Coimmobilization of Gluconolactonase with Glucose Oxidase for

Biomacromolecules 2002, 3, 625-631

625

Coimmobilization of Gluconolactonase with Glucose Oxidase for Improvement in Kinetic Property of Enzymatically Induced Volume Collapse in Ionic Gels Kazuyoshi Ogawa, Toshiaki Nakajima-Kambe, Tadaatsu Nakahara, and Etsuo Kokufuta* Institute of Applied Biochemistry, University of Tsukuba, Tsukuba, Ibaraki 305-8565, Japan Received January 27, 2002; Revised Manuscript Received March 21, 2002

The object of this paper is to provide an enzymatic means to attain faster swelling or shrinking kinetics of polyelectrolyte gels that undergo volume phase transition as an immobilized enzyme reaction sets in. For this, we studied the coimmobilization of gluconolactonase (GL) with glucose oxidase (GOD). A gel used was in the shape of a small cylinder (several hundred micrometers in diameter) and composed of a lightly cross-linked copolymer of N-isopropylacrylamide and acrylic acid. GL was isolated from Aspergillus niger and purified about 100-fold. It was found that in a substrate solution containing glucose, the gel with the coimmobilized GL and GOD shrinks very rapidly. The shrinking rate was identical to that of the enzymefree gel that undergoes a shrinking transition in response to a sudden pH change of the outer medium from 7 to 5. This indicates the rate-limiting step in the shrinking process to be diffusion of the networks, but not the enzyme reaction. In the gel with singly immobilized GOD, a very slow shrinking was observed because the process is governed by the enzyme reaction. These results were discussed in full in connection with an enzymatically induced decrease in pH within and in the vicinity of the gel phase. As a result, it has become apparent that the faster shrinking kinetics in the coimmobilized enzyme system is attained by the GLcatalyzed hydrolysis of D-glucono-δ-lactone resulting from the oxidation of glucose with GOD. Introduction Many studies have employed enzymes in the preparation of biofunctional materials based on synthetic polymers, a good example of which can be seen in the application of glucose oxidase (GOD) to biosensors,1 drug delivery devices,2 and coimmobilized biocatalysts.3 Then the systems utilized either H2O2 or D-gluconic acid (DGA) as shown in Scheme 1. For DGA, however, its formation takes place via the hydrolysis of D-glucono-δ-lactone (δ-DGL), the process of which is thermal but not enzymatic. In addition, the DGA concentration is in equilibrium with the concentrations of δ-DGL and γ-DGL. As a result, a fall in pH during the oxidation of glucose with GOD becomes fairly slow. This is undesirable for biomaterials in which a GOD-induced pH change is utilized. For example, such a disadvantage inherent in the use of GOD has been noted in our recent study4 on an enzymatically driven polyampholyte gel with immobilized GOD and urease, referred to as a “biochemomechanical system”5 capable of converting biochemical energy created as a result of an enzyme reaction into mechanical work through the swelling and shrinking of the gel. Gluconolactonase (GL; EC 3.1.1.17)6 catalyzes hydrolysis of δ-DGL to result in DGA; this fact has long been known. Also known is that GL exists in microbial cells such as yeast and E. coli and in animal tissues such as porcine liver. The purification of GL has been performed by many researchers,7-12 most of them reporting enzyme characteristics * Corresponding author. Phone: +81-298-53-6633. Fax: +81-298-534605. E-mail: [email protected].

