bacteroides thetaiotaomicron. J. Bacteriol. 156: Lowry,. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: Pazur, J. H., and T. Ando The hydrolysis of glycosyl oligosaccharides with a(1-+4) and a(1->6) bonds by fungal amyloglucosidase. J. Biol. Chem. 235: Righetti, P. G., and J. W. Drysdale. Isoelectric focusing, p In T. S. Work and E. Work (ed.), Laboratory techniques in biochemistry and molecular biology, vol. 5. Elsevier/ North-Holland Publishing Co., New York. 13. Sakano, Y., S. Hiraiwa, J. Fukushima, and T. Kobayashi Enzymatic properties and action of Thermoactinomyces vulgaris a-amylase. Agric. Biol. Chem. 46: Sakano, Y., N. Masuda, and T. Kobayashi Hydrolysis of pullulan by a novel enzyme from Aspergillus niger. Agric. Biol. Chem. 35: Salyers, A. A., and M. O'Brien Cellular location of enzymes involved in chondroitin sulfate breakdown by Bacteroides thetaiotaomicron. J. Bacteriol. 143: Salyers, A. A., J. R. Verceeloti, S. E. West, and T. D. Wilkins Fermentation of mucin and plant polysaccharides by strains of Bacteroides from the human colon. Appl. Environ. Microbiol. 33: Shimizu, M., M. Kanno, M. Tamura, and M. Suekane Purification and some properties of a novel a-amylase produced by a strain of Thermoactinomyces vulgaris. Agric. Biol. Chem.

7 2968 SMITH AND SALYERS 42: Smith, K. A., and A. A. Salyers Cell-associated pullulanase from Bacteroides thetaiotaomicron: cloning, characterization, and insertional mutagenesis to determine role in pullulan utilization. J. Bacteriol. 171: Ueda, S., K. Fujita, K. Komatsu, and N. Nakashima Polysaccharide produced by the genus Pullularia. I. Production J. BACTERIOL. of polysaccharide by growing cells. Appl. Microbiol. 11: Varel, V. H., and M. P. Bryant Nutritional features of Bacteroides fragilis subsp. fragilis. Appl. Microbiol. 28: Whitaker, J. R., and P. E. Granum An absolute method for protein determination based on difference in absorbance at 235 and 28 nm. Anal. Biochem. 19: