- Email: [email protected]
Synthesis of levan in water-miscible organic solvents
Journal of Biotechnology 114 (2004) 209–217
Synthesis of levan in water-miscible organic solvents E. Castillo, A. López-Mungu´ıa∗ Departamento de Ingenier´ıa Celular y Biocatálisis, Instituto de Biotecnolog´ıa, UNAM, Apartado Postal 510-3, Cuernavaca, Morelos 62271, Mexico Received 17 October 2003; received in revised form 16 March 2004; accepted 8 June 2004
Abstract The synthesis of levan using a levansucrase from a strain of Bacillus subtilis was studied in the presence of the water-miscible solvents: acetone, acetonitrile and 2-methyl-2-propanol (2M2P). It was found that while the enzyme activity is only slightly affected by acetone and acetonitrile, 2M2P has an activating effect increasing the total activity 35% in 40–50% (v/v) 2M2P solutions at 30 ◦ C. The enzyme is highly stable in water at 30 ◦ C; however, incubation in the presence of 15 and 50% (v/v) 2M2P reduced the half-life time to 23.6 and 1.8 days, respectively. This effect is reversed in 83% 2M2P, where a half-life time of 11.8 days is observed. The presence of 2M2P in the system increases the transfer/hydrolysis ratio of levansucrase. As the reaction proceeds with 10% (w/v) sucrose in 50/50 water/2M2P sucrose is converted to levan and an aqueous two-phase system (2M2P/Levan) is formed and more sucrose can be added in a fed batch mode. It is shown that high molecular weight levan is obtained as an hydrogel and may be easily recovered from the reaction medium. However, when high initial sucrose concentrations (40% (w/v) in 50/50 water/2M2P) are used, an aqueous two-phase system (2M2P/sucrose) is induce, where the synthesized levan has a similar molecular weight distribution as in water and remains in solution. © 2004 Elsevier B.V. All rights reserved. Keywords: Levansucrase; Bacillus subtilis; Glycosyltransferase; Levan; Organic solvents
1. Introduction In spite of the intensive research devoted to the function of enzymes in organic synthesis published during the last two decades, little attention has been paid to the synthesis of biopolymers and oligosaccharides in these media using glycosyltransferases. This is most certainly due to the high water solubility of the substrates and products of this kind of enzymes. Fructosyltransferases, a particular type of glycosyl∗ Corresponding author. Tel.: +52-55-56227637; fax: +52-777-3172388. E-mail address: [email protected] (A. L´opez-Mungu´ıa).
transferases, have been applied for the synthesis of a wide variety of fructooligosaccharides, fructosides and biopolymers (Hidaka et al., 1991, Ritsema and Smeekens, 2003). These enzymes use sucrose as substrate, transferring the fructose residue to acceptors coping the energy derived from the glycosidic bond to the formation of fructosyl linkages with a specificity that is dependent on the source of the enzyme. Fructosyltransferases are able to produce mainly two kinds of polymer: levan, a -2,6-fructose polymer, which is the product of the enzyme levansucrase, and inulin, a -2,1-fructose polymer, which is the product of inulosucrase. Recently, these products have been the subject of intensive research due to their interest-
0168-1656/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2004.06.003
210
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
ing properties as prebiotics and soluble fiber (Niness, 1999). In addition to the synthesis of fructose polymers, which are the result of the fructosyl moieties transfer to the growing fructan chain, levansucrases carry out the synthesis of low molecular weight oligosaccharides when acceptors molecules such as lactose or maltose are present in the reaction medium (Perez-Oseguera et al., 1996). Levansucrases are reported behaving essentially as transferases, however, they also possess hydrolytic activity, resulting from the transfer of the fructosyl moiety to a molecule of water. The selectivity between transferase and hydrolytic activities has been reported as a function of different factors such as the source of the enzyme (Kim et al., 1998) the temperature (Ohtsuka et al., 1992, Jang et al., 2001), the initial sucrose concentration (Korakli et al., 2003) and the presence of water-miscible organic solvents in the reaction medium (Chambert and Petit-Glatron, 1989). Enzymes have been broadly used for biotransformations in organic solvents, where they display interesting properties such as enhanced stability, alteration of their chemoselectivity or shifts in the thermodynamic equilibrium (Dordick, 1989). However, the main interest in studying the behavior of glycosyltransferases in organic solvents has been the modification of the transfer/hydrolysis ratio of the reaction. Chambert and Petit-Glatron (1989) report that Bacillus subtilis levansucrase displays only transferase activity in reactions carried out in the presence of 70% (v/v) of acetonitrile in the aqueous medium. However, Kim et al. (1998) did not found any effect when using a levansucrase from Rahnella aqualitis in the presence of 20% (v/v) of various water-miscible organic solvents. Moreover, these authors found that one limitation with R. aqualitis levansucrase is the instability of the enzyme in the presence of high concentrations of water-miscible solvents such as methanol at temperatures higher than 20 ◦ C. Indeed, in a recent paper they report a drastic decrease of levansucrase activity during enzyme incubation in methanol/buffer (60% (v/v)) at 20 ◦ C (Kim et al., 2000). Using a different approach, the behaviour of a Streptococcus mutans glucosyltransferase in the presence of water-miscible organic solvents was recently reported (Meulenbeld and Hartmans, 2000). These authors found that in reactions carried out using 15% of bis-2-methoxyethyl ether as co-solvent
and hydroxylated aromatic molecules as acceptors, the transglycosylation yields are enhanced four-fold, compared to reactions carried out in water, mainly due to a 12-fold increase in acceptor solubility. An alternative advantage of the presence of organic solvents in the synthesis of polysaccharides might be the direct recuperation of the synthesized product. In fact, the industrial recovery of dextran and levan as well as other biopolymers produced by biological processes, involve their precipitation by the addition of solvents such as low molecular alcohols to the reaction medium at the end of the synthesis in order to modify the dielectric constant and precipitate the polymers (Ohtsuka et al., 1992). As this precipitation practice involves an extra step, the development of a process involving a simultaneous enzymatic reaction and precipitation of the biopolymer would decrease the volumes of alcohol required and allow a continuous recovery of the product. In addition, the continuous removal of the product may also avoid or decrease its inhibitory effect on levansucrase activity (Steinberg et al., 2002). In this work, an enhanced method for the continuous recovering of the levan from the reaction medium as a result of the presence of an organic solvent is proposed. The effect of water-miscible organic solvents on levansucrase produced by B. subtilis (BS-LVS), on the transfer/hydrolysis ratio and on molecular weight distribution of the product is also reported.