Scheme 1

such as optimal pH and temperature, as well as a role of the enzyme in metabolic systems such as hepatic glycogen degradation. In particular, it is important that the turnover of GL is 13000 min-1 (ref 9), the value of which is much higher than the rate constant (0.01 min-1) of the thermal hydrolysis of δ-DGL.12 Therefore, with the aid of GL, the GOD-catalyzed oxidation of glucose would facilitate the formation of DGA to bring about a rapid fall in pH. This should allow us to improve the performance of GOD-based biomaterials. Indeed, a coimmobilization of GOD with GL (as a partially purified enzyme) has been studied for the purpose of improving a response characteristic of glucosesensitive biosensors.13 In the field of polymer gels, it is a long-standing subject how to attain faster swelling or shrinking kinetics. Improvements in the kinetic property have been made via alteration

10.1021/bm025512f CCC: $22.00 © 2002 American Chemical Society Published on Web 04/26/2002

626

Biomacromolecules, Vol. 3, No. 3, 2002

in the bulk size14 and the macroporosity15 of gels as well as in the structure16 of constituent polymer networks. These means were effective for preparing biochemomechanical gels that undergo a rapid swelling or shrinking change in response to an immobilized urease-induced pH change.17 In the cases of GOD18 and glucose dehydrogenase (GDH),19 however, there lie many difficulties in achieving a faster swelling or shrinking kinetics due to the slow hydrolysis of δ-DGL as mentioned previously. Thus, we attempted to study the effect of GL on the hydrolysis of δ-DGL in the gel system in which GL was coimmobilized with GOD. Particular attention was paid to the rate-determining step in the shrinking process, i.e., enzymatically induced pH change or diffusion of the networks. Through this study we intend to provide a useful tool in the preparation or modification of polyelectrolytegel-based biomaterials that are in need of GOD or GDHinduced pH changes. Experimental Section Purification of GL and Enzyme Assay. Cultures of Aspergillus niger (IAM 2094) were grown in a liquid medium with the following constituents (g‚L-1): glucose, 30; yeast extract 9; malt extract, 9; peptone, 15; and CaCO3, 20 (CaCO3 was sterilized separately). The cultures were incubated at 30 °C and for 3 days on a rotary shaker. The harvested cells were washed well with pure water, frozen, and homogenized in 0.1 M Tris-HCl buffer (pH 7.0; 500 mL). After centrifugation of the homogenate at 10000g and at 4 °C for 30 min, the supernatant was fractionated with ammonium sulfate. The precipitate from 40 to 70% saturation was collected by centrifugation and resuspended in 0.1 M sodium acetate buffer (pH 4.0; 300 mL) with stirring. After centrifugation of the suspension to remove insoluble material, the supernatant solution was made 80% saturated in ammonium sulfate. The resulting precipitate was collected by centrifugation and dissolved in 20 mL of 1mM N-2hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) buffer adjusted to pH 7 with 1 mM NaOH. This solution was subjected to the usual ultrafiltration for desalination and for concentration of the enzyme. Further purification was made using a Superose 12 HR column (1.6 cm diameter by 50 cm) and an anion-exchange resin (Resorce Q) column (0.64 cm diameter by 3 cm). The enzyme solution obtained by the ultrafiltration was loaded on the Superose column and eluted with 50 mM HEPESNaOH buffer (pH 7.0) containing 0.2 M NaCl. The fractions with GL activity were collected, combined, and treated with ammonium sulfate to bring the solution to 80% saturation. The resulting precipitate was separated by centrifugation, dissolved in the same HEPES-NaOH buffer, and again applied to the Superose column. The fractions containing GL were combined, diluted several times with pure water, and loaded on the ion-exchange resin column. The elution was made with a linear gradient from 0 to 200 mM NaCl in HEPES-NaOH buffer (20 mM; pH 7.0). The fractions with GL activity were combined and used as a final enzyme solution. The GL activity was assayed at 30 °C with δ-DGL as the substrate. A freshly prepared substrate solution (0.1 mL)

Ogawa et al.

containing 14 µmol of δ-DGL was added into a mixture of the enzyme solution (0.25 mL) and 0.5 M sodium acetate buffer (pH 5.5, 2.5 mL). After the incubation at 10 min, the concentration of the substrate remaining was determined by a colorimetric method with Hestrin alkaline hydroxylamine as a color-producing reagent. The blank test was carried out in the same way as described above, with water instead of the enzyme solution. One unit was then defined as an amount of the enzyme that needs to hydrolyze 1 µmol of δ-DGL in a 5 mM substrate solution for 1 min at pH 5.5 and at 30 °C. The GL concentration was colorimetrically determined with a micro BCA protein assay reagent kit (Pierce Chemical Co., Rockford, IL); this method is based on the chelation of two molecules of bicinchoninic acid (BAC) with one molecule of a protein-Cu+ complex to form a purple colored complex. GOD and Other Chemicals. GOD (EC 1.1.3.4, from Aspergillus niger; Mw ) 186 000) was purchased from Wako Pure Chemical Industries, Ltd. (Tokyo, Japan). This enzyme has a specific activity of 250 units/mg; one unit corresponding to the amount of GOD by which 1 µmol of δ-DGL is formed for 1 min at 25 °C and at pH 7 (3 mM phosphate buffer). All the monomers used in the preparation of ionic gels were obtained from the commercial sources indicated and used after purification in the usual way: N-isopropylacrylamide (NIPA) from Kojin Chemical Co. (Tokyo, Japan); acrylic acid (AAc) and N,N′-methylenebis(acrylamide) (MBAAm) from Wako Chemical Industries, Ltd. Ammonium persulfate (APS, initiator) and N,N,N′,N′-tetramethylethylenediamine (TEMED, accelerator) were also commercial products and used without purification. Preparation of Gels with and without Immobilized Enzymes. The enzyme immobilization was performed by the physical entrapping method using a pregel solution (1 mL) containing NIPA (78.8 mg), AAc (0.504 mg), MBAAm (1.35 mg), TEMED (4.8 µL), APS (0.4 mg), and a desired amount of enzymes. The gelation was carried out at 0 °C for 1 h using a test tube into which glass capillaries with an inner diameter of 290 µm had previously been inserted. After the gelation was completed, the gels were taken out of the capillaries and thoroughly washed with pure water. All the gel samples were cut into cylinders of approximately 2 mm in length and stored at 4 °C before use. The AAc content of the enzyme-free gel, as measured by pH titration with NaOH, was 0.084 mmol/g of dry polymer (acid form). Measurements. Phosphate buffer (pH 7.0) was prepared by mixing aqueous Na2HPO4 and NaH2PO4 solutions (1 mM each) containing 0.9% NaCl. Other buffers (pH < 7) were obtained from this buffer through pH adjustment with a slight volume of DGA solution (2.5 M). The O2-saturated buffer was used for the preparation of a glucose solution (5 mM) as the substrate for GOD. Time-dependent changes of solution pH during the GODcatalyzed oxidation of glucose in the O2-saturated buffer (pH 7) containing and not containing GL were studied at 33 °C using the usual technique of pH measurements. The pH dependence of enzyme activity was studied by measuring the consumption rate of glucose or δ-DGL (initial concentration, 5 mM each). The measurements of glucose concentration were done with Glucose C-II Test Wako (Wako Pure

Biomacromolecules, Vol. 3, No. 3, 2002 627

Coimmobilization of Gluconolactonase with Glucose Oxidase Table 1. Change in Gluconolactonase Activity at Different Stages of Purification stage

activity (units/mL)

protein (mg/mL)

volume (mL)

specific activity (units/mg)

overall purification (fold)

yield (%)

after centrifugation of cell homogenate after ultrafiltration after column chromatography

21.3 580 169.2

0.79 9.0 0.062

7950 67.5 6.2

26.9 64.4 2729

2.4 101

100 23 0.6

Figure 1. SDS-PAGE for GL after column purification. Lanes 1 and 2 denote molecular mass standards with the molecular weight (Mw) indicated on the left and purified GL, respectively. Silver staining method was employed.

Chemical Industries, Ltd., Tokyo, Japan). The δ-DGL concentration was determined as described above. For the size measurements of gels, the sample in the shape of a small cylinder was placed in a microcell together with the buffer (pH 5 or 7). The diameter was measured using a microscope (Olympus CK2-TRP-1) with a calibrated scale. The temperature was controlled to an accuracy of (0.02 °C between 32 and 35 °C with circulating water around the measuring cell. For the study of enzymatically induced gel collapse, measurements were performed in a closed system (gel:substrate solution ∼ 1:10000 in volume) without stirring. The overall concentration change of glucose during the measurement was less than 3% at maximum. The details of apparatus and measuring procedures have been described in full in our previous papers (e.g., see refs 19 and 20). Results and Discussion Purification of GL. Table 1 shows the change in GL activity at different stages of enzyme purification. In the best preparations a 100-fold purification of the hydrolytic activity was obtained; this value corresponds to a specific activity of 2730 units/mg, which is 10 times as large as the activity of GL purified from Pseudomonas fluorescens.12 The molecular weight was ca. 2.3 × 105 by gel filtration chromatography with a Superose 12 HR column. As can be seen from Figure 1, however, two protein bands with almost the same intensity were observed; the corresponding molecular weights were estimated to be 3.2 × 104 and 4.0 × 104. This contradiction may be resolved by assuming that GL consists of several of two subunits having little difference in molar mass. The purification of GL from Aspergillus niger has been attempted by Hanazato et al. (see ref 13a) using a GODcontaining crude GL from Oriental Yeast Co., Osaka, Japan. Then, they showed the following data: specific activity )

Figure 2. Time-dependent changes of hydrogen ion concentration during the oxidation of glucose with GOD (80 units/mL), GL (3.94 units/mL), and a mixture of GOD (80 units/mL) and GL (3.94 units/ mL). The reaction was studied at 33 °C in the substrate solution (5 mM glucose-containing phosphate buffer which was initially adjusted to pH 7.0).

5550 units/mg and molecular weight ) 2.7 × 104 by gel filtration chromatography with a Sephacryl S-200 column. By SDS-PAGE, however, two bands with the same intensity were observed at the positions corresponding to the molecular weights 2.3 × 104 and 2.7 × 104; thus they concluded that the purification was not complete. Comparison of the results by Hanazato et al. with ours gave a few differences in SDS-PAGE data, but a big difference in the molar mass by gel filtration chromatography. Nevertheless, the latter would be understood by the assumption that GL easily dissociates into two subunits with slightly different molar masses. In addition, one would not take seriously the fact that a specific activity of GL by Hanazato et al. is twice that by us, because they estimated the activity on the formation of DGA (but not on the consumption of δ-DGL); the amount of DGA was then calculated from the overall volume of NaOH required to maintain the system at pH 7.0 by use of a pH-stat instrument. As a result, our GL from Aspergillus niger seems to have a considerably high purity. Indeed, SDS-PAGE showed there is little protein-band corresponding to GOD which invariably appears between the molecular weights of 6.63 × 104 and 9.76 × 104 at our measuring conditions. Moreover, our GL does not result in DGA from glucose (see Figure 2). Changes in pH Due to Glucose Oxidation by GOD in the Absence and Presence of GL. The formation of H+ ions during the GOD-catalyzed oxidation of glucose in the presence of GL is a significant factor in this study. Figure 3 shows the pH dependence of the forming rate (∆[H+]/∆t) of hydrogen ions in molarity per minute. Also shown in Figure 3 is the pH-activity curves of GL and GOD (see ref

628

Biomacromolecules, Vol. 3, No. 3, 2002

Ogawa et al.

Figure 4. Change in the buffering capacity with pH. The phosphate buffer (20 mL; pH 7.0) was titrated with 100 mM DGA solution. The titration of carbonate-free 0.9% NaCl solution (20 mL) as the control sample was made with two DGA solutions: 1 mM (pH 6.98-5.72) and 100 mM (pH 5.23-3.08).

