- Email: [email protected]
Function of second glucan binding site including tyrosines 54 and 101 in Thermus aquaticus amylomaltase
JOURNAL OF BIOSCIENCE AND BIOENGINEERING Vol. 103, No. 2, 167–173. 2007 DOI: 10.1263/jbb.103.167
© 2007, The Society for Biotechnology, Japan
Function of Second Glucan Binding Site Including Tyrosines 54 and 101 in Thermus aquaticus Amylomaltase Kazutoshi Fujii,1* Hirotaka Minagawa,2 Yoshinobu Terada,1 Takeshi Takaha,1 Takashi Kuriki,1 Jiro Shimada,2 and Hiroki Kaneko3 Biochemical Research Laboratory, Ezaki Glico Co., Ltd., 4-5-6 Utajima, Nishiyodogawa-ku, Osaka 555-8502, Japan,1 Biomaterials Processing Fundamental Research Laboratories, NEC Co., Ltd., 34 Miyukigaoka, Tsukuba, Ibaraki 305-8501, Japan 2 and Department of Applied Physics, College of Humanities and Sciences, Nihon University, 3-25-40 Sakurajousui, Setagaya-ku, Tokyo 156-8550, Japan3 Received 15 September 2006/Accepted 23 November 2006
Amylomaltase from Thermus aquaticus catalyzes three types of transglycosylation reaction, as well as a weak hydrolytic reaction of α-1,4 glucan. From our previous study [Fujii et al., Appl. Environ. Microbiol., 71, 5823–5827 (2005)], tyrosine 54 (Y54) was identified as an amino acid controlling the reaction specificity of this enzyme. Since Y54 is not located around the active site but in the proposed second glucan binding site that is 14 Å away from catalytic residues, the functions of Y54 and the second glucan binding site are of great interest. In this study, we introduced mutations into another tyrosine (Y101) in the second glucan binding site. The obtained mutated enzymes were subjected to all four types of enzyme assay and the effects of mutations on the reaction specificities of these enzymes were comprehensively investigated. These studies indicated that the amino acid substitution at Y54 or Y101 for removing their aromatic side chain increases cyclization activity (intra-molecular transglycosylation reaction) but decreases disproportionation, coupling and hydrolytic activities (inter-molecular reactions). The superimposition of the reported structures of the enzyme with and without substrate analog revealed the occurrence of a conformational change in which a donor binding site becomes open. From lines of evidence, we conclude that the binding of glucan substrate to the second glucan binding site through an interaction with the aromatic side chains of Y54 and Y101 is a trigger for the enzyme to take a completely active conformation for all four types of activity, but prevents the cyclization reaction to occur since the flexibility of the glucan is restricted by such binding. [Key words: amylomaltase, Thermus aquaticus, transglycosylation, cycloamylose]
The α-amylase family is a well-established family of enzymes catalyzing the hydrolysis of an α-1,4- or α-1,6-glucosidic linkage and transglycosylation to form the α-1,4- or α-1,6-glucosidic linkage (1, 2). The α-amylase family includes extensively studied enzymes such as α-amylase (EC 3.2.1.1), isoamylase (EC 3.2.1.68), branching enzyme (EC 2.4.1.18) and cyclodextrin glycosyltransferase (CGTase; EC 2.4.1.19). Amylomaltase (EC 2.4.1.25) is also a member of the α-amylase family and is widely distributed in archaea (3), bacteria (4, 5), and plants (6, 7). This enzyme catalyzes the transfer of the α-1,4 glucan segment from one α-1,4 glucan molecule (donor) to another α-1,4 glucan molecule (acceptor), as expressed in the following equation, (α-1,4 glucan)m + (α-1,4 glucan)n ↔ (α-1,4 glucan)m–x + (α-1,4 glucan)n+x
fer reaction (cyclization reaction, Fig. 1A-2), within a single linear glucan molecule, to produce cyclic α-1,4 glucan (cycloamylose), as follows: (α-1,4 glucan)n ↔ cyclic (α-1,4 glucan)x + (α-1,4 glucan)n–x
(2)
This reaction is reversible, and the reverse reaction is often referred to as a coupling reaction (Fig. 1B-4). In addition to the three types of transglycosylation reaction described above, this enzyme showed a weak hydrolytic activity (Fig. 1A-3, 1B-5). Amylomaltase is distinct from the other enzymes of the α-amylase family, since it mainly catalyzes transglycosylation reactions and rarely catalyzes hydrolytic reactions. This enzyme has also received interest because its cyclization reaction can be used for the production of cycloamylose with a degree of polymerization (DP) of 22, which is much larger than that of conventional cyclodextrins (8). Amylomaltase from Thermus aquaticus has been extensively studied for its properties (9) and primary (9) and threedimensional structures (10). We have recently introduced a random mutation into the gene for the T. aquaticus amylo-
(1)
This is an inter-molecular transglycosylation reaction, and is often referred to as a disproportionation reaction (Fig. 1A-1). Amylomaltase also catalyzes an intra-molecular glucan trans* Corresponding author. e-mail: [email protected] phone: +81-(0)6-6477-8425 fax: +81-(0)6-6477-8362 167
168
FUJII ET AL.
FIG. 1. Models of action of amylomaltase on linear and cyclic alpha-1,4 glucans. Reactions catalyzed by amylomaltase on amylose (A) and cycloamylose (B). Lines and open circles indicate linear and cyclic alpha-1,4 glucan chains, respectively. Closed triangle, Attacking point by amylomaltase; closed circle, glucosyl residue at reducing end; R, alpha-1,4 glucan with any DPs. 1, Disproportionation; 2, cyclization; 3, 5 hydrolysis; 4, coupling.
maltase to obtain a mutated enzyme with a reduced hydrolytic activity (11). From the study of mutated enzymes, tyrosine 54 (Y54) was identified as a key amino acid controlling the reaction specificity of this enzyme. Przylas et al. (12) determined the crystal structure of this enzyme in complex with acarboses and reported two acarbose molecules bound to the enzyme, one in the active site and the other in the region 14 Å away from catalytic residues. Interestingly, Y54 lies in the region where the second acarbose molecule binds, which we proposed as the second glucan binding site (2nd GBS). Therefore, the function of the 2nd GBS is of great interest. In the 2nd GBS, another tyrosine residue, Y101, is also present and builds a hydrophobic stacking with the acarbose as well as Y54. The objective of this study is to better understand the effects of the 2nd GBS on the activity and reaction specificity of this enzyme. MATERIALS AND METHODS Chemicals and enzymes All restriction and modification enzymes used for recombinant DNA manipulations were purchased from Takara Bio (Shiga) or Toyobo (Osaka). Synthetic amylose AS-30 and AS-110 with average molecular masses of 30 kDa and 110 kDa, respectively, were purchased from Nakano Vinegar (Aichi). Maltotriose was purchased from Hayashibara Biochemical Laboratories (Okayama). Rhizopus sp. glucoamylase was purchased from Toyobo (Osaka), and porcine pancreas α-amylase was purchased from Sigma-Aldrich (St. Louis, MO, USA). A cycloamylose mixture with DP from 22 to about 50 was synthesized by T. aquaticus amylomaltase (9). Unless otherwise specified, all chemicals were purchased from Wako Pure Chemical Industries (Osaka). Construction of mutant T. aquaticus amylomaltase gene Each mutated enzyme was produced by site-directed mutagenesis, carried out using the Quick Change XL Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). pFQG8 (9) was used as the template DNA, and the following oligonucleotides were used for mutagenic primers: 5′-GGACTACGGCCTCCTCTTCGCCTGGA AGTGGCCCG-3′ (Y101F) and 5′-GGACTACGGCCTCCTCGGG GCCTGGAAGTGGCCCG-3′ (Y101G). The genes coding Y54
J. BIOSCI. BIOENG.,
