Construction of Chimeric Enzymes out of Maize Endosperm Branching Enzymes I and II:

Branching enzyme I and II isoforms from maize endosperm (mBE I and mBE II, respectively) have quite different properties, and to elucidate the domain(s) that determines the differences, chimeric genes consisting of part mBE I and part mBE II were constructed. When expressed under the control of the T7 promoter in Escherichia coli, several of the chimeric enzymes were inactive. The only fully active chimeric enzyme was mBE II-IBspHI, in which the carboxyl-terminal part of mBE II was exchanged for that of mBE I at a BspHI restriction site and was purified to homogeneity and characterized. Another chimeric enzyme, mBE I-II HindIII, in which the amino-terminal end of mBE II was replaced with that of mBE I, had very little activity and was only partially characterized. The purified mBE II-I BspHI exhibited higher activity than wild-type mBE I and mBE II when assayed by the phosphorylase a stimulation assay. mBE II-IBspHI had substrate specificity (preference for amylose rather than amylopectin) and catalytic capacity similar to mBE I, despite the fact that only the carboxyl terminus was from mBE I, suggesting that the carboxyl terminus may be involved in determining substrate specificity and catalytic capacity. In chain transfer experiments, mBE II-I BspHI transferred more short chains (with a degree of polymerization of around 6) in a fashion similar to mBE II. In contrast, mBE I-II HindIII transferred more long chains (with a degree of polymerization of around 11–12), similar to mBE I, suggesting that the amino terminus of mBEs may play a role in the size of oligosaccharide chain transferred. This study challenges the notion that the catalytic centers for branching enzymes are exclusively located in the central portion of the enzyme; it suggests instead that the amino and carboxyl termini may also be involved in determining substrate preference, catalytic capacity, and chain length transfer.

enzyme plays an important role in starch synthesis (1)(2)(3). Multiple forms of BE have been identified in many plants, including maize endosperm (4), pea seed (5), and rice endosperm (6 -7). The cDNAs encoding the genes for the BEs have been cloned from various sources, such as maize endosperm (8,9), pea seed (10), potato tuber (11,12), and rice endosperm (6,13). The cDNAs encoding mature mBE I and mBE II have been expressed in Escherichia coli using the T7 promoter (14,15), allowing the study of the structure-function relationships of mBEs using site-directed mutagenesis and the construction of chimeric enzymes of mBE I and mBE II.
Homology in the primary structures between glycogen branching enzyme and amylolytic enzymes was first reported by Romeo et al. (16). Subsequently, Baba et al. (8) established that BEs contain the four highly conserved regions in the central portion of the enzyme that are present in ␣-amylases, pullulanase, isoamylase, and cyclodextrin glucanotransferases. Neopullulanase catalyzes the hydrolysis of ␣-1,4and ␣-1,6glucosidic linkages, as well as transglycosylation to form ␣-1,4and ␣-1,6-glucosidic linkages (17,18). The introduction of several replacements of the amino residues that constitute the active center of neopullulanase indicated that one active center of the enzyme participated in all four reactions described above (19). This suggested that not only were the structures of BE, ␣-amylase, pullulanase/isoamylase, and cyclodextrin glucanotransferase similar but also that they shared common catalytic mechanisms (18,20). Based on these results, the ␣-amylase enzyme family was defined; it includes BE and other enzymes that catalyze hydrolysis and transglycosylation at ␣-1,4and ␣-1,6-glucosidic linkages (21)(22)(23)(24)(25). Structure prediction and hydrophobic cluster analysis of the enzymes mentioned above indicated that they possess a catalytic (␤/␣)8-barrel (23,26) like that seen in crystal structure of ␣-amylases (27)(28)(29) and cyclodextrin glucanotransferases (30 -31). It is currently considered that four of the ␤-strands (i.e. the four highly conserved regions in the central portion of the enzymes) make up the catalytic center (23).
Some of the amino acids found to be essential for catalytic activity of the ␣-amylase family enzymes and also present in the four conserved regions of the maize endosperm branching enzymes were mutated via site-directed mutagenesis in mBE II and were found to be important for branching enzyme activity (32,33). Thus, the conserved four active site regions found in the ␣-amylase family enzymes also play a role in branching enzyme catalysis.
The amino acid sequence of mBE I and mBE II has an identity of 58% (Fig. 1). The identity is higher (67%) in the center portion of the enzymes, which contains several highly conserved regions (regions 1-4), as previously mentioned (8,15,21,23). When amino acid residues with similar functional side chains are taken into consideration, the two enzymes are 75% similar for the total amino acid sequence and 94% similar for the center portion. In contrast, the amino acid sequences and the numbers of amino acid residues are quite different for the amino-terminal and carboxyl-terminal sides of the center portion ( Figs. 1 and 2).
It is therefore interesting to focus attention on the relationship between the enzyme specificities and the different structures in the amino terminus and carboxyl terminus. In this paper, we report on the construction and characterization of chimeric enzymes of mBE I and mBE II.

MATERIALS AND METHODS
Media-LB medium (pH 7), consisting of 10 g of tryptone, 5 g of yeast extract, and 10 g of NaCl per liter, was used for culture of E. coli. 2 ϫ YT broth (pH 7), consisting of 16 g of tryptone, 10 g of yeast extract, and 5 g of NaCl per liter, was used for the preparation of phage DNA. Ampicillin was used at final concentration of 100 g/ml.
Creation of New Restriction Endonuclease Sites-Site-directed mutagenesis (Sculptor in vitro mutagenesis system, Amersham Corp., Amersham, UK) was used for the creation of a new restriction site by the introduction of a silent mutations in the mBE I and II genes. Oligonucleotides were synthesized in an Applied Biosystems model 380A DNA synthesizer (Macro-molecular Facility, Department of Biochemistry, Michigan State University, MI). The mutation was confirmed by restriction endonuclease digestion and DNA sequencing. DNA sequencing was done by the dideoxy chain-terminating method (35); the sequence reaction was started from the M13 linker region with the universal primer or primed by internal annealing of a 17-mer synthetic oligonucleotide.
Construction of mBE II-I BspHI, a Chimeric Gene of mBE I and II-To introduce a BspHI site into the mBE I gene by site-directed mutagenesis (silent mutation), the 1156-bp EcoRI-HindIII fragment from pET-23D-mBE I was cloned into the EcoRI-HindIII sites of phage M13mp19 multiple-cloning sites. Single-stranded DNA was prepared from the phage and used as template for the site-directed mutagenesis reaction. Sequence analysis of the 1156-bp EcoRI-HindIII fragment verified that the desired nucleotide had been changed and that no second-site mutations were present. After preparation of the replicative form of the phage DNA, the 1156-bp EcoRI-HindIII fragment of pET-23D-mBE I was exchanged for the mutant fragment, resulting in the construction of pET-23D-mBE I (BspHI). The following additional steps were used to avoid methylation of adenine residues by DNA adenine methylase in the recognition sequence 5Ј-G*ATC-3Ј (asterisk indicates methylated form) when the BspHI site was used for the construction of the chimeric gene. pET-23D-mBE I (BspHI) was constructed using E. coli TG-1 (rk Ϫ mk Ϫ ) as a host. E. coli DH5 (rk Ϫ mk ϩ ) was transformed with pET-23D-mBE I (BspHI) for plasmid preparation from E. coli DH5. The plasmids obtained were then used for transformation of E. coli JM110 (dam rk ϩ mk ϩ ). Because BspHI sites on the plasmid DNAs prepared from E. coli JM110 were not affected by DNA adenine methylase, pET-23D-mBE I (BspHI) prepared from E. coli JM110 was used for construction of the chimeric gene at the BspHI site.
For the construction of the mBE II-I BspHI gene, the 687-base pair BspHI-carboxyl-terminal fragment of pET-23D-mBE II was exchanged for the 852-base pair BspHI-carboxyl-terminal fragment of pET-23D-mBE I (BspHI). Thus, pET-23D-mBE II-I BspHI, encoding the mBE II-I BspHI gene, was constructed (Fig. 2). The structure of the plasmid was confirmed by restriction endonuclease mapping.
Several other chimeric enzymes, mBE I-II NcoI, mBE II-I NcoI, mBE I-II HindIII, mBE II-I HindIII, mBE I-II BspHI, mBE I-II-I, and mBE II-I-II, were also constructed ( Fig. 2). To construct these other chimeric enzymes, two silent mutations were introduced by site-directed mutagenesis to produce HindIII and NcoI endonuclease restriction sites in mBE I and II, respectively. The procedures used were similar to those described previously.
Preparation and Purification of Wild-type mBE I, mBE II, and Chimeric BEs-E. coli BL21(DE3) carrying a recombinant plasmid encod- ing the gene for wild-type BE I, wild-type BE II, or chimeric BEs was grown overnight in LB medium containing 100 g/ml ampicillin. The preculture was then diluted 1:20 (v/v) in fresh LB medium containing 100 g/ml ampicillin, and the cells grown at 37°C to mid-log phase (A 600 nm ϭ 0.6). At this point, expression of the BE gene was induced by addition of isopropyl-1-thio-␤-D-galactopyranoside (final concentration, 0.5 mM), and were cultures transferred to 25°C for 12 h. Cells were then harvested by centrifugation (10,000 ϫ g for 10 min), and the pellet was resuspended and lysed by sonication in 50 mM Tris acetate buffer (pH 7.5) containing 10 mM EDTA and 5 mM dithiothreitol. The lysed suspension was then centrifuged at 30,000 ϫ g for 15 min, and the resulting supernatant (cell extract) was used as crude enzyme for preliminary assay of BEs. Purification of wild-type BEs and the chimeric enzymes mBE II-I BspHI and mBE I-II HindIII was carried out according to Guan et al. (14,15), with the exception that mBE I-II HindIII was expressed in AC71 (glgB Ϫ ), a glycogen BE-deficient strain of E. coli (3).
Assay of BE Activity-BE activity was measured by three different assays as described by Guan and Preiss (36).
Assay a-The phosphorylase a stimulation assay is based on the stimulation by BE of the synthesis of ␣-D-glucan from ␣-D-Glc-1-P catalyzed by rabbit phosphorylase a (37). Reaction mixtures contained, in a final volume of 200 l, 100 mM citrate (pH 7), 10 mM AMP, 0.4 mg phosphorylase a, and 50 mM D-[ 14 C]-Glc-1-P (50 dpm⅐nmol Ϫ1 ). The reaction was initiated by the addition of an appropriate amount of enzyme. One unit of enzyme activity is defined as 1 mol of Glc incorporated into ␣-D-glucan per min at 30°C.
Assay b-The branching linkage assay determines the number of branching linkages introduced by BE into the substrate, reduced amylose (38). Substrates were prepared by the reduction of enzymatically synthesized amylose (AS-320, AS-110, and AS-70, Nakano Vinegar Co., Aichi, Japan; these amylose varieties had average degrees of polymerization of 1815, 722, and 438, respectively) as described by Takeda et al. (38). Reaction mixtures contained, in a final volume of 100 l, 25 mM 4-morpholinepropanesulfonic acid (pH 7.5) and an appropriate amount of substrate. The reaction was initiated by the addition of an appropriate amount of enzyme. One unit of enzyme activity is defined as 1 mol of branching linkages formed per min at 30°C.
Assay c-The iodine stain assay is based on monitoring the decrease in absorbance of the glucan-iodine complex resulting from the branching of the substrate, amylose or amylopectin (potato type III and corn, respectively; Sigma) (4, 36). Reaction mixtures contained, in a final volume of 200 l, 50 mM citrate (pH 7) and 0.1 mg of substrate. The reaction was initiated by the addition of an appropriate amount of enzyme. One unit of enzyme activity is defined as the decrease in absorbance of 1.0 per min at 30°C.
Protein Assay-Protein concentration was measured with the BCA protein assay reagent (39), using bovine serum albumin as the standard.

Preparation of Debranched ␣-Glucan for Chain Length Distribution Analysis by High Performance Anion Exchange
Chromatography-Reduced amylose (1 mg of AS-320; degree of polymerization, 1815) was incubated in 25 mM 4-morpholinepropanesulfonic acid, pH 7.5 (200 l) at 30°C with BE (1.5 milliunits of mBE I, mBE II, mBE II-I BspHI, and mBE I-II HindIII by assay b). After 80, 160, and 320 min, the reaction was terminated by heating in a boiling water bath for 2 min, and M acetate buffer (pH 3.5, 20 l) and isoamylase (5 l of 590 units⅐ml Ϫ1 ) were added after the solution had cooled to room temperature. After 1 h of incubation at 45°C, the solution was heated in a boiling water bath for 2 min. High performance anion exchange chromatography was performed with a Dionex BioLC system as described previously (3).
Other Procedures-Plasmid or M13 replicative form DNA was prepared by either the rapid alkaline extraction method (42) or QIAGEN Plasmid Maxi Kit (QIAGEN Inc., Chatsworth, CA). Treatment of DNA with restriction enzymes and ligation of DNA were done as recommended by the manufacturer. QIAquick Gel Extraction Kit (QIAGEN Inc.) was used for recovery of DNA from agarose. Transformation of E. coli with plasmid DNA and M13 single-stranded template DNA preparation were done as described elsewhere (42).

FIG. 2. Schematic diagram representing wild-type mBE I and mBE II, indicating amino and carboxyl termini and the four conserved regions (R1-R4) in the central portion and the chimeric enzymes constructed from maize endosperm BE I and II.