Figure 3. Forming rate (a) of hydrogen ions in the GOD/GL system and the activities (b) of GOD and GL as a function of pH. The formation rate of H+ ions was obtained from data in Figure 2 (∆t ) 1 min). The relative activity was obtained from the consuming rate of glucose with GOD and of δ-DGL with GL; thus 100% activity refers the maximum rate for each of the enzyme reactions. The enzyme reaction bought about a decrease in pH; thus the horizontal axis was given so as to recognize the time-sequential pH change.

23). In this study we were unable to use a buffer-free system which allows us to utilize the pH-stat technique for a direct determination of ∆[H+]/∆t at a given pH; e.g., see ref 13a. The reason is related to the size measurement of gels, in which this technique was not useful for holding the solution pH surrounding the gel at a constant level when the immobilized enzyme reaction takes place within the gel phase. Thus, ∆[H+]/∆t was determined by graphical differentiation of the [H+] vs t curve in Figure 2. The pH dependence of ∆[H+]/∆t displays a gradual increase at pH 7-5.5, a rapid increase at pH 5.5-5, a maximum at pH around 4.7, and a drastic fall at pH near 4. The pH-activity curves show that at pH 7-5.5 both GOD and GL maintain the activities higher than 80%. Also shown from the activity curves is a drastic fall in the GL activity with decreasing pH from 5 to 4, while in this pH range the GOD activity gradually decreases. To understand the pH change of ∆[H+]/∆t in comparison with the activity curves, it should be noted that the enzyme reaction was studied in the presence of the phosphate salts as the buffer agent. Thus, one may anticipate that ∆[H+]/∆t rapidly increases when [H+] goes up to over the limit of the buffering capacity due to accumulation of DGA in the system. In understanding the fall of ∆[H+]/∆t in the vicinity of pH 4, however, we must consider the activity fall of GL. To confirm the above prediction, we performed a pH titration with DGA of the phosphate buffer (pH 7) used as the medium for the substrate solution. Also carried out as a control experiment was the titration of aqueous 0.9% NaCl

Figure 5. Temperature dependence of normalized equilibrium diameters (de/do) at pH values 5 and 7 for NIPA/AAc gel without immobilized enzyme. do (∼290 µm) denotes the gel diameter at preparation.

solution with DGA. A buffering capacity (Cb) was then calculated by Cb )

7 - (pH)1 7 - (pH)2

)

7 + log[H+]1 7 + log[H+]2

(1)

where subscripts 1 and 2 denote the systems not containing and containing the phosphates, respectively. As can be seen from Figure 4, the buffering capacity exponentially decreases when decreasing pH from 7 to 5, and at pH < 4 it approaches unity, meaning no buffering action. These results clearly indicate that the gradual increase in ∆[H+]/∆t at pH 7.05.5 should be related to the decrease in the buffering capacity. In addition, the rapid increase in ∆[H+]/∆t at pH 5.5-5.0 is due to a few buffering capacities. At pH 4 at which the system no longer holds the buffering capacity, ∆[H+]/∆t rapidly decreased because of the fall in the GL activity. As a result, it was found that both GL activity and buffering capacity play a key part in the enzymatically induced pH change in our system. Shrinking Kinetics of NIPA/AAc Gels. The gel used for the enzyme immobilization was characterized prior to the study on the enzymatically induced gel collapse. Figure 5 shows the temperature dependence of equilibrium swelling ratio at pH values 7 and 5 for the enzyme-free gel. The swelling ratio was obtained by normalization of equilibrium

Coimmobilization of Gluconolactonase with Glucose Oxidase

Biomacromolecules, Vol. 3, No. 3, 2002 629

swell or shrink is given by τ ) l 2/π2D

(2)

where l denotes the characteristic linear dimension of the gel and D is the collective diffusion coefficient of the networks. For a gel spherical in shape and when its volume change is small, τ is related to swelling ratio (r) and time (t) by r∼

1 6 exp 2 τ π

( )

(3)

