- Email: [email protected]
Characterization of the branching patterns of glycogen branching enzyme truncated on the N-terminus
ABB Archives of Biochemistry and Biophysics 418 (2003) 34–38 www.elsevier.com/locate/yabbi
Characterization of the branching patterns of glycogen branching enzyme truncated on the N-terminus Claire H. Devillers, Mary E. Piper, Miguel A. Ballicora, and Jack Preiss* Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48824, USA Received 30 May 2003, and in revised form 7 July 2003
Abstract Truncation of 112 amino acids at the N-terminus (Nd1–112 ) changes the chain transfer pattern of the Escherichia coli glycogen branching enzyme (GBE) [Arch. Biochem. Biophys. 397 (2002) 279]. We investigated further the role of the N-terminus by engineering other truncated GBEs and analyzing the branching pattern by high-performance anion-exchange chromatography. The wild type GBE transfers mainly chains with a degree of polymerization (d.p.) of 8–14, the Nd1–112 enzyme transfers a greater proportion of chains with higher d.p. 15–20, whereas the 63- and 83-amino acid deleted enzymes had an intermediate pattern of transferred chains (d.p. 10–20). These data showed that a progressive shortening of the N-terminus leads to a gradual increase in the length of the transferred chains, suggesting that the N-terminus provides a support for the glucan substrate during the processes of cleavage and transfer of the a-(1–4) glucan chains. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Branching enzyme; Glycogen synthesis; Starch; Branching pattern; a-Glucan transfer
The biosynthesis of starch in plants and glycogen in bacteria involves at least three enzymes: ADP-glucose pyrophosphorylase, starch/glycogen synthase, and branching enzyme (BE)1 [1,2]. ADP-glucose pyrophosphorylase catalyzes the regulatory step of biosynthesis, whereas starch/glycogen synthase and branching enzyme are more involved in forming the structure of the starch or glycogen [3]. Branching enzyme belongs to the a-amylase family 13 containing a central ða=bÞ8 -barrel with four conserved regions that form the catalytic site of the enzyme [4]. BE catalyzes the cleavage of an a-(1– 4) glucosidic linkage and transfers the cleaved oligosaccharide to form a new a-(1–6) branch [5]. Several BEs have been identified in bacteria [2] and plants [6–8]. In maize, spinach or rice, different isoforms encoded by different genes have also been characterized. Those *
Corresponding author. Fax: 1-517-353-9334. E-mail address: [email protected] (J. Preiss). 1 Abbreviations used: BE, branching enzyme; d.p., degree of polymerization; GBE, glycogen branching enzyme; MBEI, maize branching enzyme I; SDS, sodium dodecyl sulphate; PAGE, polyacrylamide gel electrophoresis; HPAEC, high-performance anionexchange chromatography. 0003-9861/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0003-9861(03)00341-2
isoforms present different branching patterns and substrate specificities, which probably lead to different but complementary roles in starch synthesis [6–8]. For instance, the isoform maize branching enzyme I (MBEI) transfers long chains (degree of polymerization [d.p.] 40–100) and amylose is its specific substrate, whereas MBEII transfers shorter chains (d.p. 6–14) and prefers amylopectin as a substrate [9]. It is possible that MBEI initiates branching of lesser-branched polymers similar to amylose and then MBEII uses this product for further branching to form the final end product, amylopectin [9]. Different chimeric and truncated BEs have been studied to determine the role of the N- and C-terminal domains of these enzymes. An active hybrid of the Nterminal domain and ða=bÞ8 -barrel from MBEII and the C-terminal domain from MBEI has been constructed and characterized [10]. Its substrate specificity was similar to that of MBEI and its branching pattern to that of MBEII. The reverse construct exhibited a branching pattern similar to that of MBEI showing that the N-terminus is responsible for the branching pattern. Partial digestion of the Escherichia coli glycogen branching enzyme (GBE) with proteases and analysis of
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
the kinetic parameters of the products were performed to study the role of different domains [11]. A proteolyzed enzyme with 112 residues removed from the N-terminus was 27–59% as active as the wild type enzyme (measured by different methods). A 112 N-terminal deletion was later obtained by molecular biology techniques and led to an alteration of the branching pattern without affecting the substrate specificity [11]. The truncated enzyme transferred a greater amount of longer chains than the wild type enzyme [12]. The 3-D structure of the 112-amino acid truncated GBE was determined by X-ray crystallography [13,14]. That structure comprised three domains, a 128-residue N-terminal domain (lacking the 112 amino acids) arranged in seven b-strands, a 372-residue central domain forming the ða=bÞ8 -barrel common to a-amylase family, and a 116-residue C-terminal domain. Although the central domain is common to all enzymes belonging to this family, only the isoamylases share with GBE the structure of all three domains. Unfortunately, the wild type GBE has not been crystallized yet. To understand better the role of the N-terminal domain, we investigated another set of truncated GBEs. We engineered by recombinant technology truncations of 63 and 83 amino acids. After partial purification of those enzymes, we analyzed their branching patterns by high-performance anion-exchange chromatography (HPAEC) and compared them to those of the previous 112-amino acid truncated and the wild type enzymes [12].
Materials and methods Construction of the truncated genes pEXSB is a derivative of pET-23d and carries resistance to kanamycin, a glgB ribosome-binding site, and the full-length E. coli glgB gene [15]. pEXSB was purified from Epicurian XL2 blue cells (Stratagene, La Jolla, CA) and used as a template for PCR using the following primers 83-residue deletion: upper primer 50 CGC TAT CAT ATG GCT GTT GTC TGG CAT G 30 lower primer 50 GCT AGT TAT GCT CAG CGG 30 63-residue deletion: upper primer 50 GAG TGT CAT ATG TCA CGG GGA TTC TTT A 30 lower primer 50 GCT AGT TAT GCT CAG CGG 30 The two upper primers engineered an NdeI restriction site needed for subcloning and the lower primers anneal to the T7 terminator present in pEXSB. After purification, the PCR products were digested by NdeI and SalI (New England Biolabs, Beverly, MA) and subcloned into the polycloning site of the plasmid pET-24a
35
(Novagen, Madison, WI) using the Rapid Ligation Kit (Roche, Indiana, IN). The products obtained were pNTRC63 and pNTRC83, which encoded 63- and 83amino acid N-terminal deleted enzymes, respectively, (Nd1–63 and Nd1–83 ). The coding regions of the constructs were sequenced to confirm that no mutations were introduced (Genomics facility, Michigan State University). Expression of truncated enzymes Escherichia coli BL21(DE3) cells, transformed with pNTRC63 or pNTRC83, were grown at 37 °C in LB media with 50 lg ml1 kanamycin. When the OD600 reached 0.5, cells were induced by the addition of isopropyl-D -thiogalactoside to a final concentration of 0.5 mM. After 5 h at 25 °C, the cells were harvested by centrifugation at 10,000g for 10 min. The expression was performed in 4 and 2 L culture for pNTRC63 and pNTRC83, respectively. Purification of the truncated enzymes Both enzymes were purified following a modified protocol [11]. The cell paste was resuspended in buffer A (50 mM Tris–acetate, pH 8.0, 10 mM EDTA, and 2.5 mM DTT) in a ratio of 1:4 (w/v). After sonication and centrifugation (10,000g for 20 min) of the cell suspension, the soluble fraction of the crude extract was precipitated with (NH4 )2 SO4 (40% saturation). After centrifugation (10,000g for 20 min), we resuspended and desalted the pellet in buffer A, using Econo-Pac 10DG columns (Biorad, Hercules, CA). Then, a DEAE-Fractogel column equilibrated with buffer A was run with a linear gradient of 0–0.5 M KCl in buffer A. Both truncated enzymes eluted at about 0.18 M of KCl. Protein determination The protein concentration was determined using the BCA Protein Assay Reagent (Pierce Chemical, Dallas, Texas) [16], using bovine serum albumin as standard. Electrophoresis and Western blotting analysis SDS–PAGE and Western blotting analysis were performed with 4–15% polyacrylamide gels following the method described by Laemmli [17]. The gels were either stained by Coomassie brilliant blue or transferred to nitrocellulose membrane for Western blotting. The primary antibodies were rabbit IgG anti-BE (diluted 1:2000 in a 3% gelatin solution), recognized by antirabbit IgG conjugated with alkaline phosphatase (diluted 1:10,000) using the precipitating BM Purple AP substrate (Roche).
36
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
Branching enzyme activity assays
HPAEC analysis
Assay A The phosphorylase a stimulation assay is based on the stimulation by BE of the synthesis of a-D -glucan from aD -Glc1P catalyzed by rabbit muscle phosphorylase a [8]. In this method, a very limited concentration of primers is associated with the phosphorylase a preparation. The reaction mixture contained 0.1 M citrate, pH 7.0, 10 mM adenosine monophosphate, 0.2 mg phosphorylase a (Sigma, St. Louis, MO), and 50 mM [14 C]Glc1P at 50 dpm nmol1 . The assay was initiated by adding BE in a final volume of 100 ll. Aliquots of 25 ll were taken at 30, 60, and 90 min. One unit is defined as 1 lmol of Glc incorporated into a-D -glucan per minute at 30 °C.
Reduced amylose (4 mg, AS-320, d.p. 1815) was incubated in 25 mM Mops, pH 7.5, at 30 °C with BE (4 mU of enzyme, as determined by assay B) in a final volume of 0.4 ml. After 2 h and 16 h, the reaction was stopped by heating in a boiling bath for 2 min. The samples were debranched (hydrolysis of a-(1–6) linkages) by adding Na-acetate, pH 3.5 (final concentration 100 mM) and an excess of isoamylase (118 U). After a 90-min incubation at 45 °C, the reaction was stopped by heating in a boiling water bath for 5 min. The debranched a-D -glucan was filtered through a 0.22 lm membrane and injected in the BioLC HPAEC (Dionex, Sunnyvale, CA). To separate the debranched oligosaccharide chains according to their degree of polymerization, we used a CarboPac PA-1 (250 4 mm) column equilibrated in 150 mM NaOH (eluent A) coupled to a pulsed amperometric detector. The gradient program was the following: eluent B (150 mM NaOH + 500 mM Na-acetate) at 25% at time 0, 45% at 15 min, 60% at 45 min, 70% at 80 min, and 80% at 100 min with a flow rate of 0.3 ml min1 . The standard used was a maltodextrin (d.p. 1–20 Aldrich) solution at 1 mg ml1 . The injections of standard and samples were done through a 25-ll loop [18,19].
Assay B The branching linkage assay determines the number of branching linkages introduced by BE into the substrate, reduced amylose, prepared as described previously [18]. The assay mixture contained 25 mM Mops, pH 7.5, 17 lM reduced amylose, and the reaction was initiated by the addition of BE in a final volume of 100 ll. After stopping the branching reaction by 2-min boiling, isoamylase (5.9 U) and Na-acetate, pH 3.5, to a final concentration of 100 mM, were added to the branched product. After a 90-min incubation at 45 °C, the reaction was stopped by boiling 5 min. The reducing ends, generated by the hydrolysis of the a-(1–6) linkages, were determined by a modified Park–Johnson method [18]. One unit is defined as 1 lmol of branching linkage formed per minute at 30 °C. Assay C The iodine-staining assay is based on the spectral changes of the glucan–iodine complex that occur after the branching of the substrate [9]. The 400-ll reaction mixture contained 50 mM citrate, pH 7.0, and 0.1 mg of either potato type III amylose or corn amylopectin (Sigma) used as a substrate. The reaction was initiated by adding BE and at different times, 50 ll aliquots were withdrawn, boiled for 2 min, and mixed to 950 ll of an iodine solution prepared as described previously [9]. The absorbance changes were measured at 660 and 530 nm for substrates amylose and amylopectin, respectively. One unit of activity is defined as the decrease in absorbance of 1.0 per minute at 30 °C.
Results Purification of the wild type and truncated branching enzymes The wild type and Nd1–112 GBEs were purified as described previously [11]. Nd1–83 and Nd1–63 were purified as described in Materials and methods. Nd1–83 showed a specific activity of 535 U/mg (Table 1), which is comparable to that of Nd1–112 (538 U/mg). The specific activity of Nd1–63 was 29 U/mg (Table 2). The purity and the molecular weight of the enzymes were checked by SDS–PAGE and Western blot (data not shown). The purity of Nd1–83 and Nd1–63 was estimated by densitometry of the scanned gels to be 90 and 17%, respectively. The purification procedure provided us with branching enzymes devoid of a-amylase activity (data not shown).
Table 1 Purification table of Nd1–83 , activity was measured by phosphorylase a assay (Assay A) Purification step
Volume (ml)
Protein (mg)
Total activity (U)
Specific activity (U/mg)
Yield (%)
Crude extract 0–40% (NH4 )2 SO4 DEAE–fractogel
30 7 4
270 68 6
12,312 3812 3298
45.6 100.9 535.4
100 31 27
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
37
Table 2 Purification table of Nd1–63 , activity was measured by phosphorylase a assay (Assay A) Purification step
Volume (ml)
Protein (mg)
Total activity (U)
Specific activity (U/mg)
Yield (%)
Crude extract 0–40% (NH4 )2 SO4 DEAE–fractogel
208 20 4
770 140 5.2
1539 368 148
2 2.6 28.5
100 24 10
Chain transfer pattern of the truncated enzyme The chain transfer patterns of the wild type and truncated enzymes were studied on reduced amylose AS320 as described in Materials and methods. For that purpose, the same amount of units, measured by assay B, was used for each enzyme. After stopping the reaction, the polysaccharides were debranched and analyzed by HPAEC as described in Materials and methods. Fig. 1 shows the distribution of chains transferred to a-(1–6) linkages by wild type, Nd1–112 , Nd1–83 , and Nd1–63 GBEs. The chain transfer pattern shows that the
Fig. 2. Distribution of d.p. of the debranched a-glucan after 2 h of reaction with wild type (black), Nd1–112 (grey), Nd1–83 (dash), and Nd1–63 (dot) GBEs.
greater the number of amino acids truncated from the N-terminal of the branching enzyme, the more pronounced the shift was toward formation of longer chains in the polysaccharide product. This effect was observed for different incubation times (Fig. 1). The wild type enzyme transferred mainly chains with a d.p. of 8–14 glucan units and Nd1–112 enzyme transferred a greater proportion of chains with higher d.p. (15–20). Nd1–63 and Nd1–83 GBEs had an intermediate chain transfer pattern of 10–20 glucan units (Fig. 2). For longer chains, between 31 and 40 d.p., the Nd1–112 transferred 9.5-fold more chains than the wild type enzyme and 2- to 2.2-fold more than Nd1–83 and Nd1–63 . Truncated versions Nd1–83 and Nd1–63 had intermediate effects since they transferred 4.4- and 4.9-fold more chains than the wild type GBE, respectively (Fig. 2). Nd1–112 GBE transferred chains of 26-30 d.p. 5-fold more than the wild type enzyme, whereas Nd1–83 and Nd1–63 transferred 2.7and 3.7-fold more, respectively. For chains 21–25 d.p., Nd1–112 , Nd1–83 and Nd1–63 BEs transferred 2.5-, 1.9-, and 2.1-fold more, respectively, than wild type GBE. Fig. 1. High-performance anion-exchange chromatography of debranched a-glucan formed by the action of wild type, Nd1–112 , Nd1–83 , and Nd1–63 GBEs on reduced amylose AS-320 after 2 (A) and 16 h (B) of branching. The deviation is calculated on at least two experiments. Grey: WT, red: Nd1–83 , blue: Nd1–112 , and green: Nd1–63 . The data given by HPAEC are processed in the following way: the sum of the integration of the peaks from d.p. 6 to 40 corresponds to 100%, the integration of each peak was calculated as a percentage of that total.