The HindIII, NcoI, and BspHI restriction sites used in the construction of mBE II-I BspHI are shown; the restriction sites introduced to mBE I by site-directed mutagenesis are shown in parentheses. The chimeric enzymes were constructed from mBE I and mBE II as described in the text. Molecular weights are shown in parentheses. The portion of the chimeric enzyme from the amino terminus and/or carboxyl terminus of mBE I is shown in white, and that from mBE II is shown in black. The central portion from mBE I has light shading and that from mBE II has dark shading.

RESULTS AND DISCUSSION
Activity of Branching Enzymes-Branching enzyme activity for wild-type mBE I and II and chimeric enzymes (Table I) and mBE I and II with silent mutations (BspHI and HindIII for mBE I and NcoI for mBE II) was determined in E. coli BL21(DE3) cell extracts. The silent mutations did not affect enzyme activity (results not shown). mBE II-I BspHI (17 units⅐mg Ϫ1 of protein) exhibited higher activity than wild-type mBE I (5.2 units⅐mg Ϫ1 of protein) and mBE II (11 units⅐mg Ϫ1 of protein) ( Table I). The remaining chimeric enzymes were inactive, with the exception of mBE I-II HindIII and mBE II-I HindIII, which had a very small amount of activity (0.7 and 0.6 units⅐mg Ϫ1 protein, respectively) ( Table I). To avoid destruction of the secondary structures of the chimeric enzymes, we chose all three endonuclease restriction sites (HindIII, NcoI, and BspHI) that were used for the construction of the chimeric enzyme genes, located on the highly homologous regions on the primary structures of mBE I and II (Fig. 1). Furthermore, those sites were not in the areas forming the probable secondary structures on the ␣-helix or ␤-strand. It is therefore likely that despite the sequence similarities of mBE I and II, with the exception of mBE II-I BspHI, the chimeric enzymes did not fold in a way that produced wholly functional three-dimensional arrangements.
Cell extracts were analyzed by SDS-PAGE followed by immunoblotting. The presence of immunoreactive proteins with molecular weights the same as those calculated from the amino acid sequences indicated the expression of mBE I and II and chimeric enzymes. An immunoblot of purified mBE I, mBE II, and mBE II-I BspHI is shown in Fig. 3.  Table  III). Three bands of approximately equal intensity were observed when the sample was subject to SDS-PAGE (not shown). One of the protein bands had a molecular weight and immunoreactivity consistent with mBE I-II HindIII. Due to the low activity mBE I-II HindIII, a thorough characterization could not be carried out, although useful information was gained on chain length transfer.

Purification and Characterization of Chimeric Enzymes mBE II-I BspHI and mBE I-II
The optimum temperature for catalysis by mBE II-I BspHI was 25°C, intermediate between the optimum temperatures of mBE I (30°C) and mBE II (20°C). Although the change of the carboxyl terminus had some effect on the characteristics of the enzyme, there does not appear to be a specific relationship between the carboxyl terminus of mBEs and the optimum temperature for catalysis.
The specific activity of mBE II-I BspHI measured using assay a, the phosphorylase a stimulation assay, was 3880 units⅐mg Ϫ1 of protein, which is 3-4 times higher than wild-type mBE I (1196 units⅐mg Ϫ1 of protein) and mBE II (1017 units⅐mg Ϫ1 of protein) (Ref. 36; Table IV). The partially purified mBE I-II HindIII extract had a specific activity of 3.3 units⅐mg Ϫ1 of protein, over 2 orders of magnitude lower than wild-type mBE. Using assay b, the branching linkage assay, the specific activity of mBE II-I BspHI with different reduced amyloses as substrates (AS-320, AS-110, and AS-70) at a concentration of 100 M was 1.3, 0.89, and 0.48 units⅐mg Ϫ1 of protein, respectively (Table IV). This value is similar to that of mBE I, at 2.1, 1.3, and 0.32 units⅐mg Ϫ1 of protein, respectively, but approximately 1 order of magnitude higher than that obtained from mBE II (0.4, 0.2, and 0.028 units⅐mg Ϫ1 of protein, respectively; Table IV). These results confirm that mBE I has a  higher rate in branching amylose than mBE II (36), and indicate that mBE II-I BspHI has a substrate specificity similar to mBE I for amylose. The mBE I-II HindIII chimeric enzyme preparation had insufficient activity for use in assay b. Using assay c, the iodine stain assay, mBE I and mBE II-I BspHI had higher activity with amylose (90 and 69 units⅐mg Ϫ1 of protein, respectively) than with amylopectin (2.3 units⅐mg Ϫ1 of protein), giving a high amylose to amylopectin activity ratio (Table IV). In contrast, mBE II was more active with amylopectin (97 units⅐mg Ϫ1 of protein) than with amylose (6.4 units⅐mg Ϫ1 of protein), giving a lower amylose to amylopectin activity ratio (Table IV). There was insufficient mBE I-II HindIII activity to use for assay c. The ratios of mBE I and mBE II activity with amylose as a substrate to that with amylopectin as a substrate are similar to those previously reported (36). This result provides further evidence that mBE I and mBE II-I BspHI have similar substrate specificity for amylose, as well as for amylopectin. Because the carboxyl terminus of mBE II-I BspHI is from mBE I, it is possible that in addition to the four highly conserved regions in the central portion of mBEs, the carboxyl terminus may also be involved in substrate binding. Using assay b with unbranched amylose (AS-320) as a substrate, the K m values obtained for mBE I, II, and II-I BspHI were quite similar, although the V max for mBE I and II-I BspHI (3.3 Ϯ 0.4 and 2.7 Ϯ 0.3 units⅐mg Ϫ1 of protein, respectively) were higher than the value obtained for mBE II (0.62 Ϯ 0.02 units⅐mg Ϫ1 of protein; Table V). This indicates that the catalytic capacity of mBE II-I BspHI is similar to that of mBE I, suggesting that the carboxyl terminus of BEs may also play a role in determining the catalytic efficiency of the enzymes.
