Kinetics of dextran-independent a-(1 fi 3)-glucan synthesis by Streptococcus sobrinus glucosyltransferase I Hideyuki Komatsu 1 , Yoshie Abe 1 , Kazuyuki Eguchi 1 , Hideki Matsuno 1 , Yu Matsuoka 1 , Takayuki Sadakane 1 , Tetsuyoshi Inoue 2 , Kazuhiro Fukui 2 and Takao Kodama 1 1 Department of Bioscience and Bioinformatics, Kyushu Institute of Technology, Iizuka, Japan 2 Department of Oral Microbiology, Okayama University Graduate School of Medicine, Dentistry, and Pharmaceutical Science, Japan Introduction Water-insoluble glucan, which is mainly composed of a-(1 fi 3)-glucan, enhances the colonization of cario- genic bacteria and promotes the formation of dental plaque on smooth tooth surfaces [1]. Two glucos- yltransferases (GTFs) or glucansucrases (GTF-S and GTF-I, EC 2.4.1.5) from mutans streptococci are responsible for the production of this polysaccharide [1,2]. GTF-S and GTF-I catalyze the synthesis of water-soluble a-(1 fi 6)-glucan and water-insoluble a-(1 fi 3)-glucan, respectively. They share highly homologous amino acid sequences and comprise the N-terminal glucansucrase domain (GSd) and the C-terminal glucan-binding domain (GBd) [3,4]. The GSd catalyzes glucosyl transfer from sucrose to low molecular mass acceptors such as water, isomaltose, and maltose [5]. On the other hand, the GBd binds to glucans such as dextran [6–8]. Unlike GTF-S, pre-existing dextran increases GTF-I activity [9,10]. Although truncation of the GBd of GTF-I by genetic engineering results in decreased Keywords enzyme kinetics; glucansucrase; glucosyltransferase; mutans streptococci; nigerooligosaccharide Correspondence H. Komatsu, Department of Bioscience and Bioinformatics, Kyushu Institute of Technology, Kawazu 680-4, Iizuka 820-8502, Japan Fax: +81 948 29 7801 Tel: +81 948 29 7845 E-mail: hide@bio.kyutech.ac.jp (Received 3 September 2010, revised 18 November 2010, accepted 25 November 2010) doi:10.1111/j.1742-4658.2010.07973.x Glucosyltransferase (GTF)-I from cariogenic Streptococcus sobrinus elon- gates the a-(1 fi 3)-linked glucose polymer branches on the primer dextran bound to the C-terminal glucan-binding domain. We investigated the GTF- I-catalyzed glucan synthesis reaction in the absence of the primer dextran. The time course of saccharide production during dextran-independent glu- can synthesis from sucrose was analyzed. Fructose and glucose were first produced by the sucrose hydrolysis. Leucrose was subsequently produced, followed by insoluble glucan [a-(1 fi 3)-linked glucose polymers] after a lag phase. High levels of intermediate nigerooligosaccharide series accumu- lation were characteristically not observed during the lag phase. The results from the enzymatic activity of the acceptor reaction for the nigerooligo- saccharide with a degree of polymerization of 2–6 and methyl a- D-gluco- pyranoside as a glucose analog indicate that the activity increased with an increase in the degree of polymerization. The production of insoluble glu- can was numerically simulated using the fourth-order Runge–Kutta method with the kinetic parameters estimated from the enzyme assay. The simulated time course provided a profile similar to that of experimental data. These results define the relationship between the kinetic properties of GTF-I and the time course of saccharide production. These results are discussed with respect to a mechanism that underlies efficient glucan synthesis. Abbreviations DP, degree of polymerization; GBd, glucan-binding domain; GS, glucan-binding domain-deficient glucosyltransferase-I; GSd, glucansucrase domain; GSGB, glucosyltransferase-I containing a full-length glucan-binding domain; GTF, glucosyltransferase. FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS 531 glucan synthesis [11–13], GBd-deficient GSd can still synthesize a-(1 fi 3)-glucan from sucrose in the absence of the pre-existing dextran (i.e. in a dextran- independent manner) [14], albeit at a lower rate. This suggests that the glucan synthesis reaction catalyzed by the GSd is a basic function of GTF-I. Although knowledge of the dextran-independent reaction is important for a comprehensive understand- ing of GTF-I function, its underlying mechanism remains incompletely defined. In this respect, two con- trasting mechanisms have been proposed for glucansuc- rases. One mechanism is proposed by Robyt et al. from pulse ⁄ chase experiments using [ 14 C]sucrose [15,16]. This mechanism involves processive elongation by which the enzyme continuously transfers glucose residues to only one acceptor molecule until the release of the final product. Another mechanism is suggested by Mooser et al. on the basis of the identification of the active site and the similarity with established glycosidase mecha- nisms [5,17]. The mechanism occurs via nonprocessive elongation, whereby the enzyme releases the products after each transfer of a glucose residue to the acceptor. Steady-state kinetic studies have been performed for other glucansucrases as well as fructansucrase [18–22]. In addition, the processes of insoluble glucan synthesis from sucrose, investigated with high-performance anion exchange chromatography analysis, are also reported [23,24]. However, it is not clear how changes in the intermediate oligosaccharides during the glucan syn- thesis can be interpreted with respect to the kinetic properties of enzymes. In other words, the relationship between the enzyme kinetic properties and the process of insoluble glucan synthesis remains unknown. To address these issues, we analyzed the process of dextran-independent insoluble glucan formation and examined the enzyme kinetic properties of the nigero- oligosaccharide acceptor reaction by using GBd-deficient GTF-I (GS) and GTF-I containing a full-length GBd (GSGB) from Streptococcus sobrinus 6715 (serotype g) (Fig. 1). A numerical simulation based on the enzyme kinetic parameters is in good agreement with the time course of dextran-independent insoluble glucan forma- tion. These results suggest that the dextran-independent synthesis of insoluble glucan by GTF-I proceeds via the nonprocessive elongation of a-(1 fi 3)-linked glu- cose polymers (nigerooligosaccharides). Results Functional characterization of GSGB We previously investigated the functional role of the GBd in dextran-dependent a-(1 fi 3)-glucan synthesis using GTF-I¢ (Asp85–Ile1256), which contains the GSd and the 246 N-terminal residues of the GBd (approxi- mately 50% of the GBd). The results indicate that the GBd enhances a-(1 fi 3)-branch formation on GBd- bound dextran [14]. In the present study, we character- ized newly prepared GSGB (Asp85–Asn1592) as an enzyme possessing the full-length GBd (Fig. 1). Figure 2A shows the dextran-dependent synthesis of insoluble glucan as monitored by light scattering. GSGB rapidly produced insoluble glucan in the pres- ence of dextran (filled squares in Fig. 2A), whereas GS did not (open squares in Fig. 2A). When glucosyl transfer activity (initial velocity) was measured as a function of dextran concentration, the activity of GSGB was highly dependent on dextran (filled circles in Fig. 2B); however, that of GS was nearly zero, inde- pendently of dextran concentration (open circles in Fig. 2B). On the other hand, there was no significant difference in sucrose hydrolysis activity in the absence of dextran between GS and GSGB. In addition, the sucrose hydrolysis activities of GS and GSGB as a function of sucrose concentration obeyed simple Michaelis–Menten kinetics, with similar k cat and K m values (data not shown). The k cat and K m values were as follows: GS, 11.4 ± 0.4 s )1 and 0.37 ± 0.07 mm; GSGB, 12.0 ± 0.4 s )1 and 0.29 ± 0.04 mm. These values are corroborated by our previous report [14]. Time course of dextran-independent glucan synthesis We examined the kinetics of insoluble glucan synthesis from sucrose by GS and GSGB in the absence of dex- tran (Fig. 3A,B). For this purpose, we incubated sucrose (50 mm) with GS or GSGB (0.4 lm) in the absence of dextran, and analyzed the products at arbi- trary time intervals. The time courses of the produc- tion of insoluble glucan and other sugars by GSGB were essentially the same as those for GS (Fig. 3). Fig. 1. Schematic structures of GTF-I and its variant proteins. The proteins used start at Asp85 of GTF-I and terminate at Ser1085 and Asn1592 for GS and GSGB, respectively. Both proteins contain an extra peptide, TMITNSSSVPG, from the multiple cloning site of pUC18 at their N-terminal ends. The black bars in the GBd indicate the ‘A’ repeat [6,7]. Kinetics of Streptococcus sobrinus GTF-I reaction H. Komatsu et al. 532 FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS In the initial phase ( 60 min), the concentration of fructose increased (open squares in Fig. 3) and the concentration of released glucose was lower than that of fructose (open circles in Fig. 3). As a result, the difference between the concentrations of fructose and glucose increased with time, suggesting that glucosyl transfer occurred. Leucrose, a known GTF-I product [25,26], was detected after 30 min, and its concentra- tion increased over time (open triangles in Fig. 3). Finally, the concentration of insoluble glucan increased with a lag time of 60 min (filled squares in Fig. 3). The glucosidic linkages of the insoluble glucan produced by GS or GSGB were analyzed with 13 C-NMR spectros- copy (Fig. S1). The spectra confirm that both of the glucans were a-(1 fi 3)-linked glucose polymers. The release of the sugars flattened off after 100 min, indi- cating sucrose depletion. After all of the sucrose (50 mm) had been consumed, 35 mm fructose, 12 mm glucose, 12 mm leucrose and 12 mm (glucose equivalent) insoluble glucan were pro- duced. The total concentrations of fructose and glu- cose were approximately 50 mm (resulting from 35 mm fructose and 12 mm leucrose) and 35–40 mm (resulting from 12 mm glucose, 12 mm leucrose, and 12 mm insoluble glucan), respectively. The fructose yield was equivalent to the initial amount of sucrose, whereas the glucose yield was significantly less. The 10–15 mm glucose that went undetected appears to have been incorporated into soluble nigerooligosaccharides. A B Fig. 2. Effect of dextran on the synthesis of insoluble glucan and glucosyl transfer activity by GS and GSGB. (A) Dextran-dependent insoluble glucan synthesis. The insoluble glucan synthesis reaction was initiated by adding GSGB (filled squares) or GS (open squares) to the sucrose solution in the presence of dextran. The amount of insoluble glucan was monitored by light scattering. The reaction mixture contained 50 n M GTFs, 50 lgÆmL )1 dextran T2000 and 100 m M sucrose in 10 mM Mops (pH 7.0). (B) Glucosyl transfer activity as a function of dextran concentration. The reaction mixture contained 10 n M GTFs, 50 mM sucrose, 100 mM NaCl, and 10 mM Mops (pH 6.8). Glucosyl transfer velocity was estimated as described in Experimental procedures. Filled and open circles indi- cate the GTF activity of GSGB and GS, respectively. A B Fig. 3. Saccharide production kinetics during the dextran-indepen- dent GTF-I reaction. (A) and (B) show the production of saccharide by GS and GSGB, respectively. Open squares, open circles, open triangles and filled squares indicate the production of fructose, glucose, leucrose, and insoluble glucan, respectively. The produc- tion of insoluble glucan is plotted as the glucose equivalent molar concentration. The starting solution contained 0.4 l M enzyme (GS or GSGB) and 50 m M sucrose in 10 mM potassium phosphate (pH 6.8). H. Komatsu et al. Kinetics of Streptococcus sobrinus GTF-I reaction FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS 533 To detect the soluble nigerooligosaccharides, soluble products were analyzed with HPLC, but no peak cor- responding to nigerooligosaccharides was detected (data not shown). In addition, TLC detected major spots of fructose, glucose, and leucrose; immobile spots and diffuse smears corresponding to a degree of poly- merization (DP) ‡ 4 were observed after 50 min (Fig. S2). These results suggest that the nigerooligo- saccharides were marginally accumulated and that they constituted a small proportion of the oligosaccharides with a wide range of DP values. Glucosyl transfer activity of GS and GSGB for nigerooligosaccharide acceptors Next, we estimated the glucosyl transfer activity of acceptor reactions for methyl a-d-glucopyranoside, nigerooligosaccharides (DP = 2–6), and leucrose (Fig. 4). Methyl a-d-glucopyranoside was used as a glucose analog that is not a substrate of hexokinase in a glucose assay system. The data were analyzed on the basis of the Michaelis–Menten equation; the maximum activity (k cat ) and Michaelis constant for acceptor (K m Acc ) values are listed in Table 1. Methyl a-d-glucopyr- anoside and nigerooligosaccharides served as the gluco- syl acceptors for GS and GSGB; in contrast, leucrose was not a glucosyl acceptor. The activity for nigerooligosaccharides was higher than that for methyl a-d-glucopyranoside. Furthermore, the nigerooligo- saccharides with higher DPs exhibited higher activity. The rank order of glucosyl transfer efficiencies (k cat ⁄ K m Acc ) was as follows: nigerohexaose > nigeropen- taose = nigerotetraose > nigerotriose = nigerose > methyl a- d-glucopyranoside. In addition, the kinetics of the acceptor reactions appeared to be similar between GS and GSGB. Kinetic simulation of GTF glucan synthesis To explain the time course of saccharide production during the synthesis of insoluble glucan, we assumed Scheme 1, which shows the intermediate products that are the probable acceptors for the GTF-I reaction in the absence of dextran. GTF-I was assumed to be capa- ble of catalyzing the following reactions: (a) the hydro- lysis of sucrose to glucose and fructose (reaction 1); (b) the transfer of glucose residues from sucrose to glucose and nigerooligosaccharide to produce higher- DP nigerooligosaccharide (reactions 2–7); and (c) the transfer of glucose residues from sucrose to fructose to produce leucrose (reaction 8). Here, we assumed that nigerooligosaccharides with DP ‡ 7 were insoluble, because the solubility of nigeroheptaose (DP = 7) was extraordinarily low (submillimolar level) at neu- tral pH. For numerical simulation, we made two assump- tions: (a) Michaelis–Menten kinetics for sucrose hydrolysis, as described above and indicated by Kraij et al. [21] for reuteransucrase; and (b) ping-pong bi-bi A B Fig. 4. GTF activity as a function of acceptor concentrations. The initial velocity of glucosyl transfer by GS (A) and GSGB (B) was measured in the presence of leucrose (open diamonds), methyl a- D-glucopyranoside (filled circles), nigerose (open circles), nigerotri- ose (filled squares), nigerotetraose (open squares), nigeropentaose (filled triangles), and nigerohexaose (open triangles). The lines indi- cate the curve fitted to the Michaelis–Menten equation to obtain k cat and K m Acc (Table 1), with the nonlinear regression program in the ORIGIN software package (v. 6.1J). The data from the GS nigero- pentaose reaction were analyzed by nonlinear least-squares analy- sis, but the parameters were not obtained with reasonable accuracy. Because the solubility of nigerooligosaccharides (DP ‡ 4) was very low, GTF activity could not be measured at higher con- centrations. Kinetics of Streptococcus sobrinus GTF-I reaction H. Komatsu et al. 534 FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS kinetics for glucosyl transfer from sucrose to an acceptor sugar, as indicated for other glucansucrases [5,18,19] and fructansucrase [22], and confirmed for the nigerotriose acceptor reaction by GS (Fig. S3). With these assumptions, the glucan synthesis process was numerically simulated on the basis of Scheme 1, with the experimentally estimated kinetic parameters (Table 2). The characteristic time course obtained from the experimental data was reproduced by this simulation (Fig. 5): (a) the fructose concentration increased at an appreciably higher rate than the glu- cose concentration, until all of the sucrose was con- sumed – the difference between the concentrations of fructose and glucose subsequently increased with time; (b) insoluble glucan (DP ‡ 7) production increased after a lag phase ( 50 min); (c) small quantities of each nigerooligosaccharide (DP = 2–6) were pro- duced (Fig. 5B); (d) leucrose production increased with a delay time ( 30 min); and (e) the final yield of saccharide was similar to that of the experimental data, even in terms of the total concentration of soluble nigerooligosaccharides (DP = 2–6). In other words, the difference between the total concentrations of fructose and glucose (50 mm and 35–40 mm, respectively) can be explained by this simulation. The simulation also indicates that the difference is attributable to the production of 10–15 mm soluble nigerooligosaccharides. Table 1. Kinetic parameters of the acceptor reaction. The parameters were obtained by nonlinear least-squares curve-fitting analysis for the data presented in Fig. 4. The k cat and K m Acc were estimated from the maximum activity and the Michaelis constant for the acceptor, respec- tively. ND, not determined. Acceptor GS GSGB k cat (s )1 ) K m Acc (mM) k cat ⁄ K m Acc (s )1 ÆmM )1 ) k cat (s )1 ) K m Acc (mM) k cat ⁄ K m Acc (s )1 ÆmM )1 ) Methyl a- D-glucopyranoside 39 ± 8 a 70 ± 31 0.56 ± 0.27 41 ± 25 78 ± 65 0.53 ± 0.54 Nigerose 53 ± 5 27 ± 5 2.0 ± 0.4 41 ± 4 12 ± 3 3.4 ± 0.9 Nigerotriose 57 ± 3 26 ± 3 2.2 ± 0.3 69 ± 9 33 ± 10 2.1 ± 0.7 Nigerotetraose 73 ± 8 16 ± 4 4.6 ± 1.2 65 ± 12 17 ± 8 3.8 ± 1.9 Nigeropentaose ND ND 59 ± 19 13 ± 9 4.5 ± 3.5 Nigerohexaose 53 ± 3 5 ± 1 11 ± 2 40 ± 3 6 ± 1 6.7 ± 1.2 a Errors of the estimates are indicated. Scheme 1. Kinetic model of dextran-independent insoluble glucan synthesis. H. Komatsu et al. Kinetics of Streptococcus sobrinus GTF-I reaction FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS 535 Discussion No significant difference was observed with respect to dextran-independent insoluble glucan synthesis and the nigerooligosaccharide acceptor reaction between GS and GSGB (Figs 3 and 4); therefore, the dextran- independent synthesis of a-(1 fi 3)-glucan is not an artificial reaction catalyzed by the deficient protein, but rather a basal reaction of GTF-I. In contrast, in the presence of dextran, the enzymatic activity of GSGB was much higher than that of GS (Fig. 2). Dextran- dependent glucan synthesis probably takes place via enhancement of the basal reaction of the GSd by the binding of dextran to the GBd. This result is consistent with the idea that the GBd plays an important role in dextran-dependent glucan synthesis [2,11–13]. The experimental time course of saccharide produc- tion agreed well with the simulation based on Scheme 1, using experimentally determined kinetic parameters. This agreement, in turn, supports the appropriateness of the model in Scheme 1; thus, our results provide a reasonable explanation for the dex- tran-independent glucan synthesis reaction. At the beginning of the reaction, sucrose is hydrolyzed to glu- cose and fructose. The released glucose subsequently serves as the glucosyl acceptor to form nigerose. The subsequent glucosyl transfers take place to nigerooligo- saccharides to form higher-DP nigerooligosaccharides. In this manner, insoluble glucan is formed by the elon- gation of nigerooligosaccharides. Efficient glucan synthesis, such that the accumulation of intermediate nigerooligosaccharides (DP = 2)6) is minimized, is explainable by the single-chain mecha- nism, whereby the enzyme ‘continuously’ elongates only one acceptor molecule until the release of the resulting glucan [15,16]. However, because the K m Acc values were relatively high (5–80 mm; Table 1) as compared with the concentrations of acceptors (glucose and nigerooligo- saccharides) produced under the experimental condi- tions, the enzyme tends to release the nigerooligo- saccharide during the reaction. This mode is consistent with nonprocessive elongation, whereby the enzyme releases the products after each transfer of a glucose residue to the acceptor. Hence, efficient insoluble glucan synthesis with minimal accumulation of intermediate nigerooligosaccharides is not achieved by continuous (or processive) elongation; it is, rather, achieved mainly through the enzymatic kinetic properties of GTF-I, the catalytic efficiency of which increases with an increase in the DP of nigerooligosaccharide (Table 1). In fact, we demonstrated that the enzyme kinetic model can simu- late the time course of insoluble glucan synthesis. Non- processive elongation was previously demonstrated for the formation of insoluble amylase-like polymers from sucrose by Neisseria polysaccharea amylosucrase [23]. Table 2. Kinetic parameters for the numerical simulation. Simula- tion parameters were estimated as described in Doc. S1. Reaction k cat (s )1 ) K m Suc (mM) K m Acc (mM) k (s )1 ÆmM )1 ) Sucrose hydrolysis 10 0.3 Acceptor reaction Glucose 40 2 70 Nigerose 50 2 20 Nigerotriose 60 2 30 Nigerotetraose 70 2 20 Nigeropentaose 60 2 10 Nigerohexaose 50 2 6 Fructose 0.02 A B Fig. 5. Kinetic simulation of dextran-independent glucan synthesis by GTF-I. The time course of saccharide production was simulated on the basis of the enzyme kinetic parameters listed in Table 2. (A) Time course of sucrose and major products. Molar concentra- tions of sucrose (filled circles), fructose (open squares), glucose (open circles) and leucrose (open triangles) are indicated. The con- centrations of soluble nigerooligosaccharides (open diamonds) and insoluble glucan (filled squares) are converted into glucose equivalent molar concentrations. Nigerooligosaccharides with DP ‡ 7 were defined as insoluble glucan. (B) Nigerooligosaccharide production. The amounts of nigerose (circles), nigerotriose (squares), nigerotetra- ose (diamonds), nigeropentaose (triangles) and nigerohexaose (inverted triangles) are indicated as molar concentrations. Kinetics of Streptococcus sobrinus GTF-I reaction H. Komatsu et al. 536 FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS The improved catalysis in the presence of increasing DP may be a common feature of glucansucrase. Additional studies are required to elucidate the dex- tran-dependent synthesis of glucan. As the affinity of the GBd of GTF-I for dextran is extraordinarily high (K m = 4.88 · 10 7 m )1 ) [27], the enzyme can hardly be dissociated from dextran. GTF-I may elongate the a-(1 fi 3)-glucose branches of dextran in a processive- like manner. Moulis et al. propose a semiprocessive mechanism of polymerization for Leuconostoc mesen- teroides NRRL B-512F dextransucrase and L. mesen- teroides NRRL B-1355 alternansucrase [24]. According to this mechanism, the glucan-binding domains at the C-terminal end of GTFs act as mediators of the shift between the processive and nonprocessive processes; GTF-I may otherwise randomly migrate on the dextran without dissociating during the reaction. Recently, high- speed atomic force microscopy demonstrated that Trichoderma reesei cellobiohydrolase I moves along crystalline cellulose in a processive manner [28]; there- fore, to understand the mechanism underlying the dextran-dependent GTF-I reaction, the dynamics of binding of GTF-I to dextran should be investigated with the use of single-molecule measurements, surface plas- mon resonance, and ⁄ or quartz crystal microbalances. This study defines, for the first time, the relationship between the kinetic properties of GTF and the time course of saccharide production. Owing to GTF’s kinetic property of improving activity in the presence of nigerooligosaccharides with increasing DPs, inter- mediate nigerooligosaccharides minimally accumulate during production, and insoluble glucan is efficiently produced. This feature may be favorable for dental plaque formation by mutans streptococci, because the efflux of the intermediate product may be diminished in an oral environment. Moreover, such kinetic analy- sis could also provide useful information regarding the GTF-I-catalyzed synthesis of a-(1 fi 3)-linked glucose polymers for industrial production of foods, pharma- ceuticals, and cosmetics [4,29]. Experimental procedures Plasmids Plasmid pGS [14] was used for the expression of GS (Asp85–Ser1085). To construct the GSGB (Asp85– Asn1592)-encoding plasmid pGSGB6R, an EcoRI–BglII fragment (1.5 kbp) encoding a part of the multiple cloning site and Asp85–Glu596 was isolated from pAB2 [6] and inserted into a BglII–EcoRI-digested pAB1 [6] fragment (3.4 kbp) encompassing the Asp597–Asn1592 coding region. Both plasmids contained pUC18-derived DNA (containing the lac promoter and the ampicillin resistance gene) and encoded an extra peptide, TMITNSSSVPG, from the multiple cloning site at the N-terminal end. Enzymes GS was prepared from Escherichia coli JM109 transformed with pGS. Protein expression and purification were carried out as described previously [14]. GSGB was expressed in E. coli JM109 transformed with pGSGB6R, and prepared with the same procedure as that used for GS. GSGB was further purified by gel filtration chromatography on a Toyopearl HW-55 column (TOSOH) (2.5 · 90 cm) pre- equilibrated in 0.5 m NaCl and 10 mm potassium phos- phate (pH 6.8). The protein concentration was determined from the absorbance at 280 nm, with a molar extinction coefficient calculated from the amino acid composition [30]. Preparation of a-(1 fi 3)-glucan GS was added to 300 mm sucrose solution containing 0.02% NaN 3 and 10 mm potassium phosphate (pH 6.8), giving a final concentration of 100 nm. The mixture was kept at room temperature until no more reducing sugars were produced (approximately 1 week). The resulting insol- uble material was collected by decanting and centrifugation at 600 g for 4 min, and was subsequently washed with water by suction filtration. The structure of the product was confirmed by 13 C-NMR spectroscopy. Preparation of nigerooligosaccharides The a-(1 fi 3)-glucan described above was subjected to mild acid hydrolysis (0.05 m H 2 SO 4 at 80 °C for 2 h) to release the small amount of fructose ( 1%) contained in the glucan. The fructose-free glucan was then subjected to limited acid hydrolysis (0.1 m H 2 SO 4 at 100 °C for 100 min) to produce nigerooligosaccharides. After the hydrolysis was stopped by neutralization with NaHCO 3 , the insoluble material was removed by centrifugation at 500 g for 4 min. Finally, the soluble fraction was applied to a column (4 · 22 cm) con- taining equivalent weights of activated carbon (Wako Pure Chemical, Osaka, Japan) and Celite (Wako Pure Chemical) (Fig. S4). The column was maintained at 15 °C, and was then washed with water until the glucose was completely eluted. The bound oligosaccharides were eluted with a gradi- ent of 0–8% butanol. The sugars were detected by use of the phenol ⁄ H 2 SO 4 method [31], and characterized by TLC. Measurement of glucosyl transfer velocity The glucosyl transfer rates were measured as described by Konishi et al. [14]. The initial velocities of all GTF reactions were measured for the first 4–8 min at 25 °C. The H. Komatsu et al. Kinetics of Streptococcus sobrinus GTF-I reaction FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS 537 reaction mixtures contained dextran, nigerooligosaccha- rides, methyl a-d-glucopyranoside and leucrose at various concentrations in 10 nm enzyme, 50 m m sucrose, 100 mm NaCl, and 10 mm Mops (pH 6.8). After the concentrations of the resultant glucose and fructose were measured, the extent of sucrose splitting was determined from the amount of fructose released, and the extent of glucosyl transfer was calculated by subtracting the amount of free glucose from the amount of free fructose. For kinetic analysis, the assay was carried out with at least four different concentrations of each acceptor [methyl a-d-glucopyranoside and nigerooligosaccharide (DP = 2–6)]. The kinetic parameters (velocity constant, k; Michaelis constant, K m ) were determined with the nonlinear regres- sion program in the origin software package (v. 6.1J) (OriginLab, Northampton, MA, USA). Light scattering The formation of insoluble glucan was monitored by light scattering at 90° in a thermostated cell (25 °C) at a wave- length of 350 nm. The reaction mixture contained 50 nm GTFs, 50 lgÆmL )1 dextran T2000 and 100 mm sucrose in 10 mm Mops (pH 7.0). Analysis of products during the glucan synthesis reaction GS or GSGB (0.4 lm) and 50 mm sucrose were incubated in 10 mm potassium phosphate (pH 6.8) at 25 °C. A suit- able volume of mixture was sampled at various time points, and the reaction was stopped by addition of NaOH at a final concentration of 10 mm. After the samples had been immediately centrifuged at 18,000 g for 10 min, the insolu- ble glucan was recovered as a precipitate and washed with water. The supernatants were neutralized with HCl. The precipitates were completely hydrolyzed in 1.2 m HCl at approximately 100 °C for 50 min, and this was followed by neutralization with 1.2 m NaHCO 3 . The hydrolysate was then subjected to the dinitrosalicylic acid method [32] to esti- mate the amount of insoluble glucan in terms of glucose. For the analysis of soluble sugars in the supernatants, the follow- ing methods were employed: (a) glucose and fructose concen- trations were determined with the enzymatic assay [33]; and (b) leucrose and nigerooligosaccharide were detected by TLC, and their concentrations were determined by HPLC. Simulation of the glucan synthesis reaction The synthesis of insoluble glucan was numerically analyzed on the basis of Scheme 1. We assumed Michaelis–Menten kinetics for sucrose hydrolysis [14,21], and ping-pong bi-bi kinetics for the glucosyl transfer from sucrose to an acceptor sugar [5,18,19,22]. Using the quasi-steady-state assumption, we obtained a set of 10 differential equations that describe the time course of saccharide production. The fourth-order Runge)Kutta method with a 3.6-min step was used to solve the 10 differential equations, with the kinetic parameters estimated from the experimental data using visual basic for applications in Microsoft Excel (http:// chemeng.on.coocan.jp). All parameters used are listed in Table 2. The details of these differential equations and the simulation parameter are described in Doc. S1. 