Discussion Our findings show that the N-terminus of E. coli branching enzyme is involved in determining the length of transferred oligosaccharide chains. Not only does the truncation of the N-terminus affect the branching
38
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
patterns, but also a progressive shortening of the Nterminus leads to a gradual increase in the length of the transferred oligosaccharide chains. We propose that the N-terminal region is involved or aids in the binding of the amylose or partially branched glucan substrate during the processes of cleavage and transfer of the a(1–4) glucan chains. The N-terminus may also, due to its involvement in binding of the substrate, aid in specifying the particular a-(1–4) glucosidic linkage in the amylose to be cleaved by the central domain catalytic residues. The N-terminal region may leave exposed a certain size of oligosaccharide chain to be cleaved. Longer chains would not be cleaved at the substrate site, because of binding by the N-terminal and, consequently, exposed shorter chains would be cleaved. This would imply a dynamic mechanism in which a flexible N-terminus could move and restrict exposure of longer glucan chains to the active site. The longer the N-terminus the less exposed would the longer chain portion be. The structure of the 112 amino acids from the Nterminal is not known, but it has been predicted that this missing domain is structurally homologous to the domain composed of residues 108–208 that are constituted mostly of b-strands [19]. This prediction was performed before the three-dimensional structure of the truncated enzyme was obtained [14]. It has been proposed that the two modules, N1 and N2, had originated from DNA duplication. It has been confirmed when the 3-D structure was solved that the N-terminus, in this case N2, is a b-sandwich fold. These two domains are linked by a loop that is sensitive to proteases [11]. It is possible that the elongation of the N-terminus by domain duplication is an evolutionary mechanism to control the pattern of branching according to the necessities of the organism or tissue. Divergence between enzymes in this region is higher than in the central ða=bÞ-barrel domain. The overall identity of the N-terminus is 22% among the branching enzymes from E. coli, Haemophylus influenzae, Agrobacterium tumefaciens, and Pectobacterium chrysanthemi PY35 [20], whereas the overall identity of their ða=bÞ-barrel domains is 51% (data not shown). These results indicate that the N-terminus is important in having specific interactions with the substrates and in agreement with the idea that the N-terminus is involved in determining the branching pattern. It would be of
interest to determine the 3-D structure of the missing 112 amino acids from the E. coli branching enzyme and the structure of the whole enzyme in the presence of the substrate or a substrate analogue. More crystallization experiments are currently being performed in that direction. Acknowledgment This work was supported in part by a grant from the Department of Energy (DE-FG02-93ER20121) (J.P.). References [1] J. Preiss, M.N. Sivak, Genet. Eng. 20 (1998) 177–223. [2] J. Preiss, Annu. Rev. Microbiol. 38 (1984) 419–458. [3] H.P. Guan, T. Kuriki, M.N. Sivak, J. Preiss, Proc. Natl. Acad. Sci. USA 92 (1995) 964–967. [4] T. Baba, K. Kimura, K. Mizuno, H. Etoh, Y. Ishida, O. Shida, Y. Arai, Biochem. Biophys. Res. Commun. 181 (1991) 87–94. [5] D. Borovsky, E.E. Smith, W.J. Whelan, Eur. J. Biochem. 62 (1976) 307–312. [6] C.D. Boyer, J. Preiss, Carbohyd. Res. 61 (1978) 321–334. [7] Y. Nakamura, T. Takeido, K. Kawaguchi, Y. Hiroaki, Physiol. Plant 84 (1992) 329–335. [8] J.S. Hawker, J.L. Ozbun, H. Ozaki, E. Greenberg, J. Preiss, Arch. Biochem. Biophys. 160 (1974) 530–551. [9] H.P. Guan, J. Preiss, Plant Physiol. 102 (1993) 1269–1273. [10] T. Kuriki, D.C. Stewart, J. Preiss, J. Biol. Chem. 272 (1997) 28999–29004. [11] K. Binderup, R. Mikkelsen, J. Preiss, Arch. Biochem. Biophys. 377 (2000) 366–371. [12] K. Binderup, R. Mikkelsen, J. Preiss, Arch. Biochem. Biophys. 397 (2002) 279–285. [13] M.C. Abad, K. Binderup, J. Preiss, J.H. Geiger, Acta Crystallogr. 58 (2002) 359–361. [14] M.C. Abad, K. Binderup, J. Rios-Steiner, R.K. Arni, J. Preiss, J.H. Geiger, J. Biol. Chem. 277 (2002) 42164–42170. [15] H.P. Guan, P. Li, J. Imparl-Radosevich, J. Preiss, P. Keeling, Arch. Biochem. Biophys. 342 (1997) 92–98. [16] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Anal. Biochem. 150 (1985) 76–85. [17] U.K. Laemmli, Nature 227 (1970) 680–685. [18] Y. Takeda, H.P. Guan, J. Preiss, Carbohyd. Res. 240 (1993) 253–263. [19] K. Binderup, J. Preiss, Biochemistry 37 (1998) 9033–9037.