Chain length distribution of debranched products was analyzed by high performance anion exchange chromatography following the incubation of mBE I, mBE II, mBE II-I BspHI, and mBE I-II HindIII with AS-320 amylose for 80, 160, and 320 min as described previously. The distribution of side chains transferred by mBE I and mBE II varied markedly (38). Initially (80 min), mBE I transferred chains with a broad distribution of lengths up to 46 glucose units (under the conditions employed), and at later times (160 and 320 min), mBE I began to transfer more shorter chains, with a high frequency of chains of 11-12 glucose units in length (Fig. 4). In vitro, mBE I appears to preferentially transfer longer chains, but eventually makes short branch points (38). In contrast, mBE II initially transferred short chains, with a high frequency of chains con- FIG. 4. High performance anion exchange chromatography analysis of the debranched ␣-glucan. Analysis was carried out as described in the text following incubation of reduced amylose (AS-320) with mBE I, mBE II, or mBE II-I BspHI for 80, 160, and 320 min at 30°C, with a subsequent incubation with iso-amylase for 1 h at 45°C.  (Fig. 4). The transfer of chains of 36 glucose units or longer was not detected under the conditions employed. For mBE II the ratio of short to long chains transferred did not change markedly over the time monitored (Fig. 4). mBE II-I BspHI transferred chains in a fashion similar to mBE II, since very few chains over 30 glucose units were not detected, and there was no shift over time toward transferring shorter chains. Six glucose units was the most common length transferred, although the percentage of short chains was not as high as for mBE II (Fig. 4); more mid-sized (degree of polymerization, 10 -15) chains were transferred. mBE I-II HindIII, which consists of the amino terminus of mBE I and the central portion and carboxyl terminus of mBE II, transferred chains in a similar manner to mBE I: few short chains were transferred initially, but with time, many more chains with a degree of polymerization of 11-12 were transferred (Fig. 4). Thus, the amino termini of mBEs appear to influence the length of the ␣-D-glucan chains transferred. The construction and characterization of a chimeric enzyme, mBE II-I BspHI, which consisted of the carboxyl terminus of mBE I and the amino terminus and central portion of mBE II (Fig. 2), which had substrate specificity and catalytic capacity similar to mBE I, indicates that the carboxyl terminus of mBEs may be involved to some extent in substrate binding and catalysis. mBE II-I BspHI transferred chains in a fashion similar to mBE II, suggesting that the central portion and/or amino terminus is involved in chain transfer. In the limited studies carried out with the low activity chimeric enzyme, mBE I-II HindIII, which consisted of the amino terminus of mBE I and the central portion and carboxyl terminus of mBE II (Fig. 2), chain transfer was similar to that of mBE I. Because the only portion of the mBE I-II HindIII from mBE I was the amino terminus, this would suggest that the amino terminus of mBEs is involved in chain transfer. This is also consistent with the chain transfer specificity of mBE II-I BspHI. These findings challenge previous reports that catalytic sites are conserved in the amylolytic family of enzymes (27,28,43), and in BEs these sites are located in regions 1-4 in the central portion of the enzyme (22). Because the putative catalytic centers for mBE isoforms are identical, the differing substrate specificity, catalytic capacity, and chain transfer of the isoforms may be due to either conformational changes in the central region caused by the carboxyl-terminal and amino-terminal regions or that one or more amino acid residues essential for activity are present in the terminal region(s) of mBEs.