13 C-NMR analysis To inhibit base-catalyzed reactions, the insoluble glucan was dissolved at 60 mgÆ mL )1 in 0.5 m NaOD in D 2 O con- taining 3.5 mgÆmL )1 NaBD 4 . The 13 C-NMR analyses were performed on a JEOL JNM-A500 spectrometer (Center for Instrumental Analysis, Kyushu Institute of Technology). Spectra were recorded at 125.65 MHz at room temperature, with an acquisition time of 0.9667 s and 8000 scan accumu- lations. Chemical shifts are expressed in parts per million, with 3-trimethylsilyl-1-propanesulfonic acid sodium salt as a reference. Peaks were assigned according to the method outlined in Colson et al. [34]. HPLC Leucrose and nigerooligosaccharide concentrations were measured by HPLC with a YMC-pack polyamine II column (4.6 · 250 mm) (YMC, Kyoto, Japan). The column was maintained at 30 °C, and the flow rate was 0.5 mLÆmin )1 . The eluents were water ⁄ acetonitrile (25 : 75, v ⁄ v) and water ⁄ acetonitrile (30 : 70, v ⁄ v) for the analysis of disaccha- rides (leucrose and nigerose) and nigerooligosaccharides (DP = 3–5), respectively. Carbohydrates were detected with a differential refractometer. Concentrations were calculated from the peak area of the refractive index, using a calibration curve. The sample for leucrose analysis was incubated with yeast invertase (0.2 mgÆmL )1 )at35°C for 30 min to con- sume sucrose before injection, because leucrose was eluted at 7.7 min, which was relatively close to sucrose (7.4 min). TLC After the samples were concentrated on a centrifugal evap- orator, they were separated by TLC, with three ascents on Silica Gel 70 plates (Wako Pure Chemical) in 15 : 3 : 4 (v ⁄ v ⁄ v) 1-butanol ⁄ pyridine ⁄ water [35]. The sugars were detected by spraying the plates with H 2 SO 4 , and then heat- ing at 120 °C for 5 min. Acknowledgements We thank H. Sakamoto (Kyushu Institute of Technol- ogy) for providing access to the apparatus for Kinetics of Streptococcus sobrinus GTF-I reaction H. Komatsu et al. 538 FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS biochemical experiments. We appreciate the assistance provided by M. Ikeguchi, Y. Fujita and T. Kawauchi for nigerooligosaccharide preparation and separation. This work was supported in part by grants-in-aid for Scientific Research (B) 13557161 and (C) 13671904 (to K. Fukui), (C) 13680744 (to T. Kodama) and a grant- in-aid for Young Scientists (B) 13080537 (to H. Koma- tsu) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. The initial parts of this work were performed at the Institute for Materials Chemistry and Engineering, Kyushu University. References 1 Loesche WJ (1986) Role of Streptococcus mutans in human dental decay. Microbiol Rev 50, 353–380. 2 Monchois V, Willemot RM & Monsan P (1999) Glu- cansucrase: mechanism of action and structure–function relationships. FEMS Microbiol Rev 23, 131–151. 3 Simpson CL, Giffard PM & Jacques NA (1995) Strep- tococcus salivarius ATCC 25975 possesses at least two genes coding for primer-independent glucosyltrans- ferases. Infect Immun 63, 609–621. 4 van Hijum SA, Kralj S, Ozimek LK, Dijkhuizen L & van Geel-Schutten IG (2006) Structure–function rela- tionships of glucansucrase and fructansucrase enzymes from lactic acid bacteria. Microbiol Mol Biol Rev 70, 157–176. 5 Mooser G (1992) Glycosidases and glycosyltransferases. Enzymes 20, 187–231. 6 Abo H, Matsumura T, Kodama T, Ohta H, Fukui K, Kato K & Kagawa H (1991) Peptide sequences for sucrose splitting and glucan binding within Streptococ- cus sobrinus glucosyltransferase (water-insoluble glucan synthetase). J Bacteriol 173, 989–996. 7 Ferretti JJ, Gilpin ML & Russell RR (1987) Nucleotide sequence of a glucosyltransferase gene from Strepto- coccus sobrinus MFe28. J Bacteriol 169, 4271–4278. 8 Wong C, Hefta SA, Paxton RJ, Shively JE & Mooser G (1990) Size and subdomain architecture of the glucan-binding domain of sucrose:3-alpha-D-glucosyl- transferase from Streptococcus sobrinus. Infect Immun 58, 2165–2170. 9 Fukui K, Moriyama T, Miyake Y, Mizutani K & Tanaka O (1982) Purification and properties of glucosyltransferase responsible for water-insoluble glucan synthesis from Streptococcus mutans. Infect Im- mun 37, 1–9. 10 Koga T, Sato S, Inoue M, Takeuchi K, Furuta T & Hamada S (1983) Role of primers in glucan synthesis by glucosyltransferases from Streptococcus mutans strain OMZ176. J Gen Microbiol 129, 751–754. 11 Kato C, Nakano Y, Lis M & Kuramitsu HK (1990) Carboxyl-terminal deletion analysis of the Streptococcus mutans glucosyltransferase-I enzyme. FEMS Microbiol Lett 72, 290–302. 12 Monchois V, Arguello-Morales M & Russell RRB (1999) Isolation of an active catalytic core of Streptococcus downei MFe28 GTF-I glucosyltransferase. J Bacteriol 181, 2290–2292. 13 Kingston KB, Allen DM & Jacques NA (2002) Role of the C-terminal YG repeats of the primer-dependent streptococcal glucosyltransferase, GtfJ, in binding to dextran and mutan. Microbiology 148, 549–558. 14 Konishi N, Torii Y, Yamamoto T, Miyagi A, Ohta H, Fukui K, Hanamoto S, Matsuno H, Komatsu H, Kod- ama T et al. (1999) Structure and enzymatic properties of genetically truncated forms of the water-insoluble glucan-synthesizing glucosyltransferase from Streptococ- cus sobrinus. J Biochem (Tokyo) 126, 287–295. 