mutated enzymes were previously obtained (11). All mutations were confirmed by DNA sequencing analysis. Production and purification of the wild-type and mutated T. aquaticus amylomaltases All wild-type and mutated amylomaltases were produced by the following method. Mutant constructs were transformed into E. coli strain BL21. E. coli carrying the plasmid was grown at 37°C for 18 h in 40 ml of LB medium with 50 µg/ml ampicillin. The cell culture was washed twice with 20 ml of 20 mM Tris–HCl buffer (pH 7.0). The suspended cells were broken by sonication at 4°C and centrifuged (10,000×g, 30 min) to remove cell debris. The crude extract was heated at 70°C for 30 min and centrifuged (10,000×g, 30 min) to remove aggregated proteins. The supernatant was loaded onto a Phenyl-Toyopearl 650 M (Tosoh, Tokyo) column equilibrated with the same buffer containing 0.5 M ammonium sulfate and washed with the same buffer containing 0.3 M ammonium sulfate. The enzyme was eluted with a linear 0.3 to 0 mM ammonium sulfate gradient in the same buffer. Recombinant amylomaltases were collected as active fractions. The enzymes thus prepared exhibited a single band on SDSpolyacrylamide gel electrophoresis (data not shown). Enzyme assays Sodium-acetate buffer (50 mM, pH 5.5) was used for all assays and enzyme dilutions. All incubations were carried out at 70°C and terminated at 100°C for 10 min, unless otherwise noted. The value for an activity is the mean of at least three independent assays. Amylomaltase acts on maltotriose (G3) and transfers a maltose unit of G3 to a non-reducing end of another G3 to generate maltopentaose (G5) and glucose. Disproportionation activity was assayed by measuring the amount of glucose liberated from G3 under the following conditions. G5 produced from G3 is also an effective substrate for disproportionation activity, but such a reaction can be neglected using a high G3 concentration. A 120-µl reaction mixture containing 200 mM maltotriose and the enzyme solution was incubated for 10 min. The amount of glucose produced in the reaction mixture was determined by glucose oxidase method (13). One unit of disproportionation activity was defined as the amount of enzyme producing 1 µmol of glucose per minute. Cyclization activity was determined by the following method. Amylose AS-30 (0.2 g) was dissolved in 10 ml of 90% (vol/vol) dimethyl sulfoxide. A 50-µl reaction mixture containing 0.1% AS-30 and the enzyme solution was incubated for 30 min. The reaction was terminated by the addition of 1 ml of chilled iodine solution, prepared daily by mixing 0.1 ml of iodine stock solution (0.26 g of I2 and 2.6 g of KI in 10 ml of water) and 0.1 ml of 1 N HCl and diluting the mixture to 26 ml. One unit of cyclization activity was defined as the amount of enzyme reducing the A660 of the amyloseiodine complex by 10% per minute. The activity determined in this way was considered cyclization activity, since it was considered that the decrease in A660 resulted from the decrease in the molecular weight of amylose due to the cyclization reaction. To determine coupling activity, cyclic glucan, cycloamylose, and glucose were used as donor and acceptor substrates, respectively. Amylomaltase catalyzes the linearlization of cycloamylose and transfers the linear glucan chain to glucose. Coupling activity was assayed by measuring the amount of glucose consumed under the following conditions. A 120-µl reaction mixture containing 0.5% (wt/vol) cycloamylose mixture, 0.02% (wt/vol) glucose and the enzyme solution was incubated for 10 min. The amount of glucose in the reaction mixture was determined by a glucose oxidase method (13). One unit of coupling activity was defined as the amount of enzyme reducing 1 µmol of glucose per minute. We should note that coupling activity was assayed under the acceptor limiting condition, so the value of coupling activity should be used only for comparison within the coupling activities of variant enzymes. Hydrolytic activity was assayed using cycloamylose as substrate by the following method. A100-µl reaction mixture containing 0.5%
ROLES OF Y54 AND Y101 IN T. AQUATICUS AMYLOMALTASE
VOL. 103, 2007
(wt/vol) cycloamylose mixture and the enzyme solution was incubated for 60 min, and the subsequent total reducing power of glucan in the reaction mixture was determined by a modified ParkJohnson method (14). One unit of hydrolytic activity was defined as the amount of enzyme producing 1 µmol of reducing sugar as glucose per minute. Synthesis of cycloamylose by T. aquaticus amylomaltase and mutated enzymes The amount of cycloamylose was determined as described previously (9). The synthetic amylose AS-110 (0.2 g) was dissolved in 10 ml of 90% (vol/vol) dimethyl sulfoxide. A 2.0 ml reaction mixture containing 40 U (cyclization activity) of amylomaltase, 0.2 ml of amylose solution, and 50 mM sodium acetate buffer (pH 5.5) was incubated at 70°C. The reaction was terminated by heating the mixture at 100°C for 30 min. Next, 100 µl of reaction solution was incubated with glucoamylase (3.6 U) with or without α-amylase (0.52 U) at 40°C for 3 h. The amount of released glucose was measured by a glucose oxidase method. The amount of cycloamyloses was calculated by subtracting the amount of glucose released by glucoamylase and α-amylase treatment from that released by only glucoamylase. The cycloamylose solution thus prepared was purified by ethanol precipitation, and then redissolved in 50 µl of 0.15 N NaOH, and analyzed by high-performance anion exchange chromatography (HPAEC). The conditions of HPAEC analysis were also previously reported (9). Other procedures Protein concentrations were determined using the protein assay kit (Bio-Rad, Hercules, CA, USA), with bovine gamma globulin as a standard. SDS–PAGE was carried out by using a 5% to 20% gradient precast gel (Bio-Rad), and the gel was stained with coomassie brilliant blue.
RESULTS Disproportionation and coupling activities of Y54 mutated enzymes We have already produced T. aquaticus amylomaltase mutants, where Y54 was replaced into all other amino acids, and discovered that this amino acid residue greatly affects the hydrolytic and cyclization activities
169
(11). To further investigate the function of Y54, the disproportionation activities of Y54 mutated enzymes were examined. The disproportionation activities of all the Y54 mutated enzymes, except the Y54F mutated enzyme, significantly decreased, as shown in Table 1. In most cases, the mutated enzymes exhibited the disproportionation activities lower than 25% of that of the wild-type enzyme. The effect was most pronounced in Y54G, and the disproportionation activities of the mutated enzymes markedly decreased to one-tenth of that of the wild-type enzyme. The coupling activities of these mutated enzymes were next examined and are summarized in Table 1. The coupling activities of most mutated enzymes decreased whereas those of the Y54F and Y54H mutated enzymes slightly increased. The Y54G and Y54T mutated enzymes exhibited particularly low activities, 23% and 38% of the coupling activity of the wild-type enzyme, respectively. Effect of Y54 on reaction specificity of T. aquaticus amylomaltase To visualize the effects of amino acid substitutions at Y54, four activities for each of the Y54 mutated enzymes are shown in Fig. 2. It is obvious that a mutation in Y54 does not always produce a negative effect on the activities. In most cases, the amino acid substitutions at Y54 increased cyclization activity; however, they appeared to decrease disproportionation, coupling and hydrolytic activities. Such a decrease is most apparent in disproportionation activity, but to a less extent in coupling and hydrolytic activities. Some exceptions are observed in the Y54P and Y54M mutants, where all four activities remarkably decrease. Y54F showed a very similar activity profile to the wild-type enzyme, suggesting the importance of the aro-
TABLE 1. Disproportionation and coupling activities of wild-type and Y54 mutated enzymes Disproportionation activity Coupling activity Relative Relative U/mg U/mg valuea valuea Y54A 3.39 0.23 79.9 0.87 Y54C 4.33 0.29 69.1 0.75 Y54D 2.57 0.17 73.9 0.81 Y54E 3.16 0.21 83.7 0.91 Y54F 17.0 1.13 102 1.12 Y54G 1.45 0.10 21.3 0.23 Y54H 3.92 0.26 108 1.18 Y54I 4.40 0.29 58.1 0.63 Y54K 2.56 0.17 91.3 1.00 Y54L 3.89 0.26 65.4 0.71 Y54M 3.30 0.22 48.5 0.53 Y54N 2.23 0.15 54.8 0.60 Y54P 1.18 0.08 15.2 0.17 Y54Q 3.49 0.23 88.8 0.97 Y54R 2.47 0.16 76.9 0.84 Y54S 2.58 0.17 59.9 0.65 Y54T 2.47 0.16 35.1 0.38 Y54V 3.29 0.22 51.5 0.56 Y54W 7.00 0.47 36.3 0.40 Wild-type 15.0 1.00 91.5 1.00 a Activity relative to that of wild-type enzyme.