2. Materials and methods 2.1. Materials Sucrose, fructose, glucose, polyethylenglycol (PEG) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). All high purity solvents and buffer salts were purchased from JT Baker, Co. (Phillipsburg, NJ, USA). All solvents were dried extensively with 4 Å molecular sieves (Aldrich Chemical Co. Inc., Milwaukee, WI, USA). All chemicals were of analytical grade. 2.2. Enzyme preparation Levansucrase from B. subtilis (BS-LVS) was produced from a strain derived from B. subtilis Marburg 168 (kindly provided by Dr. Farnando Valle from
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
IBT-UNAM), designated with genotype npr, apr, CmR , degU32 (Hy), which has the property of overexpressing the levansucrase gene. Fermentations were carried out in a 2 L fernbach flasks with 450 mL of fermentation media, as already described. Once the stationary phase was reached the culture is centrifugated and the extracellular enzyme precipitated with 25% (w/v) PEG (Mw: 1500) and recovered in phosphate buffer 50 mM, pH 6.0 after centrifugation (Perez-Oseguera et al., 1996; Canedo et al., 1999). 2.3. Total levansucrase activity Initial levansucrase reaction rates was determined measuring the release of reducing sugars from a 100 g L−1 sucrose solution in phosphate buffer 50 mM, pH 6.0, during the first 15 min using the dinitrosalicylic acid (DNS) method (Summer and Howell, 1935) taking glucose as standard. One total levansucrase activity unit (LVSU) is defined as the amount of enzyme liberating the equivalent of 1 mol of glucose/min. Reactions were carried out in 2.5 mL aqueous solutions (phosphate buffer 50 mM, pH 6.0) containing 1.0 LVSU mL−1 , 100 g L−1 sucrose and various concentrations of the organic solvents: acetone, acetonitrile and 2-methyl-2-propanol (2M2P), at 30 or 37 ◦ C. In order to differentiate among the transferase and hydrolase activities, glucose and fructose produced were measured by HPLC (see below). In general, statistical data is derived from experiments run in triplicates. 2.4. Levansucrase stability Levansucrase stability was evaluated incubating at 30 and 40 ◦ C 2.5 mL of phosphate buffer solutions (50 mM) with various concentrations of organic solvents, sampling at different periods of time and diluting the sample enough to measure the residual activity in water.
211
sample to a Waters HPLC system equipped with a 600E system controller and a 410 differential refractometer. The sugars were separated in a 125 Å (10 m, 3.9 mm × 300 mm) Carbohydrate Analysis column (Waters Corp. Milford MA) using a 75:25 mixture of acetonitrile–water as the mobile phase at a flow rate of 1.0 mL min−1 . The temperature both in the column and detector was regulated at 35 ◦ C. In order to calculate the transferase and hydrolytic activity, it was considered that all free fructose is the product of hydrolysis, so after substrating the amount of free fructose from the total glucose liberated, the residual glucose is equivalent to the transferase activity. 2.6. Semi-continuous production of levan in 2M2P The semi-continuous production of levan polymer was carried out in 15 mL assay tubes containing 5 mL of a 100 g L−1 sucrose solution in 50% (v/v) 2M2P and 1 U mL−1 of BS-LVS activity at 30 ◦ C. After 24 h of reaction a second load of substrate was added to restablish 100 g L−1 sucrose in the medium and the procedure repeated after 48 and 72 h. After each load of sucrose, total levansucrase activity was determined and related to its initial value. After 96 h, the residual concentration of sucrose was evaluated by HPLC and the final conversion of substrate was calculated. The levan produced was recovered by centrifugation and lyophilized; the final levan concentration was determined gravimetrically. The molar mass distribution of levan was estimated by gel-permeation chromatography in a Waters 600E HPLC system controller (Waters Corp. Milford MA) equipped with a Waters 410 differential refractometer, a serial set of ultrahydrogel columns (UG 500 and linear, Waters) and eluted at a flow rate of 0.9 mL min−1 with a 100 mM sodium nitrate solution. The columns were calibrated by a standard calibration method using pullulans of Aureobasidium pullulan as standard.
2.5. Determination of the transfer/hydrolysis ratio
3. Results and discussion
The transferase and hydrolytic activities were evaluated at different concentrations of the water-miscible 2M2P. For this purpose, after 20 h of reaction, the concentration of free glucose, free fructose and residual sucrose were measured by direct injection of the
3.1. Effect of organic solvents on BS-LEV performance The total BS-LVS activity was measured in the presence of up to 50% (v/v) of acetone, acetonitrile
212
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
Fig. 1. Levansucrase activity at 30 ◦ C as a function of acetone, acetonitrile or 2-methyl-2-propanol concentration. Reactions were carried out with 100 g L−1 initial sucrose concentration.
and 2M2P at 30 ◦ C. The three solvents are able to solubilize at least 100 g L−1 of sucrose in 50% water mixtures and are common solvents in enzyme engineering. In these conditions a single and stable homogeneous phase is formed during the early stages of the reaction. In Fig. 1, it may be observed that in the presence of acetone and acetonitrile, total levansucrase activity decreased slightly and linearly with increasing solvent concentration. As a result, in the presence of 50% (v/v) of either acetone or acetonitrile, levansucrase expresses only 75% of its activity in water. However, it is interesting to note that in the presence of low concentrations of 2M2P after a slight decrease of around 10% in total LVS activity, at concentrations higher than 30% (v/v) there is an activating effect, LVS reaching the highest activity at 40–50% 2M2P (v/v), 35% higher than the activity measured in water. A similar effect was found by Chambert and Petit-Glatron (1989), who reported a three-fold increase in total activity for a B. subtilis levansucrase in the presence of acetonitrile 60% (v/v). In contrast, Kim et al. (2000) reported a drastic loss of activity for a levansucrase from R. aquatilis, when assayed in a medium containing high methanol concentrations at temperatures above 20 ◦ C, while Girard and Legoy (1999) found an important loss of activity with dextransucrase from L. mesenteroides B-512, when evaluated at 30 ◦ C in concentrations above 20% (v/v) of acetone, acetonitrile or dimethylsulfoxide. A more extensive study of the influence of 2M2P as co-solvent on total levansucrase activity was car-
ried out exploring concentrations as high as 95% (v/v). The experiments of this study were also performed at 37 ◦ C where the enzyme has higher activity but poor water stability. Total levansucrase activity in water at 37 ◦ C is 1.5 times higher than the activity at 30 ◦ C. In Fig. 2, the relative levansucrase reaction rate as a function of 2M2P evaluated at 30 and 37 ◦ C is shown. It may be observed that the already described activating effect observed at 30 ◦ C above 20% 2M2P and up to 50% is also observed at 37 ◦ C. Interestingly, at solvent concentrations higher than 50% (v/v) the activating effect of 2M2P is lost and the initial reaction rate is reduced to the rate measured in water. Again, a similar behaviour was observed at both temperatures. When solvent concentration reached 90% a drastic
Fig. 2. Effect of 2M2P concentration on the activity of BS-LEV measured at 30 and 37 ◦ C and 100 g L−1 of initial sucrose concentration (100% of relative activity is 1 and 1.5 U mL−1 at 30 and 37 ◦ C, respectively).
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
213
decrease in total activity was observed. However, it is important to point out that at high solvent concentration, sucrose has a poor solubility, so the decrease in the measured initial rates may be influenced by the low sucrose concentration, also reducing the substrate protecting effect. 3.2. Effect of organic solvents on BS-LVS stability In order to study this phenomenon in detail, the effect of 2M2P on BS-LVS stability was studied. For this purpose, levansucrase solutions containing 1 LVSU mL−1 were incubated in the presence of different 2M2P concentrations at 30 ◦ C. After different incubation times samples were withdrawn for initial reaction rate determinations. The effect of solvent concentration on the stability of the enzyme is shown in Fig. 3, where it may be observed that the enzyme is stable for several hours in the presence of up to 15% 2M2P. However, a decrease in BS-LVS activity is observed when the enzyme is incubated at concentrations of 2M2P higher than 30% (v/v). Nevertheless, a better stability of the enzyme was observed at 83% 2M2P. Due to the enhanced stability of the enzyme observed at high concentrations of 2M2P, the experiments were also performed at 40 ◦ C for concentrations of 2M2P up to 94%. The half-life time of the enzyme for both temperatures is reported in Fig. 4 as a function of solvent concentration. In both figures, a decrease in t1/2 of BS-LVS is observed
Fig. 4. Half-life time of levansucrase as a function of 2M2P concentration calculated at (A) 30 ◦ C and (B) 40 ◦ C.
when the enzyme is incubated in 2M2P concentrations up to 75% (v/v). Obviously, as shown in Fig. 4B for the half-life times of BS-LVS, the deactivating effect of the solvent is higher at 40 ◦ C. Nevertheless, in both cases the effect is reduced at high 2M2P concentrations. Stronger intramolecular forces such as hydrogen bonds and salt bridges and lower flexibility of proteins in organic solvents explain these results. Indeed, more rigid proteins are reported to be more stable than flexible proteins (Colombo and Carrea, 2002).
Fig. 3. Levansucrase stability as a function of 2M2P concentration at 30 ◦ C (initial activity = 1 U mL−1 ).
214
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
Fig. 5. Profile of glucose, fructose and transfer/hydrolysis ratio, during levan synthesis in the presence of different 2M2P concentrations. Sugars were determined by HPLC after 20 h of reaction with 100 g L−1 initial sucrose concentration.
3.3. The effect of solvent concentration on the transfer/hydrolysis ratio The use of aqueous-miscible organic solvents can decrease the water activity (aw ) of a given system through water interaction with the organic phase. Considering that the hydrolytic properties of glycosyltransferases are dependent on the capacity of the enzyme to transfer the glycosyl moiety to water, a low aw in a system should limit the availability of water as an acceptor for the glycosyl transfer, limiting the hydrolytic process. The transfer/hydrolysis ratio of the BS-LVS was evaluated at different concentrations of 2M2P. For these experiments, the specific transfer and hydrolysis
products were quantified by HPLC. In Fig. 5, the profile of free glucose and fructose after 20 h of reaction is shown, free fructose being the product of sucrose hydrolysis and free glucose the product of both activities. As expected, transferase activities increase as 2M2P concentration is increased with the concomitant decrease in free fructose. Thus, the transfer capacity of BS-LVS is increased from 45% in reactions carried out in buffer to almost 85% in reactions carried out in 50 % (v/v) 2M2P. At higher 2M2P concentrations the transfer capacity remains constant. These results confirm that the capacity of the BS-LVS for transferring fructose moieties to water is considerably reduced as water becomes less available. In the design of an
Fig. 6. Visual aspect of levan synthesis in three systems. Tubes positioned up-side-down showing: (A) a monophasic 400 g L−1 sucrose batch reaction carried out in water, (B) a biphasic liquid/liquid 400 g L−1 sucrose batch reaction carried out in 50% (v/v) 2M2P; and (C) a monophasic 400 g L−1 sucrose fed-batch reaction in 50% (v/v) 2M2P, where levan is found as an hydrogel.
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
oligosaccharide synthesis process from sucrose, it is clear that the use of 2M2P concentrations higher than 50 % is recommended. 3.4. Continuous production of levan in 2M2P The experiments performed up to now refer exclusively to initial reaction rates. However, when the reaction proceeds for longer periods, particularly at high solvent concentrations, it was observed that two phases were formed, an upper 2M2P rich phase and a lower levan rich phase. Depending on reaction conditions
215
this lower levan phase becomes a very viscous hydrogel which could be easily recovered (see Fig. 6C). Considering that a usual non-inhibitory substrate concentration in BS-LVS synthesis is 100 g L−1 (Euzenat et al., 1998) fed-batch reactions at this sucrose concentration were carried out. Experiments were performed in 50% (v/v) 2M2P where the enzyme shows high activity, sucrose is still soluble and levan could be recovered as hydrogel in the lower viscous phase. The reaction was carried out at 30 ◦ C with 1 LVSU mL−1 . After 24 h of reaction, a second load of substrate was added to re-establish 100 g L−1 sucrose in the medium
Fig. 7. Effect of 2M2P and reaction mode on levan molar mass distribution (M). (A) Batch reaction carried out in water with 400 g L−1 initial sucrose concentration. (B) Batch reaction carried out in 50% (v/v) 2M2P with 400 g L−1 initial sucrose concentration. (C) Fed-batch reaction carried out in 50% (v/v) 2M2P adjusting the sucrose concentration to 100 g L−1 in each step.
216
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
and the procedure was repeated at 48 and 72 h. After each load of sucrose, total levansucrase activity was determined and related to its initial value. It was found that although 30% of the initial BS-LVS activity was lost during the first batch, in the subsequent reactions the activity remained constant, most probably due to the combined stabilizing effect of both solvent and substrate. After the fourth batch, 78% of the total 400 g L−1 of sucrose added had been converted, and although more than 90% of levan produced was recovered, the medium became difficult to handle for further sucrose addition. After a few hours of reaction, solvent and levan induced the formation of a two phase system: no levan was found in the upper 2M2P-rich phase, which according to HPLC analysis contains small amounts of fructose and glucose. The levan rich phase can be easily recovered by sedimentation or centrifugation, converting the process in an interesting alternative for the simultaneous production and recovery of the biopolymer. Most of the glucose, a by-product of the reaction, is found in the polymer phase. The industrial recovery of glycopolysaccharides by precipitation, with ethanol or iso-propanol, requires volumes of three–four times the volume of the aqueous phase: it is worthy noting that in this work four times less solvent is required to precipitate and wash the levan produced, demonstrating the feasibility of the semi-continuous production. Levan produced in the already described fed-batch process was compared with levans produced in batch experiments carried out in water and in a 50% 2M2P (v/v) medium with an initial sucrose concentration of 400 g L−1 Fig. 6A and B, respectively). In this case, the high sucrose concentration in 50% (v/v) 2M2P induced a two phase system, changing the behaviour of the process as may be observed in Fig. 6B. Finally, in Fig. 7, the molecular weight distribution of levan produced in the three conditions is reported. As it is may be observed in the figure, levan profile in the range of 102 to 104 Da is similar for the three conditions assayed. However, in the experiments carried out in the presence of 2M2P, levans having a molar weight between 104 and 106 Da were obtained, the synthesis of levan of high molecular weight being more abundant in the fed-batch reaction. It is important to point out that, the high initial sucrose concentration (400 g L−1 ) used in the batch experiment with 2M2P induced the formation of the two phases instantaneously: the lower
phase containing most of the sucrose and the enzyme, and the upper 2M2P phase. Reaction conditions in the lower phase resemble those of the batch reaction carried out in buffer. On the contrary, in the fed-batch experiment, the lower sucrose concentration allowed to work in a homogeneous system, at least in the early stages of the reaction, favoring the synthesis of high molecular weight levan, and later on, the continuous separation of the biopolymer once two phases formation conditions are reached (Fig. 6C). It may be concluded that levan may be synthesized in water soluble solvents. In particular, 2M2P increases levansucrase activity at concentrations of 40–50% (v/v), improve the transfer ability of the enzyme and favors the synthesis of a high molecular weight polymers, but most interestingly, it allows the continuous recovery of the product when sucrose is added in a fed-batch mode. This reaction system may be extended to processes such as dextran or inulin synthesis carried out with glycosyltransferases.
Acknowledgements This Project was partially financed by CONACYT 40609-Z. The authors thank Dr. Carlos Peña for molecular weight determinations and Fernando Gonzalez for technical assistance.
References Canedo, M., Jiménez-Estrada, M., Cassani, J., Lopez-Munguia, A., 1999. Production of maltosylfructose (erlose) with levansucrase from Bacillus subtilis. Biocatal. Biotransformation 16, 475– 485. Chambert, R., Petit-Glatron, M.F., 1989. Study of the effect of organic solvents on the synthesis of levan. Carbohydr. Res. 191, 117–123. Colombo, G., Carrea, G., 2002. Modeling enzyme reactivity in organic solvents and water through computer simulations. J. Biotechnol. 96, 23–33. Dordick, J.S., 1989. Enzymatic catalysis in monophasic organic solvents. Enzym. Microb. Technol. 11, 194–210. Euzenat, O., Guibert, A., Combes, D., 1998. Production and purification of Bacillus subtilis C4 levansucrase—kinetic characterization of the enzyme. Ann. N.Y. Acad. Sci. 864, 288–294. Girard, E., Legoy, M.D., 1999. Activity and stability of dextransucrase from Leuconostoc mesenteroides NRRL B-512F
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217 in the presence of organic solvents. Enzym. Microb. Technol. 24, 425–432. Hidaka, H., Hirayama, M., Yamada, K., 1991. Fructooligosaccharides enzymatic preparation and biofunctions. J. Carbohydr. Chem. 10, 509–522. Jang, K.H., Song, K.B., Park, B.S., Kim, C.H., Chung, B.H., Choue, R.W., Lee, K.S., Lee, C., Chun, U.H., Rhee, S.K., 2001. Levan production by use of the recombinant levansucrase immobilized on titanium-activated magnetite. Process Biochem. 37, 339–343. Kim, M.G., Kim, C.H., Lee, J.S., Song, K.B., Rhee, S.K., 2000. Synthesis of methyl beta-d-fructoside catalyzed by levansucrase from Rahnella aquatilis. Enzym. Microb. Technol. 27, 641–646. Kim, M.G., Seo, J.W., Song, K.B., Kim, C.H., Chung, B.H., Rhee, S.K., 1998. Levan and fructosyl derivatives formation by a recombinant levansucrase from Rahnella aquatilis. Biotechnol. Lett. 20, 333–336. Korakli, M., Pavlovic, M., Gänzle, M.G., Vogel, R.F., 2003. Exopolysaccharide and kestose production by Lactobacillus sanfranciscensis LTH2590. Appl. Environ. Microbiol. 69, 2073–2079.
217
Meulenbeld, G.H., Hartmans, S., 2000. Transglycosylation by Streptococcus mutans GS-5 glucosyltransferase-d: acceptor specificity and engineering of reaction conditions. Biotechnol. Bioeng. 70, 363–369. Niness, K.R., 1999. Inulin and oligofructose: what are they? J. Nutr. 129, 1402S–1406S. Ohtsuka, K., Hino, S., Fukushima, T., Ozawa, O., Kanematsu, T., Uchida, T., 1992. Characterization of levansucrase from Rahnella aquatilis JCM-1638. Biosci. Biotechnol. Biochem. 56, 1373–1377. Perez-Oseguera, M.A., Guereca, L., Lopez-Munguia, A., 1996. Properties of levansucrase from Bacillus circulans. Appl. Microbiol. Biotechnol. 45, 465–471. Ritsema, T., Smeekens, S., 2003. Fructans: beneficial for plants and humans. Curr. Opin. Plant Biol. 6, 223–230. Steinberg, D., Rozen, R., Bromshteym, M., Zaks, B., Gedalia, I., Bachrach, G., 2002. Regulation of fructosyltransferase activity by carbohydrates, in solution and immobilized on hydroxyapatite surfaces. Carbohydr. Res. 337, 701– 710. Summer, J.B., Howell, S.F., 1935. A method for determination of saccharase activity. J. Biol. Chem. 108, 51–54.
Synthesis of levan in water-miscible organic solvents E. Castillo, A. López-Mungu´ıa∗ Departamento de Ingenier´ıa Celular y Biocatálisis, Instituto de Biotecnolog´ıa, UNAM, Apartado Postal 510-3, Cuernavaca, Morelos 62271, Mexico Received 17 October 2003; received in revised form 16 March 2004; accepted 8 June 2004
Abstract The synthesis of levan using a levansucrase from a strain of Bacillus subtilis was studied in the presence of the water-miscible solvents: acetone, acetonitrile and 2-methyl-2-propanol (2M2P). It was found that while the enzyme activity is only slightly affected by acetone and acetonitrile, 2M2P has an activating effect increasing the total activity 35% in 40–50% (v/v) 2M2P solutions at 30 ◦ C. The enzyme is highly stable in water at 30 ◦ C; however, incubation in the presence of 15 and 50% (v/v) 2M2P reduced the half-life time to 23.6 and 1.8 days, respectively. This effect is reversed in 83% 2M2P, where a half-life time of 11.8 days is observed. The presence of 2M2P in the system increases the transfer/hydrolysis ratio of levansucrase. As the reaction proceeds with 10% (w/v) sucrose in 50/50 water/2M2P sucrose is converted to levan and an aqueous two-phase system (2M2P/Levan) is formed and more sucrose can be added in a fed batch mode. It is shown that high molecular weight levan is obtained as an hydrogel and may be easily recovered from the reaction medium. However, when high initial sucrose concentrations (40% (w/v) in 50/50 water/2M2P) are used, an aqueous two-phase system (2M2P/sucrose) is induce, where the synthesized levan has a similar molecular weight distribution as in water and remains in solution. © 2004 Elsevier B.V. All rights reserved. Keywords: Levansucrase; Bacillus subtilis; Glycosyltransferase; Levan; Organic solvents
1. Introduction In spite of the intensive research devoted to the function of enzymes in organic synthesis published during the last two decades, little attention has been paid to the synthesis of biopolymers and oligosaccharides in these media using glycosyltransferases. This is most certainly due to the high water solubility of the substrates and products of this kind of enzymes. Fructosyltransferases, a particular type of glycosyl∗ Corresponding author. Tel.: +52-55-56227637; fax: +52-777-3172388. E-mail address: [email protected] (A. L´opez-Mungu´ıa).
transferases, have been applied for the synthesis of a wide variety of fructooligosaccharides, fructosides and biopolymers (Hidaka et al., 1991, Ritsema and Smeekens, 2003). These enzymes use sucrose as substrate, transferring the fructose residue to acceptors coping the energy derived from the glycosidic bond to the formation of fructosyl linkages with a specificity that is dependent on the source of the enzyme. Fructosyltransferases are able to produce mainly two kinds of polymer: levan, a -2,6-fructose polymer, which is the product of the enzyme levansucrase, and inulin, a -2,1-fructose polymer, which is the product of inulosucrase. Recently, these products have been the subject of intensive research due to their interest-
0168-1656/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jbiotec.2004.06.003
210
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
ing properties as prebiotics and soluble fiber (Niness, 1999). In addition to the synthesis of fructose polymers, which are the result of the fructosyl moieties transfer to the growing fructan chain, levansucrases carry out the synthesis of low molecular weight oligosaccharides when acceptors molecules such as lactose or maltose are present in the reaction medium (Perez-Oseguera et al., 1996). Levansucrases are reported behaving essentially as transferases, however, they also possess hydrolytic activity, resulting from the transfer of the fructosyl moiety to a molecule of water. The selectivity between transferase and hydrolytic activities has been reported as a function of different factors such as the source of the enzyme (Kim et al., 1998) the temperature (Ohtsuka et al., 1992, Jang et al., 2001), the initial sucrose concentration (Korakli et al., 2003) and the presence of water-miscible organic solvents in the reaction medium (Chambert and Petit-Glatron, 1989). Enzymes have been broadly used for biotransformations in organic solvents, where they display interesting properties such as enhanced stability, alteration of their chemoselectivity or shifts in the thermodynamic equilibrium (Dordick, 1989). However, the main interest in studying the behavior of glycosyltransferases in organic solvents has been the modification of the transfer/hydrolysis ratio of the reaction. Chambert and Petit-Glatron (1989) report that Bacillus subtilis levansucrase displays only transferase activity in reactions carried out in the presence of 70% (v/v) of acetonitrile in the aqueous medium. However, Kim et al. (1998) did not found any effect when using a levansucrase from Rahnella aqualitis in the presence of 20% (v/v) of various water-miscible organic solvents. Moreover, these authors found that one limitation with R. aqualitis levansucrase is the instability of the enzyme in the presence of high concentrations of water-miscible solvents such as methanol at temperatures higher than 20 ◦ C. Indeed, in a recent paper they report a drastic decrease of levansucrase activity during enzyme incubation in methanol/buffer (60% (v/v)) at 20 ◦ C (Kim et al., 2000). Using a different approach, the behaviour of a Streptococcus mutans glucosyltransferase in the presence of water-miscible organic solvents was recently reported (Meulenbeld and Hartmans, 2000). These authors found that in reactions carried out using 15% of bis-2-methoxyethyl ether as co-solvent
and hydroxylated aromatic molecules as acceptors, the transglycosylation yields are enhanced four-fold, compared to reactions carried out in water, mainly due to a 12-fold increase in acceptor solubility. An alternative advantage of the presence of organic solvents in the synthesis of polysaccharides might be the direct recuperation of the synthesized product. In fact, the industrial recovery of dextran and levan as well as other biopolymers produced by biological processes, involve their precipitation by the addition of solvents such as low molecular alcohols to the reaction medium at the end of the synthesis in order to modify the dielectric constant and precipitate the polymers (Ohtsuka et al., 1992). As this precipitation practice involves an extra step, the development of a process involving a simultaneous enzymatic reaction and precipitation of the biopolymer would decrease the volumes of alcohol required and allow a continuous recovery of the product. In addition, the continuous removal of the product may also avoid or decrease its inhibitory effect on levansucrase activity (Steinberg et al., 2002). In this work, an enhanced method for the continuous recovering of the levan from the reaction medium as a result of the presence of an organic solvent is proposed. The effect of water-miscible organic solvents on levansucrase produced by B. subtilis (BS-LVS), on the transfer/hydrolysis ratio and on molecular weight distribution of the product is also reported.
2. Materials and methods 2.1. Materials Sucrose, fructose, glucose, polyethylenglycol (PEG) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). All high purity solvents and buffer salts were purchased from JT Baker, Co. (Phillipsburg, NJ, USA). All solvents were dried extensively with 4 Å molecular sieves (Aldrich Chemical Co. Inc., Milwaukee, WI, USA). All chemicals were of analytical grade. 2.2. Enzyme preparation Levansucrase from B. subtilis (BS-LVS) was produced from a strain derived from B. subtilis Marburg 168 (kindly provided by Dr. Farnando Valle from
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
IBT-UNAM), designated with genotype npr, apr, CmR , degU32 (Hy), which has the property of overexpressing the levansucrase gene. Fermentations were carried out in a 2 L fernbach flasks with 450 mL of fermentation media, as already described. Once the stationary phase was reached the culture is centrifugated and the extracellular enzyme precipitated with 25% (w/v) PEG (Mw: 1500) and recovered in phosphate buffer 50 mM, pH 6.0 after centrifugation (Perez-Oseguera et al., 1996; Canedo et al., 1999). 2.3. Total levansucrase activity Initial levansucrase reaction rates was determined measuring the release of reducing sugars from a 100 g L−1 sucrose solution in phosphate buffer 50 mM, pH 6.0, during the first 15 min using the dinitrosalicylic acid (DNS) method (Summer and Howell, 1935) taking glucose as standard. One total levansucrase activity unit (LVSU) is defined as the amount of enzyme liberating the equivalent of 1 mol of glucose/min. Reactions were carried out in 2.5 mL aqueous solutions (phosphate buffer 50 mM, pH 6.0) containing 1.0 LVSU mL−1 , 100 g L−1 sucrose and various concentrations of the organic solvents: acetone, acetonitrile and 2-methyl-2-propanol (2M2P), at 30 or 37 ◦ C. In order to differentiate among the transferase and hydrolase activities, glucose and fructose produced were measured by HPLC (see below). In general, statistical data is derived from experiments run in triplicates. 2.4. Levansucrase stability Levansucrase stability was evaluated incubating at 30 and 40 ◦ C 2.5 mL of phosphate buffer solutions (50 mM) with various concentrations of organic solvents, sampling at different periods of time and diluting the sample enough to measure the residual activity in water.
211
sample to a Waters HPLC system equipped with a 600E system controller and a 410 differential refractometer. The sugars were separated in a 125 Å (10 m, 3.9 mm × 300 mm) Carbohydrate Analysis column (Waters Corp. Milford MA) using a 75:25 mixture of acetonitrile–water as the mobile phase at a flow rate of 1.0 mL min−1 . The temperature both in the column and detector was regulated at 35 ◦ C. In order to calculate the transferase and hydrolytic activity, it was considered that all free fructose is the product of hydrolysis, so after substrating the amount of free fructose from the total glucose liberated, the residual glucose is equivalent to the transferase activity. 2.6. Semi-continuous production of levan in 2M2P The semi-continuous production of levan polymer was carried out in 15 mL assay tubes containing 5 mL of a 100 g L−1 sucrose solution in 50% (v/v) 2M2P and 1 U mL−1 of BS-LVS activity at 30 ◦ C. After 24 h of reaction a second load of substrate was added to restablish 100 g L−1 sucrose in the medium and the procedure repeated after 48 and 72 h. After each load of sucrose, total levansucrase activity was determined and related to its initial value. After 96 h, the residual concentration of sucrose was evaluated by HPLC and the final conversion of substrate was calculated. The levan produced was recovered by centrifugation and lyophilized; the final levan concentration was determined gravimetrically. The molar mass distribution of levan was estimated by gel-permeation chromatography in a Waters 600E HPLC system controller (Waters Corp. Milford MA) equipped with a Waters 410 differential refractometer, a serial set of ultrahydrogel columns (UG 500 and linear, Waters) and eluted at a flow rate of 0.9 mL min−1 with a 100 mM sodium nitrate solution. The columns were calibrated by a standard calibration method using pullulans of Aureobasidium pullulan as standard.
2.5. Determination of the transfer/hydrolysis ratio
3. Results and discussion
The transferase and hydrolytic activities were evaluated at different concentrations of the water-miscible 2M2P. For this purpose, after 20 h of reaction, the concentration of free glucose, free fructose and residual sucrose were measured by direct injection of the
3.1. Effect of organic solvents on BS-LEV performance The total BS-LVS activity was measured in the presence of up to 50% (v/v) of acetone, acetonitrile
212
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
Fig. 1. Levansucrase activity at 30 ◦ C as a function of acetone, acetonitrile or 2-methyl-2-propanol concentration. Reactions were carried out with 100 g L−1 initial sucrose concentration.
and 2M2P at 30 ◦ C. The three solvents are able to solubilize at least 100 g L−1 of sucrose in 50% water mixtures and are common solvents in enzyme engineering. In these conditions a single and stable homogeneous phase is formed during the early stages of the reaction. In Fig. 1, it may be observed that in the presence of acetone and acetonitrile, total levansucrase activity decreased slightly and linearly with increasing solvent concentration. As a result, in the presence of 50% (v/v) of either acetone or acetonitrile, levansucrase expresses only 75% of its activity in water. However, it is interesting to note that in the presence of low concentrations of 2M2P after a slight decrease of around 10% in total LVS activity, at concentrations higher than 30% (v/v) there is an activating effect, LVS reaching the highest activity at 40–50% 2M2P (v/v), 35% higher than the activity measured in water. A similar effect was found by Chambert and Petit-Glatron (1989), who reported a three-fold increase in total activity for a B. subtilis levansucrase in the presence of acetonitrile 60% (v/v). In contrast, Kim et al. (2000) reported a drastic loss of activity for a levansucrase from R. aquatilis, when assayed in a medium containing high methanol concentrations at temperatures above 20 ◦ C, while Girard and Legoy (1999) found an important loss of activity with dextransucrase from L. mesenteroides B-512, when evaluated at 30 ◦ C in concentrations above 20% (v/v) of acetone, acetonitrile or dimethylsulfoxide. A more extensive study of the influence of 2M2P as co-solvent on total levansucrase activity was car-
ried out exploring concentrations as high as 95% (v/v). The experiments of this study were also performed at 37 ◦ C where the enzyme has higher activity but poor water stability. Total levansucrase activity in water at 37 ◦ C is 1.5 times higher than the activity at 30 ◦ C. In Fig. 2, the relative levansucrase reaction rate as a function of 2M2P evaluated at 30 and 37 ◦ C is shown. It may be observed that the already described activating effect observed at 30 ◦ C above 20% 2M2P and up to 50% is also observed at 37 ◦ C. Interestingly, at solvent concentrations higher than 50% (v/v) the activating effect of 2M2P is lost and the initial reaction rate is reduced to the rate measured in water. Again, a similar behaviour was observed at both temperatures. When solvent concentration reached 90% a drastic
Fig. 2. Effect of 2M2P concentration on the activity of BS-LEV measured at 30 and 37 ◦ C and 100 g L−1 of initial sucrose concentration (100% of relative activity is 1 and 1.5 U mL−1 at 30 and 37 ◦ C, respectively).