We used the gel cylindrical in shape, so that eq 3 may not be directly applied to our data in the swelling and shrinking processes. As shown in Figure 6b, however, there is a good linear relation in the time course of natural logarithm of the swelling degree (Sd) given by Figure 6. Time courses of normalized gel diameter (a) and swelling degree (b) in the logarithmic expression after a sudden pH change of the outer medium from 7 to 5 and vice versa for NIPA/AAc gel without immobilized enzyme at 33 °C.

gel diameter (de) at a given temperature using the inner diameter (do, 290 µm) of the capillary used in the gel preparation. It is found that at 33 °C at which the enzymatic pH change has been studied, the gel undergoes a shrinking or swelling change in response to the ambient pH change from 7 to 5 or vice versa. We have accounted for the volume phase transition in gels by hypothesizing a balance between the repulsion and attraction among functional groups attached to the crosslinked polymers which arise from a combination of four intermolecular forces:21 ionic, hydrophobic, van der Waals, and hydrogen bonding. When a repulsive force, usually electrostatic in nature, overcomes a nonionic attractive force such as hydrogen bonding and hydrophobic interaction, the gel volume should increase discontinuously in some cases and continuously in others. Thus, the pH-induced change of de/do at a constant temperature (33 °C) may be understood as an alteration in the ionization degree of COOH with pH while hydrophobic interaction mainly due to the NIPA moiety is kept constant. From our previous study22 on pH titration of NIPA/AAc gels, it has been found that the degree of ionization is about 0.45 at pH 5 and is about 0.95 at pH 7. The time courses of shrinking and swelling changes in response to a sudden pH change of the outer medium from 7 to 5 or vice versa were studied. For this, diameter (dt) at time t was normalized by do, and dt/do was plotted against time (see Figure 6a). An abrupt change in pH from 7 to 5 causes a collapse transition which takes place very rapidly for the first 30 min and then slowly for the next 60 min. About 90 min is required to reach a fully collapsed state at which dt/do ∼ 0.7. In contrast to such shrinking changes, a sudden increase in pH from 5 to 7 allows for the gel to swell rapidly until a fully swollen state is reached. With respect to the swelling-shrinking kinetics of gels, Tanaka et al.24,25 have shown that the characteristic time (τ) taken for a gel to

Sd )

dt - dec de - de s

c

∼

( ){( ) } 1 a

dt -b do

(4)

where des and dec represent an equilibrium gel diameter under a given swollen state and a collapse state, respectively. Since a ) (des - dec)/do and b ) dec/do; we may obtain a ∼ 0.3 and b ∼ 0.7 from Figure 6a. As a result, τ was estimated from the slope of ln Sd vs t plots; i.e., τ ∼ 24 min in the shrinking process and τ ∼ 19 min in the swelling process. These values should be an indication for characterizing diffusion of the networks composed of cross-linked NIPA/ AAc copolymer chains. Since the enzymatically induced gel collapse would be governed by either diffusion of the networks or ∆[H+]/∆t, a comparison of τ values for the gels with and without the immobilized enzyme allows us to investigate which factor to govern. Shrinking Kinetics of NIPA/AAc Gels with Coimmobilized GL and GOD. Figure 7 shows the time course of enzymatically induced diameter change of the gel with the coimmobilized GOD and GL. The gel sample (3.24 × 10-5 mL) in the shape of a small cylinder was placed in a “large” volume (0.4 mL) of the buffer solution. After that, the outer medium was quickly replaced with the substrate solution by means of oxygen gas pressure. This replacement leads to the gel collapse, in which the diameter exponentially decreases for the first 90 min, remains almost constant over a period of the next 50 min, and then decreases again with a very slow rate. Since the shrinking stopped at once (or became very slow) between 90 and 140 min and at this period dt/do was 0.7, we plotted ln Sd against time as done in the previous section. The results are shown in Figure 7b, from which we obtained τ ∼ 24 min. This value is equal to that of the enzyme-free gel. Therefore, it is clear that the shrinking observed in the GL/GOD system is governed by diffusion of the gel networks but not by the enzyme reaction. It is important to investigate the role of GL in the shrinking process of the coimmobilized enzyme system by comparison with the NIPA/AAc gel in which GOD was singly immobilized. Our previous study18 has dealt with such a singly immobilized GOD system, but there are several differences between the previous and present studies, e.g., amount of