Characterization of the branching patterns of glycogen branching enzyme truncated on the N-terminus Claire H. Devillers, Mary E. Piper, Miguel A. Ballicora, and Jack Preiss* Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, MI 48824, USA Received 30 May 2003, and in revised form 7 July 2003
Abstract Truncation of 112 amino acids at the N-terminus (Nd1–112 ) changes the chain transfer pattern of the Escherichia coli glycogen branching enzyme (GBE) [Arch. Biochem. Biophys. 397 (2002) 279]. We investigated further the role of the N-terminus by engineering other truncated GBEs and analyzing the branching pattern by high-performance anion-exchange chromatography. The wild type GBE transfers mainly chains with a degree of polymerization (d.p.) of 8–14, the Nd1–112 enzyme transfers a greater proportion of chains with higher d.p. 15–20, whereas the 63- and 83-amino acid deleted enzymes had an intermediate pattern of transferred chains (d.p. 10–20). These data showed that a progressive shortening of the N-terminus leads to a gradual increase in the length of the transferred chains, suggesting that the N-terminus provides a support for the glucan substrate during the processes of cleavage and transfer of the a-(1–4) glucan chains. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Branching enzyme; Glycogen synthesis; Starch; Branching pattern; a-Glucan transfer
The biosynthesis of starch in plants and glycogen in bacteria involves at least three enzymes: ADP-glucose pyrophosphorylase, starch/glycogen synthase, and branching enzyme (BE)1 [1,2]. ADP-glucose pyrophosphorylase catalyzes the regulatory step of biosynthesis, whereas starch/glycogen synthase and branching enzyme are more involved in forming the structure of the starch or glycogen [3]. Branching enzyme belongs to the a-amylase family 13 containing a central ða=bÞ8 -barrel with four conserved regions that form the catalytic site of the enzyme [4]. BE catalyzes the cleavage of an a-(1– 4) glucosidic linkage and transfers the cleaved oligosaccharide to form a new a-(1–6) branch [5]. Several BEs have been identified in bacteria [2] and plants [6–8]. In maize, spinach or rice, different isoforms encoded by different genes have also been characterized. Those *
Corresponding author. Fax: 1-517-353-9334. E-mail address: [email protected] (J. Preiss). 1 Abbreviations used: BE, branching enzyme; d.p., degree of polymerization; GBE, glycogen branching enzyme; MBEI, maize branching enzyme I; SDS, sodium dodecyl sulphate; PAGE, polyacrylamide gel electrophoresis; HPAEC, high-performance anionexchange chromatography. 0003-9861/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0003-9861(03)00341-2
isoforms present different branching patterns and substrate specificities, which probably lead to different but complementary roles in starch synthesis [6–8]. For instance, the isoform maize branching enzyme I (MBEI) transfers long chains (degree of polymerization [d.p.] 40–100) and amylose is its specific substrate, whereas MBEII transfers shorter chains (d.p. 6–14) and prefers amylopectin as a substrate [9]. It is possible that MBEI initiates branching of lesser-branched polymers similar to amylose and then MBEII uses this product for further branching to form the final end product, amylopectin [9]. Different chimeric and truncated BEs have been studied to determine the role of the N- and C-terminal domains of these enzymes. An active hybrid of the Nterminal domain and ða=bÞ8 -barrel from MBEII and the C-terminal domain from MBEI has been constructed and characterized [10]. Its substrate specificity was similar to that of MBEI and its branching pattern to that of MBEII. The reverse construct exhibited a branching pattern similar to that of MBEI showing that the N-terminus is responsible for the branching pattern. Partial digestion of the Escherichia coli glycogen branching enzyme (GBE) with proteases and analysis of
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
the kinetic parameters of the products were performed to study the role of different domains [11]. A proteolyzed enzyme with 112 residues removed from the N-terminus was 27–59% as active as the wild type enzyme (measured by different methods). A 112 N-terminal deletion was later obtained by molecular biology techniques and led to an alteration of the branching pattern without affecting the substrate specificity [11]. The truncated enzyme transferred a greater amount of longer chains than the wild type enzyme [12]. The 3-D structure of the 112-amino acid truncated GBE was determined by X-ray crystallography [13,14]. That structure comprised three domains, a 128-residue N-terminal domain (lacking the 112 amino acids) arranged in seven b-strands, a 372-residue central domain forming the ða=bÞ8 -barrel common to a-amylase family, and a 116-residue C-terminal domain. Although the central domain is common to all enzymes belonging to this family, only the isoamylases share with GBE the structure of all three domains. Unfortunately, the wild type GBE has not been crystallized yet. To understand better the role of the N-terminal domain, we investigated another set of truncated GBEs. We engineered by recombinant technology truncations of 63 and 83 amino acids. After partial purification of those enzymes, we analyzed their branching patterns by high-performance anion-exchange chromatography (HPAEC) and compared them to those of the previous 112-amino acid truncated and the wild type enzymes [12].