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
213
decrease in total activity was observed. However, it is important to point out that at high solvent concentration, sucrose has a poor solubility, so the decrease in the measured initial rates may be influenced by the low sucrose concentration, also reducing the substrate protecting effect. 3.2. Effect of organic solvents on BS-LVS stability In order to study this phenomenon in detail, the effect of 2M2P on BS-LVS stability was studied. For this purpose, levansucrase solutions containing 1 LVSU mL−1 were incubated in the presence of different 2M2P concentrations at 30 ◦ C. After different incubation times samples were withdrawn for initial reaction rate determinations. The effect of solvent concentration on the stability of the enzyme is shown in Fig. 3, where it may be observed that the enzyme is stable for several hours in the presence of up to 15% 2M2P. However, a decrease in BS-LVS activity is observed when the enzyme is incubated at concentrations of 2M2P higher than 30% (v/v). Nevertheless, a better stability of the enzyme was observed at 83% 2M2P. Due to the enhanced stability of the enzyme observed at high concentrations of 2M2P, the experiments were also performed at 40 ◦ C for concentrations of 2M2P up to 94%. The half-life time of the enzyme for both temperatures is reported in Fig. 4 as a function of solvent concentration. In both figures, a decrease in t1/2 of BS-LVS is observed
Fig. 4. Half-life time of levansucrase as a function of 2M2P concentration calculated at (A) 30 ◦ C and (B) 40 ◦ C.
when the enzyme is incubated in 2M2P concentrations up to 75% (v/v). Obviously, as shown in Fig. 4B for the half-life times of BS-LVS, the deactivating effect of the solvent is higher at 40 ◦ C. Nevertheless, in both cases the effect is reduced at high 2M2P concentrations. Stronger intramolecular forces such as hydrogen bonds and salt bridges and lower flexibility of proteins in organic solvents explain these results. Indeed, more rigid proteins are reported to be more stable than flexible proteins (Colombo and Carrea, 2002).
Fig. 3. Levansucrase stability as a function of 2M2P concentration at 30 ◦ C (initial activity = 1 U mL−1 ).
214
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
Fig. 5. Profile of glucose, fructose and transfer/hydrolysis ratio, during levan synthesis in the presence of different 2M2P concentrations. Sugars were determined by HPLC after 20 h of reaction with 100 g L−1 initial sucrose concentration.
3.3. The effect of solvent concentration on the transfer/hydrolysis ratio The use of aqueous-miscible organic solvents can decrease the water activity (aw ) of a given system through water interaction with the organic phase. Considering that the hydrolytic properties of glycosyltransferases are dependent on the capacity of the enzyme to transfer the glycosyl moiety to water, a low aw in a system should limit the availability of water as an acceptor for the glycosyl transfer, limiting the hydrolytic process. The transfer/hydrolysis ratio of the BS-LVS was evaluated at different concentrations of 2M2P. For these experiments, the specific transfer and hydrolysis
products were quantified by HPLC. In Fig. 5, the profile of free glucose and fructose after 20 h of reaction is shown, free fructose being the product of sucrose hydrolysis and free glucose the product of both activities. As expected, transferase activities increase as 2M2P concentration is increased with the concomitant decrease in free fructose. Thus, the transfer capacity of BS-LVS is increased from 45% in reactions carried out in buffer to almost 85% in reactions carried out in 50 % (v/v) 2M2P. At higher 2M2P concentrations the transfer capacity remains constant. These results confirm that the capacity of the BS-LVS for transferring fructose moieties to water is considerably reduced as water becomes less available. In the design of an
Fig. 6. Visual aspect of levan synthesis in three systems. Tubes positioned up-side-down showing: (A) a monophasic 400 g L−1 sucrose batch reaction carried out in water, (B) a biphasic liquid/liquid 400 g L−1 sucrose batch reaction carried out in 50% (v/v) 2M2P; and (C) a monophasic 400 g L−1 sucrose fed-batch reaction in 50% (v/v) 2M2P, where levan is found as an hydrogel.
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
oligosaccharide synthesis process from sucrose, it is clear that the use of 2M2P concentrations higher than 50 % is recommended. 3.4. Continuous production of levan in 2M2P The experiments performed up to now refer exclusively to initial reaction rates. However, when the reaction proceeds for longer periods, particularly at high solvent concentrations, it was observed that two phases were formed, an upper 2M2P rich phase and a lower levan rich phase. Depending on reaction conditions
215
this lower levan phase becomes a very viscous hydrogel which could be easily recovered (see Fig. 6C). Considering that a usual non-inhibitory substrate concentration in BS-LVS synthesis is 100 g L−1 (Euzenat et al., 1998) fed-batch reactions at this sucrose concentration were carried out. Experiments were performed in 50% (v/v) 2M2P where the enzyme shows high activity, sucrose is still soluble and levan could be recovered as hydrogel in the lower viscous phase. The reaction was carried out at 30 ◦ C with 1 LVSU mL−1 . After 24 h of reaction, a second load of substrate was added to re-establish 100 g L−1 sucrose in the medium
Fig. 7. Effect of 2M2P and reaction mode on levan molar mass distribution (M). (A) Batch reaction carried out in water with 400 g L−1 initial sucrose concentration. (B) Batch reaction carried out in 50% (v/v) 2M2P with 400 g L−1 initial sucrose concentration. (C) Fed-batch reaction carried out in 50% (v/v) 2M2P adjusting the sucrose concentration to 100 g L−1 in each step.