Mutated enzyme
FIG. 2. Effects of amino acid substitution at Y54 on disproportionation activity (A), coupling activity (B), cyclization activity (C) and hydrolytic activity (D). All Y54 mutated enzymes, including the wildtype enzyme, are represented by a single letter of the amino acid residue at the 54th position (Y represents the wild-type enzyme). Each activity is expressed relative to that of the wild-type enzyme and the column is in the ascending order of cyclization activity. Black and white bars indicate values smaller and larger than 1.0 of the relative value, respectively.
170
J. BIOSCI. BIOENG.,
FUJII ET AL.
TABLE 2. Activities of Y101 mutated enzymes Hydrolytic activity Relative mU/mg value 1.00 Wild-type 50.4a 0.96 Y54F 48.4a 0.16 Y54G 8.19a Y101F 51.7 1.02 Y101G 7.56 0.15 – Y54G/Y101G N.D.b a These values are derived from a previous study (11). b N.D., Not detectable. Mutated enzyme
Cyclization activity Relative U/mg value 1120a 1.00 1020a 0.91 1860a 1.67 1100 0.98 1710 1.53 1110 1.01
matic side chain. These results clearly indicate that Y54 plays a very important role in determining the reaction specificity of this enzyme. Effect of Y101 on reaction specificities To characterize the function of Y101, which is another tyrosine in the 2nd GBS, the Y101F, Y101G and Y54G/Y101G mutated enzymes were produced and subjected to all four types of enzyme assay. The four activities of amylomaltase were measured and are summarized in Table 2 with those of the Y54F, Y54G and wild-type enzymes. The Y101G mutation significantly affected the reaction specificity of T. aquaticus amylomaltase, whereas the Y101F mutation hardly affected it. In the Y101G mutated enzyme, hydrolytic and disproportionation activities decreased to one-seventh of those of the wild-type enzyme, and coupling activity also decreased to one-fourth of that of the wild-type enzyme. In contrast, cyclization activity increased by approximately 1.5-fold that of the wild-type enzyme. The effects of the Y101G mutation on reaction specificities of the enzymes were very similar to those of the Y54G mutation. Composition of cycloamylose produced by mutated enzymes T. aquaticus amylomaltase produced cyclic glucan with DP starting from 22 and never produced smaller cyclic glucans. In our previous study, we demonstrated that the introduction of mutations into Y54 affects reaction specificity, but not the smallest size of cycloamylose produced (11). To confirm whether Y101 affects the cycloamylose synthesis, the Y101G or Y54G/Y101G mutated enzyme was incubated with a synthetic amylose and the cyclic glucan thus prepared was analyzed by HPAEC. No differences were observed in the smallest size and the distribution of cyclic glucan produced among the Y54G, Y101G, Y54G/Y101G and wild-type enzymes (data not shown). These results demonstrate that Y54 and Y101 located in the 2nd GBS are not involved in determining the size of cycloamylose. Yield of cycloamylose produced by mutated enzymes T. aquaticus amylomaltase has a weak hydrolytic activity, but this activity is not preferable for the synthesis of cycloamylose. We have already reported the good correlation between the yield of cycloamylose and the ratio of hydrolytic activity to cyclization activity (11). For cycloamylose production, the enzyme with the lowest hydrolytic activity and the highest cyclization activity is preferable. Both Y54G and Y101G mutated enzymes show a decreased hydrolytic activity and an increased cyclization activity, thus meeting these requirements. To engineer more effective enzymes for cycloamylose production, an enzyme with both amino acid
Disproportionation activity Relative U/mg value 15.0 1.00 17.0 1.13 1.45 0.10 14.7 0.98 2.08 0.14 1.18 0.08
Coupling activity Relative U/mg value 91.5 1.00 102 1.12 21.3 0.23 92.2 1.01 22.8 0.25 9.23 0.10
substitutions, Y54G/Y101G, was produced. The yield of cycloamylose produced by the enzyme with Y54G and Y101G drastically increased to 90% and did not decrease even in an extended reaction time (Fig. 3). DISCUSSION We investigated the reaction specificities of the Y54 and Y101 mutated enzymes in our previous and current studies. The obtained results clearly indicate the possibility of altering the reaction specificity of amylomaltase by introducing an amino acid substitution into particular amino acids of this enzyme. In most cases, an amino acid substitution at Y54 or Y101 of T. aquaticus amylomaltase increased cyclization activity but decreased disproportionation, coupling and hydrolytic activities (Fig. 2 and Table 2). Y54 and Y101 have already been identified as amino acids bound to an acarbose molecule by hydrophobic stacking interaction (15). All lines of evidence strongly suggest that Y54 and Y101 are involved in binding with not only acarbose but also with a natural substrate and that the introduction of mutations to remove aromatic side chains significantly affects the reaction specificity of amylomaltase. In CGTases, resembling amylomaltase in terms of catalytic reactions but differing in structures, some aromatic amino acid residues involved in the reaction specificity have been reported (16–19). However such amino acid residues locate near catalytic residues, while two tyrosine residues (Y54 and Y101) of T. aquaticus amylomaltase locate away from the catalytic residues. The interesting question here is why an amino acid far from the active center affects the reaction specificity of this enzyme so significantly. Amino acid substitutions at Y54 and Y101 for removing their aromatic side chains increased cyclization activity (Table 2). The cyclization reaction is different from three other reactions since cyclization is an intra-molecular reaction but the other reactions are inter-molecular reactions (Fig. 1). In other words, the amino acid substitution of Y54 or Y101 may positively affect the intra-molecular transglycosylation reaction but negatively affect the inter-molecular reactions. However, this view is unlikely because all four reactions should be operated by the same catalytic mechanism but using different donor and acceptor molecules. Therefore, we believe that the amino acid substitutions at Y54 and Y101 should produce a negative effect for the enzyme to catalyze any of these four reactions, but specifically produce a positive effect for the enzyme to catalyze cycliza-
VOL. 103, 2007
ROLES OF Y54 AND Y101 IN T. AQUATICUS AMYLOMALTASE
171
FIG. 3. Time courses of cycloamylose production by wild-type amylomaltase (diamonds), Y54G (squares), Y101G (triangles) and Y54G/Y101G (circles) mutated enzymes. Cycloamylose yield and reducing power in the reaction solution are shown in panels A and B, respectively. The reaction conditions and the determination of the amount of cycloamylose are described in Materials and Methods. The reducing power when all amyloses are broken down to glucose is defined as 100%.
tion reaction. We discuss separately both the negative and positive effects of the amino acid substitutions at Y54 and Y101. The negative effect appears to be most pronounced in disproportionation activity. Among the four reactions of this enzyme, the disproportionating reaction differed from the other three reactions since a low-molecular-weight glucan, G3, is used both as a donor and an acceptor, whereas a high-molecular-weight substrate, amylose or cycloamylose, is used as a donor in the other three reactions. A high-molecularweight glucan is expected to bind to this enzyme from the active site to the 2nd GBS, as has been suggested by Przylas et al. (12). In contrast, G3 is not sufficiently long to interact simultaneously with both the active site and the 2nd GBS, judging from the distance. This indicates that the 2nd GBS is not be directly involved in the catalytic mechanism and also indicates the presence of an indirect mechanism for the 2nd GBS to affect the catalytic properties of this enzyme, since disproportionation activity was dramatically decreased by introducing mutations into Y54 or Y101. Furthermore, Y54 and Y101 are involved in the binding of the glucan molecule through the hydrophobic stacking interaction with the glucose ring, as has been described. Therefore, it is considered that the binding of maltotriose to Y54 and Y101 or to the 2nd GBS is essential for this enzyme to express its activity. In the mutated enzymes, such as Y54G and Y101G, the interaction of glucan to these residues or to the 2nd GBS are weakened or diminished, and as a consequence, the enzyme shows limited catalytic properties. It is important to understand how the binding of the glucan molecule to the 2nd GBS affects the catalytic properties of this enzyme. To investigate whether the conformational change of this enzyme is caused by the binding of the glucan molecule to the 2nd GBS, the structures of amylomaltase having two acarbose molecules (12), one in the 2nd GBS and the other in the active site, were carefully compared with those of the native form (10). As illustrated in Fig. 4, the superimposition of these two structures reveals the overlapping of the main chains of the two structures (a root-mean-square deviation of Cα 0.36 Å), and differences
around the 2nd GBS and acceptor binding site. In the 2nd GBS, the Cα atoms of Y54 and G55 shift by 0.86 and 0.93 Å, respectively, and the side chain of Y54 rotates around the Cα-Cβ bond by 25.2 degrees. As a result, Y54 becomes nearer the acarbose molecule. Around the acceptor binding site, we found significant changes in two loops, L342-P347 and D249-S252. The most marked difference was observed in G343, V344 and I345, which moved 1.61, 2.91 and 3.98 Å in Cα, respectively. Y250 and F251, corresponding approximately to the position of subsite +2 or +3, also move significantly with the distances of Cα of 1.32 and 1.69 Å, respectively. The conformational changes of these two loops result in the creation of a large space within the acceptor binding site, and these observations allow us to suppose that a conformational change is deeply involved in the expression of an activity. From all these available lines of evidence, we consider that the binding of a substrate to the 2nd GBS with the help of Y54 and Y101 is a trigger for this enzyme to take a completely active conformation. The increased cyclization activity can be explained as follows. To complete a cyclization reaction, a non-reducing end of the glucan chain bound to the enzyme must be turned back to the active site. In the cyclization reaction of the wild-type enzyme, an amylose substrate probably binds to both the active site and the 2nd GBS. In the mutated enzymes lacking an aromatic residue at the 54th or 101st position, the glucan chain cannot be held by the 2nd GBS; thus, the glucan chain is expected to be more flexible than that in the wild-type enzyme. As a result, since the non-reducing end of glucan has more opportunities to turn back to the catalytic site, cyclization activity increases. The apparent increase in cyclization activity is probably a result of the balance of this positive effect from the negative effect discussed earlier. In conclusion, our results strongly indicate that Y54 and Y101 away from catalytic residues significantly affect the reaction specificity of this enzyme. Our results also suggest that the conformational change caused by the binding of the glucan molecule to the 2nd GBS is important for expressing the activity of this enzyme. Although it is still unclear how
172
J. BIOSCI. BIOENG.,
FUJII ET AL.