Materials and methods Construction of the truncated genes pEXSB is a derivative of pET-23d and carries resistance to kanamycin, a glgB ribosome-binding site, and the full-length E. coli glgB gene [15]. pEXSB was purified from Epicurian XL2 blue cells (Stratagene, La Jolla, CA) and used as a template for PCR using the following primers 83-residue deletion: upper primer 50 CGC TAT CAT ATG GCT GTT GTC TGG CAT G 30 lower primer 50 GCT AGT TAT GCT CAG CGG 30 63-residue deletion: upper primer 50 GAG TGT CAT ATG TCA CGG GGA TTC TTT A 30 lower primer 50 GCT AGT TAT GCT CAG CGG 30 The two upper primers engineered an NdeI restriction site needed for subcloning and the lower primers anneal to the T7 terminator present in pEXSB. After purification, the PCR products were digested by NdeI and SalI (New England Biolabs, Beverly, MA) and subcloned into the polycloning site of the plasmid pET-24a
35
(Novagen, Madison, WI) using the Rapid Ligation Kit (Roche, Indiana, IN). The products obtained were pNTRC63 and pNTRC83, which encoded 63- and 83amino acid N-terminal deleted enzymes, respectively, (Nd1–63 and Nd1–83 ). The coding regions of the constructs were sequenced to confirm that no mutations were introduced (Genomics facility, Michigan State University). Expression of truncated enzymes Escherichia coli BL21(DE3) cells, transformed with pNTRC63 or pNTRC83, were grown at 37 °C in LB media with 50 lg ml1 kanamycin. When the OD600 reached 0.5, cells were induced by the addition of isopropyl-D -thiogalactoside to a final concentration of 0.5 mM. After 5 h at 25 °C, the cells were harvested by centrifugation at 10,000g for 10 min. The expression was performed in 4 and 2 L culture for pNTRC63 and pNTRC83, respectively. Purification of the truncated enzymes Both enzymes were purified following a modified protocol [11]. The cell paste was resuspended in buffer A (50 mM Tris–acetate, pH 8.0, 10 mM EDTA, and 2.5 mM DTT) in a ratio of 1:4 (w/v). After sonication and centrifugation (10,000g for 20 min) of the cell suspension, the soluble fraction of the crude extract was precipitated with (NH4 )2 SO4 (40% saturation). After centrifugation (10,000g for 20 min), we resuspended and desalted the pellet in buffer A, using Econo-Pac 10DG columns (Biorad, Hercules, CA). Then, a DEAE-Fractogel column equilibrated with buffer A was run with a linear gradient of 0–0.5 M KCl in buffer A. Both truncated enzymes eluted at about 0.18 M of KCl. Protein determination The protein concentration was determined using the BCA Protein Assay Reagent (Pierce Chemical, Dallas, Texas) [16], using bovine serum albumin as standard. Electrophoresis and Western blotting analysis SDS–PAGE and Western blotting analysis were performed with 4–15% polyacrylamide gels following the method described by Laemmli [17]. The gels were either stained by Coomassie brilliant blue or transferred to nitrocellulose membrane for Western blotting. The primary antibodies were rabbit IgG anti-BE (diluted 1:2000 in a 3% gelatin solution), recognized by antirabbit IgG conjugated with alkaline phosphatase (diluted 1:10,000) using the precipitating BM Purple AP substrate (Roche).
36
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
Branching enzyme activity assays
HPAEC analysis
Assay A The phosphorylase a stimulation assay is based on the stimulation by BE of the synthesis of a-D -glucan from aD -Glc1P catalyzed by rabbit muscle phosphorylase a [8]. In this method, a very limited concentration of primers is associated with the phosphorylase a preparation. The reaction mixture contained 0.1 M citrate, pH 7.0, 10 mM adenosine monophosphate, 0.2 mg phosphorylase a (Sigma, St. Louis, MO), and 50 mM [14 C]Glc1P at 50 dpm nmol1 . The assay was initiated by adding BE in a final volume of 100 ll. Aliquots of 25 ll were taken at 30, 60, and 90 min. One unit is defined as 1 lmol of Glc incorporated into a-D -glucan per minute at 30 °C.
Reduced amylose (4 mg, AS-320, d.p. 1815) was incubated in 25 mM Mops, pH 7.5, at 30 °C with BE (4 mU of enzyme, as determined by assay B) in a final volume of 0.4 ml. After 2 h and 16 h, the reaction was stopped by heating in a boiling bath for 2 min. The samples were debranched (hydrolysis of a-(1–6) linkages) by adding Na-acetate, pH 3.5 (final concentration 100 mM) and an excess of isoamylase (118 U). After a 90-min incubation at 45 °C, the reaction was stopped by heating in a boiling water bath for 5 min. The debranched a-D -glucan was filtered through a 0.22 lm membrane and injected in the BioLC HPAEC (Dionex, Sunnyvale, CA). To separate the debranched oligosaccharide chains according to their degree of polymerization, we used a CarboPac PA-1 (250 4 mm) column equilibrated in 150 mM NaOH (eluent A) coupled to a pulsed amperometric detector. The gradient program was the following: eluent B (150 mM NaOH + 500 mM Na-acetate) at 25% at time 0, 45% at 15 min, 60% at 45 min, 70% at 80 min, and 80% at 100 min with a flow rate of 0.3 ml min1 . The standard used was a maltodextrin (d.p. 1–20 Aldrich) solution at 1 mg ml1 . The injections of standard and samples were done through a 25-ll loop [18,19].