216
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217
and the procedure was repeated at 48 and 72 h. After each load of sucrose, total levansucrase activity was determined and related to its initial value. It was found that although 30% of the initial BS-LVS activity was lost during the first batch, in the subsequent reactions the activity remained constant, most probably due to the combined stabilizing effect of both solvent and substrate. After the fourth batch, 78% of the total 400 g L−1 of sucrose added had been converted, and although more than 90% of levan produced was recovered, the medium became difficult to handle for further sucrose addition. After a few hours of reaction, solvent and levan induced the formation of a two phase system: no levan was found in the upper 2M2P-rich phase, which according to HPLC analysis contains small amounts of fructose and glucose. The levan rich phase can be easily recovered by sedimentation or centrifugation, converting the process in an interesting alternative for the simultaneous production and recovery of the biopolymer. Most of the glucose, a by-product of the reaction, is found in the polymer phase. The industrial recovery of glycopolysaccharides by precipitation, with ethanol or iso-propanol, requires volumes of three–four times the volume of the aqueous phase: it is worthy noting that in this work four times less solvent is required to precipitate and wash the levan produced, demonstrating the feasibility of the semi-continuous production. Levan produced in the already described fed-batch process was compared with levans produced in batch experiments carried out in water and in a 50% 2M2P (v/v) medium with an initial sucrose concentration of 400 g L−1 Fig. 6A and B, respectively). In this case, the high sucrose concentration in 50% (v/v) 2M2P induced a two phase system, changing the behaviour of the process as may be observed in Fig. 6B. Finally, in Fig. 7, the molecular weight distribution of levan produced in the three conditions is reported. As it is may be observed in the figure, levan profile in the range of 102 to 104 Da is similar for the three conditions assayed. However, in the experiments carried out in the presence of 2M2P, levans having a molar weight between 104 and 106 Da were obtained, the synthesis of levan of high molecular weight being more abundant in the fed-batch reaction. It is important to point out that, the high initial sucrose concentration (400 g L−1 ) used in the batch experiment with 2M2P induced the formation of the two phases instantaneously: the lower
phase containing most of the sucrose and the enzyme, and the upper 2M2P phase. Reaction conditions in the lower phase resemble those of the batch reaction carried out in buffer. On the contrary, in the fed-batch experiment, the lower sucrose concentration allowed to work in a homogeneous system, at least in the early stages of the reaction, favoring the synthesis of high molecular weight levan, and later on, the continuous separation of the biopolymer once two phases formation conditions are reached (Fig. 6C). It may be concluded that levan may be synthesized in water soluble solvents. In particular, 2M2P increases levansucrase activity at concentrations of 40–50% (v/v), improve the transfer ability of the enzyme and favors the synthesis of a high molecular weight polymers, but most interestingly, it allows the continuous recovery of the product when sucrose is added in a fed-batch mode. This reaction system may be extended to processes such as dextran or inulin synthesis carried out with glycosyltransferases.
Acknowledgements This Project was partially financed by CONACYT 40609-Z. The authors thank Dr. Carlos Peña for molecular weight determinations and Fernando Gonzalez for technical assistance.
References Canedo, M., Jiménez-Estrada, M., Cassani, J., Lopez-Munguia, A., 1999. Production of maltosylfructose (erlose) with levansucrase from Bacillus subtilis. Biocatal. Biotransformation 16, 475– 485. Chambert, R., Petit-Glatron, M.F., 1989. Study of the effect of organic solvents on the synthesis of levan. Carbohydr. Res. 191, 117–123. Colombo, G., Carrea, G., 2002. Modeling enzyme reactivity in organic solvents and water through computer simulations. J. Biotechnol. 96, 23–33. Dordick, J.S., 1989. Enzymatic catalysis in monophasic organic solvents. Enzym. Microb. Technol. 11, 194–210. Euzenat, O., Guibert, A., Combes, D., 1998. Production and purification of Bacillus subtilis C4 levansucrase—kinetic characterization of the enzyme. Ann. N.Y. Acad. Sci. 864, 288–294. Girard, E., Legoy, M.D., 1999. Activity and stability of dextransucrase from Leuconostoc mesenteroides NRRL B-512F
E. Castillo, A. L´opez-Mungu´ıa / Journal of Biotechnology 114 (2004) 209–217 in the presence of organic solvents. Enzym. Microb. Technol. 24, 425–432. Hidaka, H., Hirayama, M., Yamada, K., 1991. Fructooligosaccharides enzymatic preparation and biofunctions. J. Carbohydr. Chem. 10, 509–522. Jang, K.H., Song, K.B., Park, B.S., Kim, C.H., Chung, B.H., Choue, R.W., Lee, K.S., Lee, C., Chun, U.H., Rhee, S.K., 2001. Levan production by use of the recombinant levansucrase immobilized on titanium-activated magnetite. Process Biochem. 37, 339–343. Kim, M.G., Kim, C.H., Lee, J.S., Song, K.B., Rhee, S.K., 2000. Synthesis of methyl beta-d-fructoside catalyzed by levansucrase from Rahnella aquatilis. Enzym. Microb. Technol. 27, 641–646. Kim, M.G., Seo, J.W., Song, K.B., Kim, C.H., Chung, B.H., Rhee, S.K., 1998. Levan and fructosyl derivatives formation by a recombinant levansucrase from Rahnella aquatilis. Biotechnol. Lett. 20, 333–336. Korakli, M., Pavlovic, M., Gänzle, M.G., Vogel, R.F., 2003. Exopolysaccharide and kestose production by Lactobacillus sanfranciscensis LTH2590. Appl. Environ. Microbiol. 69, 2073–2079.
217
Meulenbeld, G.H., Hartmans, S., 2000. Transglycosylation by Streptococcus mutans GS-5 glucosyltransferase-d: acceptor specificity and engineering of reaction conditions. Biotechnol. Bioeng. 70, 363–369. Niness, K.R., 1999. Inulin and oligofructose: what are they? J. Nutr. 129, 1402S–1406S. Ohtsuka, K., Hino, S., Fukushima, T., Ozawa, O., Kanematsu, T., Uchida, T., 1992. Characterization of levansucrase from Rahnella aquatilis JCM-1638. Biosci. Biotechnol. Biochem. 56, 1373–1377. Perez-Oseguera, M.A., Guereca, L., Lopez-Munguia, A., 1996. Properties of levansucrase from Bacillus circulans. Appl. Microbiol. Biotechnol. 45, 465–471. Ritsema, T., Smeekens, S., 2003. Fructans: beneficial for plants and humans. Curr. Opin. Plant Biol. 6, 223–230. Steinberg, D., Rozen, R., Bromshteym, M., Zaks, B., Gedalia, I., Bachrach, G., 2002. Regulation of fructosyltransferase activity by carbohydrates, in solution and immobilized on hydroxyapatite surfaces. Carbohydr. Res. 337, 701– 710. Summer, J.B., Howell, S.F., 1935. A method for determination of saccharase activity. J. Biol. Chem. 108, 51–54.