FIG. 4. (A) Superimposition of overall structures of T. aquaticus amylomaltase in native (PDB ID: 1CWY) and acarbose-complex (1ESW) forms. The main chains of the native and acarbose-complex forms are drawn in white and blue, respectively. The side chains of three catalytic residues, D293, E340 and D395, are colored purple. Two acarbose molecules are shown by green and red sticks. The 2nd GBS is indicated by a red arrow. (B) Enlarged view of proposed acceptor binding site which is enclosed with dashed line in panel A. The red and yellow arrows indicate the D249-S252 and L342-P347 loops, respectively. (C) Enlarged view around 2nd GBS. Orange and purple sticks indicate the side chains of Y54 and Y101, respectively. Gray sticks show the side chains of Y54 and Y101 in native form. All figures are produced with Onyx2 (SGI Japan, Tokyo) and Insight II (Accelrys, San Diego, CA, USA).
the conformational change is involved in the expression of the activity, a likely speculation is that an acceptor molecule may easily bind to the acceptor binding site opened by the conformational change. ACKNOWLEDGMENTS This work was supported by a grant for the development of the next generation of bioreactor systems from the Society for TechnoInnovation of Agriculture, Forestry and Fisheries (STAFF), Tokyo, Japan, to Ezaki Glico Co., Ltd.
REFERENCES 1. Kuriki, T.: The concept of α-amylase family, p. 107–113. In Glilbert, H. J., Davies, G. J., Henrissat, B., and Svensson B. (ed.), Recent advances in carbohydrate bioengineering. Royal Society of Chemistry, Cambridge (1999). 2. Takata, H., Kuriki, T., Okada, S., Takesada, Y., Iizuka, M., Minamiura, N., and Imanaka, T.: Action of neopullulanase. Neopullulanase catalyzes both hydrolysis and transglycosylation at alpha-(1→4)- and alpha-(1→6)-glucosidic linkages. J. Biol. Chem., 267, 18447–18452 (1992). 3. Jeon, B. S., Taguchi, H., Sakai, H., Ohshima, T., Wakagi, T., and Matsuzawa, H.: 4-Alpha-glucanotransferase from
VOL. 103, 2007
4.
5. 6. 7. 8. 9.
10.
11.
the hyperthermophilic archaeon Thermococcus litoralis—enzyme purification and characterization, and gene cloning, sequencing and expression in Escherichia coli. Eur. J. Biochem., 248, 171–178 (1997). Goda, S. K., Eissa, O., Akhtar, M., and Minton, N. P.: Molecular analysis of a Clostridium butyricum NCIMB 7423 gene encoding 4-alpha-glucanotransferase and characterization of the recombinant enzyme produced in Escherichia coli. Microbiology, 143, 3287–3294 (1997). Pugsley, A. P. and Dubreuil, C.: Molecular characterization of malQ, the structural gene for the Escherichia coli enzyme amylomaltase. Mol. Microbiol., 2, 473–479 (1988). Jones, G. and Whelan, W. J.: The action pattern of D-enzyme, a transmaltodextrinylase from potato. Carbohydr. Res., 9, 483–490 (1969). Yoshio, N., Maeda, I., Taniguchi, H., and Nakamura, M.: Purification and properties of D-enzyme from malted barley. J. Jpn. Soc. Starch Sci., 33, 244–252 (1986). Takaha, T. and Smith, S. M.: The function of 4-α-glucanotransferases and their use for the production of cyclic glucans. Biotechnol. Genet. Eng. Rev., 16, 257–280 (1999). Terada, Y., Fujii, K., Takaha, T., and Okada, S.: Thermus aquaticus ATCC 33923 amylomaltase gene cloning and expression and enzyme characterization: production of cycloamylose. Appl. Environ. Microbiol., 65, 910–915 (1999). Przylas, I., Tomoo, K., Terada, Y., Takaha, T., Fujii, K., Saenger, W., and Sträter, N.: Crystal structure of amylomaltase from Thermus aquaticus, a glycosyltransferase catalyzing the production of large cyclic glucans. J. Mol. Biol., 296, 873–886 (2000). Fujii, K., Minagawa, H., Terada, Y., Takaha, T., Kuriki, T., Shimada, J., and Kaneko, H.: Use of random and saturation mutageneses to improve the properties of Thermus aquaticus amylomaltase for efficient production of cycloamyloses. Appl. Environ. Microbiol., 71, 5823–5827 (2005).
ROLES OF Y54 AND Y101 IN T. AQUATICUS AMYLOMALTASE
173
12. Przylas, I., Terada, Y., Fujii, K., Takaha, T., Saenger, W., and Sträter, N.: X-ray structure of acarbose bound to amylomaltase from Thermus aquaticus. Eur. J. Biochem., 267, 1–12 (2000). 13. Miwa, I., Okuda, J., Maeda, K., and Okuda, G.: Mutarotase effect on colorimetric determination of blood glucose with β-D-glucose oxidase. Clin. Chim. Acta, 37, 538–540 (1972). 14. Takeda, Y., Guan, H.-P., and Preiss, J.: Branching of amylose by the branching isoenzymes of maize endosperm. Carbohydr. Res., 240, 1600–1606 (1993). 15. Sträter, N., Przylas, I., Saenger, W., Terada, Y., Fujii, K., and Takaha, T.: Structural basis of the synthesis of large cycloamyloses by amylomaltase. Biologia, Bratislava, 57, 93– 99 (2002). 16. Nakamura, A., Haga, K., and Yamane, K.: Four aromatic resides in the active center of cyclodextrin glucanotransferase from alkalophilic Bacillus sp. 1011: effects of replacements on substrate binding and cyclization characteristics. Biochemistry, 33, 9929–9936 (1994). 17. Fujiwara, S., Kakihara, H., Sakaguchi, K., and Imanaka, T.: Analysis of mutations in cyclodextrin glucanotransferase from Bacillus stearothermophilus which affect cyclization characteristics and thermostability. J. Bacteriol., 174, 7478– 7481 (1992). 18. van der Veen, B. A., Leemhius, H., Kralj, S., Uitdehaag, J. C. M., Dijkstra, B. W., and Dijkhuizen, L.: Hydrophobic amino acid residues in the acceptor binding site are main determinants for reaction mechanism and specificity of cyclodextrin-glycosyltransferase. J. Biol. Chem., 276, 44557–44562 (2001). 19. Leemhuis, H., Dijkstra, B. W., and Dijkhuizen, L.: Mutations converting cyclodextrin glycosyltransferase from a transglycosylase into a starch hydrolase. FEBS Lett., 514, 189– 192 (2002).