Assay B The branching linkage assay determines the number of branching linkages introduced by BE into the substrate, reduced amylose, prepared as described previously [18]. The assay mixture contained 25 mM Mops, pH 7.5, 17 lM reduced amylose, and the reaction was initiated by the addition of BE in a final volume of 100 ll. After stopping the branching reaction by 2-min boiling, isoamylase (5.9 U) and Na-acetate, pH 3.5, to a final concentration of 100 mM, were added to the branched product. After a 90-min incubation at 45 °C, the reaction was stopped by boiling 5 min. The reducing ends, generated by the hydrolysis of the a-(1–6) linkages, were determined by a modified Park–Johnson method [18]. One unit is defined as 1 lmol of branching linkage formed per minute at 30 °C. Assay C The iodine-staining assay is based on the spectral changes of the glucan–iodine complex that occur after the branching of the substrate [9]. The 400-ll reaction mixture contained 50 mM citrate, pH 7.0, and 0.1 mg of either potato type III amylose or corn amylopectin (Sigma) used as a substrate. The reaction was initiated by adding BE and at different times, 50 ll aliquots were withdrawn, boiled for 2 min, and mixed to 950 ll of an iodine solution prepared as described previously [9]. The absorbance changes were measured at 660 and 530 nm for substrates amylose and amylopectin, respectively. One unit of activity is defined as the decrease in absorbance of 1.0 per minute at 30 °C.
Results Purification of the wild type and truncated branching enzymes The wild type and Nd1–112 GBEs were purified as described previously [11]. Nd1–83 and Nd1–63 were purified as described in Materials and methods. Nd1–83 showed a specific activity of 535 U/mg (Table 1), which is comparable to that of Nd1–112 (538 U/mg). The specific activity of Nd1–63 was 29 U/mg (Table 2). The purity and the molecular weight of the enzymes were checked by SDS–PAGE and Western blot (data not shown). The purity of Nd1–83 and Nd1–63 was estimated by densitometry of the scanned gels to be 90 and 17%, respectively. The purification procedure provided us with branching enzymes devoid of a-amylase activity (data not shown).
Table 1 Purification table of Nd1–83 , activity was measured by phosphorylase a assay (Assay A) Purification step
Volume (ml)
Protein (mg)
Total activity (U)
Specific activity (U/mg)
Yield (%)
Crude extract 0–40% (NH4 )2 SO4 DEAE–fractogel
30 7 4
270 68 6
12,312 3812 3298
45.6 100.9 535.4
100 31 27
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
37
Table 2 Purification table of Nd1–63 , activity was measured by phosphorylase a assay (Assay A) Purification step
Volume (ml)
Protein (mg)
Total activity (U)
Specific activity (U/mg)
Yield (%)
Crude extract 0–40% (NH4 )2 SO4 DEAE–fractogel
208 20 4
770 140 5.2
1539 368 148
2 2.6 28.5
100 24 10
Chain transfer pattern of the truncated enzyme The chain transfer patterns of the wild type and truncated enzymes were studied on reduced amylose AS320 as described in Materials and methods. For that purpose, the same amount of units, measured by assay B, was used for each enzyme. After stopping the reaction, the polysaccharides were debranched and analyzed by HPAEC as described in Materials and methods. Fig. 1 shows the distribution of chains transferred to a-(1–6) linkages by wild type, Nd1–112 , Nd1–83 , and Nd1–63 GBEs. The chain transfer pattern shows that the
Fig. 2. Distribution of d.p. of the debranched a-glucan after 2 h of reaction with wild type (black), Nd1–112 (grey), Nd1–83 (dash), and Nd1–63 (dot) GBEs.
greater the number of amino acids truncated from the N-terminal of the branching enzyme, the more pronounced the shift was toward formation of longer chains in the polysaccharide product. This effect was observed for different incubation times (Fig. 1). The wild type enzyme transferred mainly chains with a d.p. of 8–14 glucan units and Nd1–112 enzyme transferred a greater proportion of chains with higher d.p. (15–20). Nd1–63 and Nd1–83 GBEs had an intermediate chain transfer pattern of 10–20 glucan units (Fig. 2). For longer chains, between 31 and 40 d.p., the Nd1–112 transferred 9.5-fold more chains than the wild type enzyme and 2- to 2.2-fold more than Nd1–83 and Nd1–63 . Truncated versions Nd1–83 and Nd1–63 had intermediate effects since they transferred 4.4- and 4.9-fold more chains than the wild type GBE, respectively (Fig. 2). Nd1–112 GBE transferred chains of 26-30 d.p. 5-fold more than the wild type enzyme, whereas Nd1–83 and Nd1–63 transferred 2.7and 3.7-fold more, respectively. For chains 21–25 d.p., Nd1–112 , Nd1–83 and Nd1–63 BEs transferred 2.5-, 1.9-, and 2.1-fold more, respectively, than wild type GBE. Fig. 1. High-performance anion-exchange chromatography of debranched a-glucan formed by the action of wild type, Nd1–112 , Nd1–83 , and Nd1–63 GBEs on reduced amylose AS-320 after 2 (A) and 16 h (B) of branching. The deviation is calculated on at least two experiments. Grey: WT, red: Nd1–83 , blue: Nd1–112 , and green: Nd1–63 . The data given by HPAEC are processed in the following way: the sum of the integration of the peaks from d.p. 6 to 40 corresponds to 100%, the integration of each peak was calculated as a percentage of that total.