630

Biomacromolecules, Vol. 3, No. 3, 2002

Figure 7. Time courses of normalized gel diameter (a) and swelling degree (b) in the logarithmic expression observed at 33 °C in the gel system with coimmobilized GOD and GL after quick replacement of the outer medium (0.4 mL) with the same volume of the substrate solution. The amount of immobilized enzyme in units per mL of gel at preparation: 1600 for GOD and 39.4 for GL from which the overall concentrations of the enzyme in a measuring cell were estimated to be 0.13 units/mL for GOD and 0.0032 units/mL for GL. Thus, the difference in GOD or GL concentration between the free and immobilized enzyme reactions should be considered in a comparison of the results in Figures 2 and 7.

Figure 8. Time courses of normalized gel diameter (a) and swelling degree (b) in the logarithmic expression observed at 33 °C in the gel system with singly immobilized GOD (1600 units/mL gel at preparation) after quick replacement of the outer medium with the substrate solution. Volume of the outer medium was ∼0.4 mL.

the immobilized GOD, buffer solution, and measurement conditions.26 Thus, we examined again the GOD system which was prepared under the same conditions used for the GL/GOD system, except that GL was not immobilized. The results obtained in the substrate solution are shown by open circles in Figure 8, together with those (closed circles) in a buffer-free substrate solution (i.e., aqueous 0.9% NaCl solution containing 5 mM glucose). In the substrate solution, the gel shrinks slowly in the first stage, shrinks smoothly but not rapidly in the second stage, and finally reaches a fully collapsed state. In the buffer-free system, no slow

Ogawa et al.

shrinking change at the first stage was observed. The τ values estimated are 250 h (15 000 min) at the first stage and 2.8 h (168 min) at the second stage, both in the buffered system, and 1.8 h (108 min) in the buffer-free system. These results may be interpreted as follows: (i) A large portion of DGA produced is immediately neutralized with bases in the buffer, not causing a smooth and large change in the solution pH both within and in the vicinity of the gel phase; this would be the reason for the gradual shrinking at the first stage (Note that such a gradual shrinking was not observed in the bufferfree system). (ii) When the amount of the acid formed is in excess of the buffering capacity, the pH fall in the vicinity of the gel becomes larger, allowing for the gel to shrink smoothly at the second stage. Therefore, the shrinking rate of the gel with singly immobilized GOD should be governed by the enzyme reaction, but not by diffusion of the networks. Another important result from Figure 8 is that the dt/do value (∼0.65) for the gel fully collapsed in the substrate solution is little different from that (∼0.67) in the bufferfree system. Also of importance is that both dt/do values are very close to that (∼0.63) of the GOD/GL system at times >400 min (see Figure 7). By taking these into account, let us go back to Figure 2. Then we may notice that an increase in [H+] levels off at around 10-4 mol/L (pH 4.0), regardless of the presence or the absence of GL. This means that both in the GL/GOD system and in the GOD system the pH level of the whole medium surrounding the gel under a fully collapsed state would be identical with each other. Indeed, the de/do becomes 0.65 when the gel was equilibrated at 33 °C in the buffer adjusted to pH 4 with DGA. The above discussion allows us to account for why the dt/do vs t curve of the GL/GOD system passed through a plateau between 90 and 140 min before reaching a fully collapse state. As was mentioned previously, the dt/do at the plateau is close to the de/do at pH 5 and at 33 °C. This means that the solution pH within and in the vicinity of the gel quickly decreases from 7 to 5 as if the outer medium with pH 7 is physically replaced with the medium with pH 5. Thus, a slight and slow gel collapse after reaching the plateau may be due to a fall in pH from 5 to 4. At this pH range, the medium maintains a few buffering capacities, and the GL exhibits a drastic fall in the activity (see Figures 3 and 4). As a result, the hydrolysis of δ-DGL which is not an enzymatic but a thermal change causes a slow gel collapse, even when GL was coimmobilized. Conclusions We have prepared a gel-entrapped enzyme system in which GL was coimmobilized with GOD, using an ionic gel consisting of lightly cross-linked copolymer chains of NIPA and AAc. GL was isolated from Aspergillus niger and purified about 100-fold. Shrinking kinetics of the gel was studied at 33 °C in the substrate solution adjusted to pH 7.0 with phosphate salts. It was found that the coimmobilization of GL with GOD allows the gel to shrink rapidly as the enzyme reaction sets in. The shrinking rate of the GL/GOD system was identical to that of the enzyme-free gel in response to a sudden pH change of the outer medium from