© 2007, The Society for Biotechnology, Japan
Function of Second Glucan Binding Site Including Tyrosines 54 and 101 in Thermus aquaticus Amylomaltase Kazutoshi Fujii,1* Hirotaka Minagawa,2 Yoshinobu Terada,1 Takeshi Takaha,1 Takashi Kuriki,1 Jiro Shimada,2 and Hiroki Kaneko3 Biochemical Research Laboratory, Ezaki Glico Co., Ltd., 4-5-6 Utajima, Nishiyodogawa-ku, Osaka 555-8502, Japan,1 Biomaterials Processing Fundamental Research Laboratories, NEC Co., Ltd., 34 Miyukigaoka, Tsukuba, Ibaraki 305-8501, Japan 2 and Department of Applied Physics, College of Humanities and Sciences, Nihon University, 3-25-40 Sakurajousui, Setagaya-ku, Tokyo 156-8550, Japan3 Received 15 September 2006/Accepted 23 November 2006
Amylomaltase from Thermus aquaticus catalyzes three types of transglycosylation reaction, as well as a weak hydrolytic reaction of α-1,4 glucan. From our previous study [Fujii et al., Appl. Environ. Microbiol., 71, 5823–5827 (2005)], tyrosine 54 (Y54) was identified as an amino acid controlling the reaction specificity of this enzyme. Since Y54 is not located around the active site but in the proposed second glucan binding site that is 14 Å away from catalytic residues, the functions of Y54 and the second glucan binding site are of great interest. In this study, we introduced mutations into another tyrosine (Y101) in the second glucan binding site. The obtained mutated enzymes were subjected to all four types of enzyme assay and the effects of mutations on the reaction specificities of these enzymes were comprehensively investigated. These studies indicated that the amino acid substitution at Y54 or Y101 for removing their aromatic side chain increases cyclization activity (intra-molecular transglycosylation reaction) but decreases disproportionation, coupling and hydrolytic activities (inter-molecular reactions). The superimposition of the reported structures of the enzyme with and without substrate analog revealed the occurrence of a conformational change in which a donor binding site becomes open. From lines of evidence, we conclude that the binding of glucan substrate to the second glucan binding site through an interaction with the aromatic side chains of Y54 and Y101 is a trigger for the enzyme to take a completely active conformation for all four types of activity, but prevents the cyclization reaction to occur since the flexibility of the glucan is restricted by such binding. [Key words: amylomaltase, Thermus aquaticus, transglycosylation, cycloamylose]
The α-amylase family is a well-established family of enzymes catalyzing the hydrolysis of an α-1,4- or α-1,6-glucosidic linkage and transglycosylation to form the α-1,4- or α-1,6-glucosidic linkage (1, 2). The α-amylase family includes extensively studied enzymes such as α-amylase (EC 3.2.1.1), isoamylase (EC 3.2.1.68), branching enzyme (EC 2.4.1.18) and cyclodextrin glycosyltransferase (CGTase; EC 2.4.1.19). Amylomaltase (EC 2.4.1.25) is also a member of the α-amylase family and is widely distributed in archaea (3), bacteria (4, 5), and plants (6, 7). This enzyme catalyzes the transfer of the α-1,4 glucan segment from one α-1,4 glucan molecule (donor) to another α-1,4 glucan molecule (acceptor), as expressed in the following equation, (α-1,4 glucan)m + (α-1,4 glucan)n ↔ (α-1,4 glucan)m–x + (α-1,4 glucan)n+x
fer reaction (cyclization reaction, Fig. 1A-2), within a single linear glucan molecule, to produce cyclic α-1,4 glucan (cycloamylose), as follows: (α-1,4 glucan)n ↔ cyclic (α-1,4 glucan)x + (α-1,4 glucan)n–x
(2)
This reaction is reversible, and the reverse reaction is often referred to as a coupling reaction (Fig. 1B-4). In addition to the three types of transglycosylation reaction described above, this enzyme showed a weak hydrolytic activity (Fig. 1A-3, 1B-5). Amylomaltase is distinct from the other enzymes of the α-amylase family, since it mainly catalyzes transglycosylation reactions and rarely catalyzes hydrolytic reactions. This enzyme has also received interest because its cyclization reaction can be used for the production of cycloamylose with a degree of polymerization (DP) of 22, which is much larger than that of conventional cyclodextrins (8). Amylomaltase from Thermus aquaticus has been extensively studied for its properties (9) and primary (9) and threedimensional structures (10). We have recently introduced a random mutation into the gene for the T. aquaticus amylo-
(1)
This is an inter-molecular transglycosylation reaction, and is often referred to as a disproportionation reaction (Fig. 1A-1). Amylomaltase also catalyzes an intra-molecular glucan trans* Corresponding author. e-mail: [email protected] phone: +81-(0)6-6477-8425 fax: +81-(0)6-6477-8362 167
168
FUJII ET AL.
FIG. 1. Models of action of amylomaltase on linear and cyclic alpha-1,4 glucans. Reactions catalyzed by amylomaltase on amylose (A) and cycloamylose (B). Lines and open circles indicate linear and cyclic alpha-1,4 glucan chains, respectively. Closed triangle, Attacking point by amylomaltase; closed circle, glucosyl residue at reducing end; R, alpha-1,4 glucan with any DPs. 1, Disproportionation; 2, cyclization; 3, 5 hydrolysis; 4, coupling.
maltase to obtain a mutated enzyme with a reduced hydrolytic activity (11). From the study of mutated enzymes, tyrosine 54 (Y54) was identified as a key amino acid controlling the reaction specificity of this enzyme. Przylas et al. (12) determined the crystal structure of this enzyme in complex with acarboses and reported two acarbose molecules bound to the enzyme, one in the active site and the other in the region 14 Å away from catalytic residues. Interestingly, Y54 lies in the region where the second acarbose molecule binds, which we proposed as the second glucan binding site (2nd GBS). Therefore, the function of the 2nd GBS is of great interest. In the 2nd GBS, another tyrosine residue, Y101, is also present and builds a hydrophobic stacking with the acarbose as well as Y54. The objective of this study is to better understand the effects of the 2nd GBS on the activity and reaction specificity of this enzyme. MATERIALS AND METHODS Chemicals and enzymes All restriction and modification enzymes used for recombinant DNA manipulations were purchased from Takara Bio (Shiga) or Toyobo (Osaka). Synthetic amylose AS-30 and AS-110 with average molecular masses of 30 kDa and 110 kDa, respectively, were purchased from Nakano Vinegar (Aichi). Maltotriose was purchased from Hayashibara Biochemical Laboratories (Okayama). Rhizopus sp. glucoamylase was purchased from Toyobo (Osaka), and porcine pancreas α-amylase was purchased from Sigma-Aldrich (St. Louis, MO, USA). A cycloamylose mixture with DP from 22 to about 50 was synthesized by T. aquaticus amylomaltase (9). Unless otherwise specified, all chemicals were purchased from Wako Pure Chemical Industries (Osaka). Construction of mutant T. aquaticus amylomaltase gene Each mutated enzyme was produced by site-directed mutagenesis, carried out using the Quick Change XL Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). pFQG8 (9) was used as the template DNA, and the following oligonucleotides were used for mutagenic primers: 5′-GGACTACGGCCTCCTCTTCGCCTGGA AGTGGCCCG-3′ (Y101F) and 5′-GGACTACGGCCTCCTCGGG GCCTGGAAGTGGCCCG-3′ (Y101G). The genes coding Y54
J. BIOSCI. BIOENG.,