Discussion Our findings show that the N-terminus of E. coli branching enzyme is involved in determining the length of transferred oligosaccharide chains. Not only does the truncation of the N-terminus affect the branching
38
C.H. Devillers et al. / Archives of Biochemistry and Biophysics 418 (2003) 34–38
patterns, but also a progressive shortening of the Nterminus leads to a gradual increase in the length of the transferred oligosaccharide chains. We propose that the N-terminal region is involved or aids in the binding of the amylose or partially branched glucan substrate during the processes of cleavage and transfer of the a(1–4) glucan chains. The N-terminus may also, due to its involvement in binding of the substrate, aid in specifying the particular a-(1–4) glucosidic linkage in the amylose to be cleaved by the central domain catalytic residues. The N-terminal region may leave exposed a certain size of oligosaccharide chain to be cleaved. Longer chains would not be cleaved at the substrate site, because of binding by the N-terminal and, consequently, exposed shorter chains would be cleaved. This would imply a dynamic mechanism in which a flexible N-terminus could move and restrict exposure of longer glucan chains to the active site. The longer the N-terminus the less exposed would the longer chain portion be. The structure of the 112 amino acids from the Nterminal is not known, but it has been predicted that this missing domain is structurally homologous to the domain composed of residues 108–208 that are constituted mostly of b-strands [19]. This prediction was performed before the three-dimensional structure of the truncated enzyme was obtained [14]. It has been proposed that the two modules, N1 and N2, had originated from DNA duplication. It has been confirmed when the 3-D structure was solved that the N-terminus, in this case N2, is a b-sandwich fold. These two domains are linked by a loop that is sensitive to proteases [11]. It is possible that the elongation of the N-terminus by domain duplication is an evolutionary mechanism to control the pattern of branching according to the necessities of the organism or tissue. Divergence between enzymes in this region is higher than in the central ða=bÞ-barrel domain. The overall identity of the N-terminus is 22% among the branching enzymes from E. coli, Haemophylus influenzae, Agrobacterium tumefaciens, and Pectobacterium chrysanthemi PY35 [20], whereas the overall identity of their ða=bÞ-barrel domains is 51% (data not shown). These results indicate that the N-terminus is important in having specific interactions with the substrates and in agreement with the idea that the N-terminus is involved in determining the branching pattern. It would be of
interest to determine the 3-D structure of the missing 112 amino acids from the E. coli branching enzyme and the structure of the whole enzyme in the presence of the substrate or a substrate analogue. More crystallization experiments are currently being performed in that direction. Acknowledgment This work was supported in part by a grant from the Department of Energy (DE-FG02-93ER20121) (J.P.). References [1] J. Preiss, M.N. Sivak, Genet. Eng. 20 (1998) 177–223. [2] J. Preiss, Annu. Rev. Microbiol. 38 (1984) 419–458. [3] H.P. Guan, T. Kuriki, M.N. Sivak, J. Preiss, Proc. Natl. Acad. Sci. USA 92 (1995) 964–967. [4] T. Baba, K. Kimura, K. Mizuno, H. Etoh, Y. Ishida, O. Shida, Y. Arai, Biochem. Biophys. Res. Commun. 181 (1991) 87–94. [5] D. Borovsky, E.E. Smith, W.J. Whelan, Eur. J. Biochem. 62 (1976) 307–312. [6] C.D. Boyer, J. Preiss, Carbohyd. Res. 61 (1978) 321–334. [7] Y. Nakamura, T. Takeido, K. Kawaguchi, Y. Hiroaki, Physiol. Plant 84 (1992) 329–335. [8] J.S. Hawker, J.L. Ozbun, H. Ozaki, E. Greenberg, J. Preiss, Arch. Biochem. Biophys. 160 (1974) 530–551. [9] H.P. Guan, J. Preiss, Plant Physiol. 102 (1993) 1269–1273. [10] T. Kuriki, D.C. Stewart, J. Preiss, J. Biol. Chem. 272 (1997) 28999–29004. [11] K. Binderup, R. Mikkelsen, J. Preiss, Arch. Biochem. Biophys. 377 (2000) 366–371. [12] K. Binderup, R. Mikkelsen, J. Preiss, Arch. Biochem. Biophys. 397 (2002) 279–285. [13] M.C. Abad, K. Binderup, J. Preiss, J.H. Geiger, Acta Crystallogr. 58 (2002) 359–361. [14] M.C. Abad, K. Binderup, J. Rios-Steiner, R.K. Arni, J. Preiss, J.H. Geiger, J. Biol. Chem. 277 (2002) 42164–42170. [15] H.P. Guan, P. Li, J. Imparl-Radosevich, J. Preiss, P. Keeling, Arch. Biochem. Biophys. 342 (1997) 92–98. [16] P.K. Smith, R.I. Krohn, G.T. Hermanson, A.K. Mallia, F.H. Gartner, M.D. Provenzano, E.K. Fujimoto, N.M. Goeke, B.J. Olson, D.C. Klenk, Anal. Biochem. 150 (1985) 76–85. [17] U.K. Laemmli, Nature 227 (1970) 680–685. [18] Y. Takeda, H.P. Guan, J. Preiss, Carbohyd. Res. 240 (1993) 253–263. [19] K. Binderup, J. Preiss, Biochemistry 37 (1998) 9033–9037.