15 Robyt JF, Kimble BK & Walseth TF (1974) The mech- anism of dextransucrase action. Direction of dextran biosynthesis. Arch Biochem Biophys 165, 634–640. 16 Robyt JF, Yoon SH & Mukerjea R (2008) Dextransucr- ase and the mechanism for dextran biosynthesis. Carbo- hydr Res 343, 3039–3048. 17 Mooser G, Hefta SA, Paxton RJ, Shively JE & Lee TD (1991) Isolation and sequence of an active-site peptide containing a catalytic aspartic acid from two Strepto- coccus sobrinus alpha-glucosyltransferases. J Biol Chem 266, 8916–8922. 18 Kobayashi M & Matsuda K (1978) Inhibition of dextran synthesis by glucoamylase and endodextranase. Carbohydr Res 66, 277–288. 19 Mooser G, Shur D, Lyou M & Watanabe C (1985) Kinetic studies on dextransucrase from the cariogenic oral bacterium Streptococcus mutans. J Biol Chem 260, 6907–6915. 20 Mukasa H, Shimamura A & Tsumori H (2000) Nige- rooligosaccharide acceptor reaction of Streptococcus sobrinus glucosyltransferase GTF-I. Carbohydr Res 326, 98–103. 21 Kralj S, van Geel-Schutten GH, van der Maarel MJ & Dijkhuizen L (2004) Biochemical and molecular charac- terization of Lactobacillus reuteri 121 reuteransucrase. Microbiology 150, 2099–2112. 22 Chambert R, Treboul G & Dedonder R (1974) Kinetic studies of levansucrase of Bacillus subtilis. Eur J Biochem 41, 285–300. 23 Albenne C, Skov LK, Mirza O, Gajhede M, Feller G, D’Amico S, Andre ´ G, Potocki-Ve ´ rone ` se G, van der Veen BA, Monsan P et al. (2004) Molecular basis of the amylose-like polymer formation catalyzed by Neisseria polysaccharea amylosucrase. J Biol Chem 279, 726–734. 24 Moulis C, Joucla G, Harrison D, Fabre E, Potocki- Veronese G, Monsan P & Remaud-Simeon M (2006) Understanding the polymerization mechanism of H. Komatsu et al. Kinetics of Streptococcus sobrinus GTF-I reaction FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS 539 glycoside-hydrolase family 70 glucansucrases. J Biol Chem 281, 31254–31267. 25 Monchois V, Vignon M, Escalier PC, Svensson B & Russell RR (2000) Involvement of Gln937 of Strepto- coccus downei GTF-I glucansucrase in transition-state stabilization. Eur J Biochem 267, 4127–4136. 26 Monchois V, Vignon M & Russell RR (2000) Mutagen- esis of asp-569 of glucosyltransferase I glucansucrase modulates glucan and oligosaccharide synthesis. Appl Environ Microbiol 66, 1923–1927. 27 Komatsu H, Katayama M, Sawada M, Hirata Y, Mori M, Inoue T, Fukui K, Fukada H & Kodama T (2007) Thermodynamics of the binding of the C-terminal repeat domain of Streptococcus sobrinus glucosyltrans- ferase-I to dextran. Biochemistry 46, 8436–8444. 28 Igarashi K, Koivula A, Wada M, Kimura S, Penttila ¨ M & Samejima M (2009) High speed atomic force micros- copy visualizes processive movement of Trichoderma reesei cellobiohydrolase I on crystalline cellulose. J Biol Chem 284, 36186–36190. 29 Hellmuth H, Wittrock S, Kralj S, Dijkhuizen L, Hofer B & Seibel J (2008) Engineering the glucansucrase GTFR enzyme reaction and glycosidic bond specificity: toward tailor-made polymer and oligosaccharide prod- ucts. Biochemistry 47, 6678–6684. 30 Gill SC & von Hippel PH (1982) Calculation of protein extinction coefficients from amino acid sequence data. Anal Biochem 182, 319–326. 31 Dubois M, Gilles KA, Hamilton JK, Robers P & Smith F (1956) Colorimetric method for determination of sugars and related substances. Anal Chem 28, 350– 356. 32 Chaplin MF (1986) Monosaccharides. In Carbohydrate Analysis: Practical Approach (Chaplin MF & Kennedy JF eds), pp 1–36. IRL Press, Oxford. 33 Schmidt FH (1961) Die enzymatische Bestimmung von Glucose und Fructose nebeneinander. Klin Wochenschr 39, 1244–1247. 34 Colson P, Jarrell HC, Lamberts BL & Smith IC (1979) Determination, by carbon-13 nuclear magnetic reso- nance spectroscopy, of the composition of glucans syn- thesized by enzymes of the cariogenic organism Streptococcus mutans. Carbohydr Res 71, 265–272. 35 Mukasa H, Tsumori H & Shimamura A (2001) Dextran acceptor reaction of Streptococcus sobrinus glucosyl- transferase GTF-I as revealed by using uniformly 13 C-labeled sucrose. Carbohydr Res 333, 19–26. Supporting information The following supplementary material is available: Fig. S1. 13 C-NMR analysis of the insoluble glucan produced by GS and GSGB. Fig. S2. TLC analysis of saccharide production by GS and GSGB. Fig. S3. Double reciprocal plot of nigerotriose accep- tor reaction of GS. Fig. S4. Separation of nigerooligosaccharides by acti- vated carbon chromatography. Doc. S1. Simulation of the glucan synthesis reaction. This supplementary material can be found in the online version of this article. Please note: As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Kinetics of Streptococcus sobrinus GTF-I reaction H. Komatsu et al. 540 FEBS Journal 278 (2011) 531–540 ª 2010 The Authors Journal compilation ª 2010 FEBS . Kinetics of dextran-independent a-(1 fi 3)-glucan synthesis by Streptococcus sobrinus glucosyltransferase I Hideyuki Komatsu 1 , Yoshie Abe 1 , Kazuyuki. Owing to GTF’s kinetic property of improving activity in the presence of nigerooligosaccharides with increasing DPs, inter- mediate nigerooligosaccharides