mutated enzymes were previously obtained (11). All mutations were confirmed by DNA sequencing analysis. Production and purification of the wild-type and mutated T. aquaticus amylomaltases All wild-type and mutated amylomaltases were produced by the following method. Mutant constructs were transformed into E. coli strain BL21. E. coli carrying the plasmid was grown at 37°C for 18 h in 40 ml of LB medium with 50 µg/ml ampicillin. The cell culture was washed twice with 20 ml of 20 mM Tris–HCl buffer (pH 7.0). The suspended cells were broken by sonication at 4°C and centrifuged (10,000×g, 30 min) to remove cell debris. The crude extract was heated at 70°C for 30 min and centrifuged (10,000×g, 30 min) to remove aggregated proteins. The supernatant was loaded onto a Phenyl-Toyopearl 650 M (Tosoh, Tokyo) column equilibrated with the same buffer containing 0.5 M ammonium sulfate and washed with the same buffer containing 0.3 M ammonium sulfate. The enzyme was eluted with a linear 0.3 to 0 mM ammonium sulfate gradient in the same buffer. Recombinant amylomaltases were collected as active fractions. The enzymes thus prepared exhibited a single band on SDSpolyacrylamide gel electrophoresis (data not shown). Enzyme assays Sodium-acetate buffer (50 mM, pH 5.5) was used for all assays and enzyme dilutions. All incubations were carried out at 70°C and terminated at 100°C for 10 min, unless otherwise noted. The value for an activity is the mean of at least three independent assays. Amylomaltase acts on maltotriose (G3) and transfers a maltose unit of G3 to a non-reducing end of another G3 to generate maltopentaose (G5) and glucose. Disproportionation activity was assayed by measuring the amount of glucose liberated from G3 under the following conditions. G5 produced from G3 is also an effective substrate for disproportionation activity, but such a reaction can be neglected using a high G3 concentration. A 120-µl reaction mixture containing 200 mM maltotriose and the enzyme solution was incubated for 10 min. The amount of glucose produced in the reaction mixture was determined by glucose oxidase method (13). One unit of disproportionation activity was defined as the amount of enzyme producing 1 µmol of glucose per minute. Cyclization activity was determined by the following method. Amylose AS-30 (0.2 g) was dissolved in 10 ml of 90% (vol/vol) dimethyl sulfoxide. A 50-µl reaction mixture containing 0.1% AS-30 and the enzyme solution was incubated for 30 min. The reaction was terminated by the addition of 1 ml of chilled iodine solution, prepared daily by mixing 0.1 ml of iodine stock solution (0.26 g of I2 and 2.6 g of KI in 10 ml of water) and 0.1 ml of 1 N HCl and diluting the mixture to 26 ml. One unit of cyclization activity was defined as the amount of enzyme reducing the A660 of the amyloseiodine complex by 10% per minute. The activity determined in this way was considered cyclization activity, since it was considered that the decrease in A660 resulted from the decrease in the molecular weight of amylose due to the cyclization reaction. To determine coupling activity, cyclic glucan, cycloamylose, and glucose were used as donor and acceptor substrates, respectively. Amylomaltase catalyzes the linearlization of cycloamylose and transfers the linear glucan chain to glucose. Coupling activity was assayed by measuring the amount of glucose consumed under the following conditions. A 120-µl reaction mixture containing 0.5% (wt/vol) cycloamylose mixture, 0.02% (wt/vol) glucose and the enzyme solution was incubated for 10 min. The amount of glucose in the reaction mixture was determined by a glucose oxidase method (13). One unit of coupling activity was defined as the amount of enzyme reducing 1 µmol of glucose per minute. We should note that coupling activity was assayed under the acceptor limiting condition, so the value of coupling activity should be used only for comparison within the coupling activities of variant enzymes. Hydrolytic activity was assayed using cycloamylose as substrate by the following method. A100-µl reaction mixture containing 0.5%
ROLES OF Y54 AND Y101 IN T. AQUATICUS AMYLOMALTASE
VOL. 103, 2007
(wt/vol) cycloamylose mixture and the enzyme solution was incubated for 60 min, and the subsequent total reducing power of glucan in the reaction mixture was determined by a modified ParkJohnson method (14). One unit of hydrolytic activity was defined as the amount of enzyme producing 1 µmol of reducing sugar as glucose per minute. Synthesis of cycloamylose by T. aquaticus amylomaltase and mutated enzymes The amount of cycloamylose was determined as described previously (9). The synthetic amylose AS-110 (0.2 g) was dissolved in 10 ml of 90% (vol/vol) dimethyl sulfoxide. A 2.0 ml reaction mixture containing 40 U (cyclization activity) of amylomaltase, 0.2 ml of amylose solution, and 50 mM sodium acetate buffer (pH 5.5) was incubated at 70°C. The reaction was terminated by heating the mixture at 100°C for 30 min. Next, 100 µl of reaction solution was incubated with glucoamylase (3.6 U) with or without α-amylase (0.52 U) at 40°C for 3 h. The amount of released glucose was measured by a glucose oxidase method. The amount of cycloamyloses was calculated by subtracting the amount of glucose released by glucoamylase and α-amylase treatment from that released by only glucoamylase. The cycloamylose solution thus prepared was purified by ethanol precipitation, and then redissolved in 50 µl of 0.15 N NaOH, and analyzed by high-performance anion exchange chromatography (HPAEC). The conditions of HPAEC analysis were also previously reported (9). Other procedures Protein concentrations were determined using the protein assay kit (Bio-Rad, Hercules, CA, USA), with bovine gamma globulin as a standard. SDS–PAGE was carried out by using a 5% to 20% gradient precast gel (Bio-Rad), and the gel was stained with coomassie brilliant blue.
RESULTS Disproportionation and coupling activities of Y54 mutated enzymes We have already produced T. aquaticus amylomaltase mutants, where Y54 was replaced into all other amino acids, and discovered that this amino acid residue greatly affects the hydrolytic and cyclization activities
169
(11). To further investigate the function of Y54, the disproportionation activities of Y54 mutated enzymes were examined. The disproportionation activities of all the Y54 mutated enzymes, except the Y54F mutated enzyme, significantly decreased, as shown in Table 1. In most cases, the mutated enzymes exhibited the disproportionation activities lower than 25% of that of the wild-type enzyme. The effect was most pronounced in Y54G, and the disproportionation activities of the mutated enzymes markedly decreased to one-tenth of that of the wild-type enzyme. The coupling activities of these mutated enzymes were next examined and are summarized in Table 1. The coupling activities of most mutated enzymes decreased whereas those of the Y54F and Y54H mutated enzymes slightly increased. The Y54G and Y54T mutated enzymes exhibited particularly low activities, 23% and 38% of the coupling activity of the wild-type enzyme, respectively. Effect of Y54 on reaction specificity of T. aquaticus amylomaltase To visualize the effects of amino acid substitutions at Y54, four activities for each of the Y54 mutated enzymes are shown in Fig. 2. It is obvious that a mutation in Y54 does not always produce a negative effect on the activities. In most cases, the amino acid substitutions at Y54 increased cyclization activity; however, they appeared to decrease disproportionation, coupling and hydrolytic activities. Such a decrease is most apparent in disproportionation activity, but to a less extent in coupling and hydrolytic activities. Some exceptions are observed in the Y54P and Y54M mutants, where all four activities remarkably decrease. Y54F showed a very similar activity profile to the wild-type enzyme, suggesting the importance of the aro-
TABLE 1. Disproportionation and coupling activities of wild-type and Y54 mutated enzymes Disproportionation activity Coupling activity Relative Relative U/mg U/mg valuea valuea Y54A 3.39 0.23 79.9 0.87 Y54C 4.33 0.29 69.1 0.75 Y54D 2.57 0.17 73.9 0.81 Y54E 3.16 0.21 83.7 0.91 Y54F 17.0 1.13 102 1.12 Y54G 1.45 0.10 21.3 0.23 Y54H 3.92 0.26 108 1.18 Y54I 4.40 0.29 58.1 0.63 Y54K 2.56 0.17 91.3 1.00 Y54L 3.89 0.26 65.4 0.71 Y54M 3.30 0.22 48.5 0.53 Y54N 2.23 0.15 54.8 0.60 Y54P 1.18 0.08 15.2 0.17 Y54Q 3.49 0.23 88.8 0.97 Y54R 2.47 0.16 76.9 0.84 Y54S 2.58 0.17 59.9 0.65 Y54T 2.47 0.16 35.1 0.38 Y54V 3.29 0.22 51.5 0.56 Y54W 7.00 0.47 36.3 0.40 Wild-type 15.0 1.00 91.5 1.00 a Activity relative to that of wild-type enzyme.
Mutated enzyme
FIG. 2. Effects of amino acid substitution at Y54 on disproportionation activity (A), coupling activity (B), cyclization activity (C) and hydrolytic activity (D). All Y54 mutated enzymes, including the wildtype enzyme, are represented by a single letter of the amino acid residue at the 54th position (Y represents the wild-type enzyme). Each activity is expressed relative to that of the wild-type enzyme and the column is in the ascending order of cyclization activity. Black and white bars indicate values smaller and larger than 1.0 of the relative value, respectively.
170
J. BIOSCI. BIOENG.,
FUJII ET AL.
TABLE 2. Activities of Y101 mutated enzymes Hydrolytic activity Relative mU/mg value 1.00 Wild-type 50.4a 0.96 Y54F 48.4a 0.16 Y54G 8.19a Y101F 51.7 1.02 Y101G 7.56 0.15 – Y54G/Y101G N.D.b a These values are derived from a previous study (11). b N.D., Not detectable. Mutated enzyme
Cyclization activity Relative U/mg value 1120a 1.00 1020a 0.91 1860a 1.67 1100 0.98 1710 1.53 1110 1.01
matic side chain. These results clearly indicate that Y54 plays a very important role in determining the reaction specificity of this enzyme. Effect of Y101 on reaction specificities To characterize the function of Y101, which is another tyrosine in the 2nd GBS, the Y101F, Y101G and Y54G/Y101G mutated enzymes were produced and subjected to all four types of enzyme assay. The four activities of amylomaltase were measured and are summarized in Table 2 with those of the Y54F, Y54G and wild-type enzymes. The Y101G mutation significantly affected the reaction specificity of T. aquaticus amylomaltase, whereas the Y101F mutation hardly affected it. In the Y101G mutated enzyme, hydrolytic and disproportionation activities decreased to one-seventh of those of the wild-type enzyme, and coupling activity also decreased to one-fourth of that of the wild-type enzyme. In contrast, cyclization activity increased by approximately 1.5-fold that of the wild-type enzyme. The effects of the Y101G mutation on reaction specificities of the enzymes were very similar to those of the Y54G mutation. Composition of cycloamylose produced by mutated enzymes T. aquaticus amylomaltase produced cyclic glucan with DP starting from 22 and never produced smaller cyclic glucans. In our previous study, we demonstrated that the introduction of mutations into Y54 affects reaction specificity, but not the smallest size of cycloamylose produced (11). To confirm whether Y101 affects the cycloamylose synthesis, the Y101G or Y54G/Y101G mutated enzyme was incubated with a synthetic amylose and the cyclic glucan thus prepared was analyzed by HPAEC. No differences were observed in the smallest size and the distribution of cyclic glucan produced among the Y54G, Y101G, Y54G/Y101G and wild-type enzymes (data not shown). These results demonstrate that Y54 and Y101 located in the 2nd GBS are not involved in determining the size of cycloamylose. Yield of cycloamylose produced by mutated enzymes T. aquaticus amylomaltase has a weak hydrolytic activity, but this activity is not preferable for the synthesis of cycloamylose. We have already reported the good correlation between the yield of cycloamylose and the ratio of hydrolytic activity to cyclization activity (11). For cycloamylose production, the enzyme with the lowest hydrolytic activity and the highest cyclization activity is preferable. Both Y54G and Y101G mutated enzymes show a decreased hydrolytic activity and an increased cyclization activity, thus meeting these requirements. To engineer more effective enzymes for cycloamylose production, an enzyme with both amino acid
Disproportionation activity Relative U/mg value 15.0 1.00 17.0 1.13 1.45 0.10 14.7 0.98 2.08 0.14 1.18 0.08
Coupling activity Relative U/mg value 91.5 1.00 102 1.12 21.3 0.23 92.2 1.01 22.8 0.25 9.23 0.10
substitutions, Y54G/Y101G, was produced. The yield of cycloamylose produced by the enzyme with Y54G and Y101G drastically increased to 90% and did not decrease even in an extended reaction time (Fig. 3). DISCUSSION We investigated the reaction specificities of the Y54 and Y101 mutated enzymes in our previous and current studies. The obtained results clearly indicate the possibility of altering the reaction specificity of amylomaltase by introducing an amino acid substitution into particular amino acids of this enzyme. In most cases, an amino acid substitution at Y54 or Y101 of T. aquaticus amylomaltase increased cyclization activity but decreased disproportionation, coupling and hydrolytic activities (Fig. 2 and Table 2). Y54 and Y101 have already been identified as amino acids bound to an acarbose molecule by hydrophobic stacking interaction (15). All lines of evidence strongly suggest that Y54 and Y101 are involved in binding with not only acarbose but also with a natural substrate and that the introduction of mutations to remove aromatic side chains significantly affects the reaction specificity of amylomaltase. In CGTases, resembling amylomaltase in terms of catalytic reactions but differing in structures, some aromatic amino acid residues involved in the reaction specificity have been reported (16–19). However such amino acid residues locate near catalytic residues, while two tyrosine residues (Y54 and Y101) of T. aquaticus amylomaltase locate away from the catalytic residues. The interesting question here is why an amino acid far from the active center affects the reaction specificity of this enzyme so significantly. Amino acid substitutions at Y54 and Y101 for removing their aromatic side chains increased cyclization activity (Table 2). The cyclization reaction is different from three other reactions since cyclization is an intra-molecular reaction but the other reactions are inter-molecular reactions (Fig. 1). In other words, the amino acid substitution of Y54 or Y101 may positively affect the intra-molecular transglycosylation reaction but negatively affect the inter-molecular reactions. However, this view is unlikely because all four reactions should be operated by the same catalytic mechanism but using different donor and acceptor molecules. Therefore, we believe that the amino acid substitutions at Y54 and Y101 should produce a negative effect for the enzyme to catalyze any of these four reactions, but specifically produce a positive effect for the enzyme to catalyze cycliza-
VOL. 103, 2007
ROLES OF Y54 AND Y101 IN T. AQUATICUS AMYLOMALTASE
171
FIG. 3. Time courses of cycloamylose production by wild-type amylomaltase (diamonds), Y54G (squares), Y101G (triangles) and Y54G/Y101G (circles) mutated enzymes. Cycloamylose yield and reducing power in the reaction solution are shown in panels A and B, respectively. The reaction conditions and the determination of the amount of cycloamylose are described in Materials and Methods. The reducing power when all amyloses are broken down to glucose is defined as 100%.
tion reaction. We discuss separately both the negative and positive effects of the amino acid substitutions at Y54 and Y101. The negative effect appears to be most pronounced in disproportionation activity. Among the four reactions of this enzyme, the disproportionating reaction differed from the other three reactions since a low-molecular-weight glucan, G3, is used both as a donor and an acceptor, whereas a high-molecular-weight substrate, amylose or cycloamylose, is used as a donor in the other three reactions. A high-molecularweight glucan is expected to bind to this enzyme from the active site to the 2nd GBS, as has been suggested by Przylas et al. (12). In contrast, G3 is not sufficiently long to interact simultaneously with both the active site and the 2nd GBS, judging from the distance. This indicates that the 2nd GBS is not be directly involved in the catalytic mechanism and also indicates the presence of an indirect mechanism for the 2nd GBS to affect the catalytic properties of this enzyme, since disproportionation activity was dramatically decreased by introducing mutations into Y54 or Y101. Furthermore, Y54 and Y101 are involved in the binding of the glucan molecule through the hydrophobic stacking interaction with the glucose ring, as has been described. Therefore, it is considered that the binding of maltotriose to Y54 and Y101 or to the 2nd GBS is essential for this enzyme to express its activity. In the mutated enzymes, such as Y54G and Y101G, the interaction of glucan to these residues or to the 2nd GBS are weakened or diminished, and as a consequence, the enzyme shows limited catalytic properties. It is important to understand how the binding of the glucan molecule to the 2nd GBS affects the catalytic properties of this enzyme. To investigate whether the conformational change of this enzyme is caused by the binding of the glucan molecule to the 2nd GBS, the structures of amylomaltase having two acarbose molecules (12), one in the 2nd GBS and the other in the active site, were carefully compared with those of the native form (10). As illustrated in Fig. 4, the superimposition of these two structures reveals the overlapping of the main chains of the two structures (a root-mean-square deviation of Cα 0.36 Å), and differences
around the 2nd GBS and acceptor binding site. In the 2nd GBS, the Cα atoms of Y54 and G55 shift by 0.86 and 0.93 Å, respectively, and the side chain of Y54 rotates around the Cα-Cβ bond by 25.2 degrees. As a result, Y54 becomes nearer the acarbose molecule. Around the acceptor binding site, we found significant changes in two loops, L342-P347 and D249-S252. The most marked difference was observed in G343, V344 and I345, which moved 1.61, 2.91 and 3.98 Å in Cα, respectively. Y250 and F251, corresponding approximately to the position of subsite +2 or +3, also move significantly with the distances of Cα of 1.32 and 1.69 Å, respectively. The conformational changes of these two loops result in the creation of a large space within the acceptor binding site, and these observations allow us to suppose that a conformational change is deeply involved in the expression of an activity. From all these available lines of evidence, we consider that the binding of a substrate to the 2nd GBS with the help of Y54 and Y101 is a trigger for this enzyme to take a completely active conformation. The increased cyclization activity can be explained as follows. To complete a cyclization reaction, a non-reducing end of the glucan chain bound to the enzyme must be turned back to the active site. In the cyclization reaction of the wild-type enzyme, an amylose substrate probably binds to both the active site and the 2nd GBS. In the mutated enzymes lacking an aromatic residue at the 54th or 101st position, the glucan chain cannot be held by the 2nd GBS; thus, the glucan chain is expected to be more flexible than that in the wild-type enzyme. As a result, since the non-reducing end of glucan has more opportunities to turn back to the catalytic site, cyclization activity increases. The apparent increase in cyclization activity is probably a result of the balance of this positive effect from the negative effect discussed earlier. In conclusion, our results strongly indicate that Y54 and Y101 away from catalytic residues significantly affect the reaction specificity of this enzyme. Our results also suggest that the conformational change caused by the binding of the glucan molecule to the 2nd GBS is important for expressing the activity of this enzyme. Although it is still unclear how
172
J. BIOSCI. BIOENG.,
FUJII ET AL.
FIG. 4. (A) Superimposition of overall structures of T. aquaticus amylomaltase in native (PDB ID: 1CWY) and acarbose-complex (1ESW) forms. The main chains of the native and acarbose-complex forms are drawn in white and blue, respectively. The side chains of three catalytic residues, D293, E340 and D395, are colored purple. Two acarbose molecules are shown by green and red sticks. The 2nd GBS is indicated by a red arrow. (B) Enlarged view of proposed acceptor binding site which is enclosed with dashed line in panel A. The red and yellow arrows indicate the D249-S252 and L342-P347 loops, respectively. (C) Enlarged view around 2nd GBS. Orange and purple sticks indicate the side chains of Y54 and Y101, respectively. Gray sticks show the side chains of Y54 and Y101 in native form. All figures are produced with Onyx2 (SGI Japan, Tokyo) and Insight II (Accelrys, San Diego, CA, USA).
the conformational change is involved in the expression of the activity, a likely speculation is that an acceptor molecule may easily bind to the acceptor binding site opened by the conformational change. ACKNOWLEDGMENTS This work was supported by a grant for the development of the next generation of bioreactor systems from the Society for TechnoInnovation of Agriculture, Forestry and Fisheries (STAFF), Tokyo, Japan, to Ezaki Glico Co., Ltd.
REFERENCES 1. Kuriki, T.: The concept of α-amylase family, p. 107–113. In Glilbert, H. J., Davies, G. J., Henrissat, B., and Svensson B. (ed.), Recent advances in carbohydrate bioengineering. Royal Society of Chemistry, Cambridge (1999). 2. Takata, H., Kuriki, T., Okada, S., Takesada, Y., Iizuka, M., Minamiura, N., and Imanaka, T.: Action of neopullulanase. Neopullulanase catalyzes both hydrolysis and transglycosylation at alpha-(1→4)- and alpha-(1→6)-glucosidic linkages. J. Biol. Chem., 267, 18447–18452 (1992). 3. Jeon, B. S., Taguchi, H., Sakai, H., Ohshima, T., Wakagi, T., and Matsuzawa, H.: 4-Alpha-glucanotransferase from
VOL. 103, 2007
4.
5. 6. 7. 8. 9.
10.
11.
the hyperthermophilic archaeon Thermococcus litoralis—enzyme purification and characterization, and gene cloning, sequencing and expression in Escherichia coli. Eur. J. Biochem., 248, 171–178 (1997). Goda, S. K., Eissa, O., Akhtar, M., and Minton, N. P.: Molecular analysis of a Clostridium butyricum NCIMB 7423 gene encoding 4-alpha-glucanotransferase and characterization of the recombinant enzyme produced in Escherichia coli. Microbiology, 143, 3287–3294 (1997). Pugsley, A. P. and Dubreuil, C.: Molecular characterization of malQ, the structural gene for the Escherichia coli enzyme amylomaltase. Mol. Microbiol., 2, 473–479 (1988). Jones, G. and Whelan, W. J.: The action pattern of D-enzyme, a transmaltodextrinylase from potato. Carbohydr. Res., 9, 483–490 (1969). Yoshio, N., Maeda, I., Taniguchi, H., and Nakamura, M.: Purification and properties of D-enzyme from malted barley. J. Jpn. Soc. Starch Sci., 33, 244–252 (1986). Takaha, T. and Smith, S. M.: The function of 4-α-glucanotransferases and their use for the production of cyclic glucans. Biotechnol. Genet. Eng. Rev., 16, 257–280 (1999). Terada, Y., Fujii, K., Takaha, T., and Okada, S.: Thermus aquaticus ATCC 33923 amylomaltase gene cloning and expression and enzyme characterization: production of cycloamylose. Appl. Environ. Microbiol., 65, 910–915 (1999). Przylas, I., Tomoo, K., Terada, Y., Takaha, T., Fujii, K., Saenger, W., and Sträter, N.: Crystal structure of amylomaltase from Thermus aquaticus, a glycosyltransferase catalyzing the production of large cyclic glucans. J. Mol. Biol., 296, 873–886 (2000). Fujii, K., Minagawa, H., Terada, Y., Takaha, T., Kuriki, T., Shimada, J., and Kaneko, H.: Use of random and saturation mutageneses to improve the properties of Thermus aquaticus amylomaltase for efficient production of cycloamyloses. Appl. Environ. Microbiol., 71, 5823–5827 (2005).
ROLES OF Y54 AND Y101 IN T. AQUATICUS AMYLOMALTASE
173
12. Przylas, I., Terada, Y., Fujii, K., Takaha, T., Saenger, W., and Sträter, N.: X-ray structure of acarbose bound to amylomaltase from Thermus aquaticus. Eur. J. Biochem., 267, 1–12 (2000). 13. Miwa, I., Okuda, J., Maeda, K., and Okuda, G.: Mutarotase effect on colorimetric determination of blood glucose with β-D-glucose oxidase. Clin. Chim. Acta, 37, 538–540 (1972). 14. Takeda, Y., Guan, H.-P., and Preiss, J.: Branching of amylose by the branching isoenzymes of maize endosperm. Carbohydr. Res., 240, 1600–1606 (1993). 15. Sträter, N., Przylas, I., Saenger, W., Terada, Y., Fujii, K., and Takaha, T.: Structural basis of the synthesis of large cycloamyloses by amylomaltase. Biologia, Bratislava, 57, 93– 99 (2002). 16. Nakamura, A., Haga, K., and Yamane, K.: Four aromatic resides in the active center of cyclodextrin glucanotransferase from alkalophilic Bacillus sp. 1011: effects of replacements on substrate binding and cyclization characteristics. Biochemistry, 33, 9929–9936 (1994). 17. Fujiwara, S., Kakihara, H., Sakaguchi, K., and Imanaka, T.: Analysis of mutations in cyclodextrin glucanotransferase from Bacillus stearothermophilus which affect cyclization characteristics and thermostability. J. Bacteriol., 174, 7478– 7481 (1992). 18. van der Veen, B. A., Leemhius, H., Kralj, S., Uitdehaag, J. C. M., Dijkstra, B. W., and Dijkhuizen, L.: Hydrophobic amino acid residues in the acceptor binding site are main determinants for reaction mechanism and specificity of cyclodextrin-glycosyltransferase. J. Biol. Chem., 276, 44557–44562 (2001). 19. Leemhuis, H., Dijkstra, B. W., and Dijkhuizen, L.: Mutations converting cyclodextrin glycosyltransferase from a transglycosylase into a starch hydrolase. FEBS Lett., 514, 189– 192 (2002).