Coimmobilization of Gluconolactonase with Glucose Oxidase

7 to 5. In the gel with the singly immobilized GOD, however, we observed very slow shrinking changes both in buffered and in buffer-free substrate solutions. These results were discussed in terms of the concentration change of H+ ions during the GOD-catalyzed oxidation of glucose in the presence of GL. Then the pH dependence of the enzyme activity as well as of the buffering capacity was considered. As a result, it was demonstrated that the difference in the shrinking rate between the gels with and without coimmobilized GL is related to whether the hydrolysis of δ-DGL takes place enzymatically or not. Acknowledgment. This work was supported in part by a Grant-in-Aid for Scientific Research to E.K. from the Ministry of Education, Japan (# 08558092). References and Notes (1) Andrei, B. K.; Maya, Z.; Amir, L.; Eugenii, K.; Itamar, W. Sens. Actuators, B 2000, 70, 222 and references therein. (2) (a) Brown, L. R.; Edelman, E. R.; Fischel-Ghodsian. F.; Langer, R. J. Pharm. Sci. 1996, 85, 1341. (b) Kim, C. K.; Im, E. B.; Lim, S. J.; Oh, Y. K.; Han, S. K. Int. J. Pharm. 1994, 101, 191. (c) Fiitz, E.; Brandenburg, D. Proc. Natl. Acad. Sci. U.S.A.. 1988, 85, 2403 and references therein. (3) (a) Van de Velde, F.; Lourenco, N. D.; Bakker, M.; Van Rantwijk, F.; Sheldon, R. A. Biotechnol. Bioeng. 2000, 69, 286. (b) Bartlett, P. N.; Ali, Z.; Eastwickfield, V. J. Chem. Soc., Faraday Trans. 1992, 88, 2677. (c) Gestrelius, S.; Msttiasson, B.; Mosbach, K. Eur. J. Biochem. 1973, 36, 89 and references therein. (4) Ogawa, Y.; Ogawa, K.; Wang, B.; Kokufuta, E. Langmuir 2001, 17, 2670. (5) (a) Kokufuta, E. Functional Immobilized Biocatalysts Prepared Using Stimulus-sensitive Polymer Gels. In The Polymeric Materials EncyclopediasSynthesis, Properties and Applications; Salamone, J. C., Ed.; CRC Press: Boca Raton, FL, 1996; Vol. 4, F-G, pp 26152621. (b) Kokufuta, E. AdV. Polym. Sci. 1993, 110, 159. (c) Kokufuta, E. Prog. Polym. Sci. 1992, 16, 647. (d) Kokufuta, E.; Zhang, Y.-Q.; Tanaka, T. Nature 1991, 351, 302. (6) Schomburg, D., Salzmann, M., Eds. Enzyme Handbook 3; SpringerVerlag: Berlin, 1991

Biomacromolecules, Vol. 3, No. 3, 2002 631 (7) Bailey, G. D.; Roberts, B. D.; Buess, C. M.; Carper, W. R. Arch. Biochem. Biophys. 1979, 192 (2), 482. (8) Roberts, B. D.; Bailey, G. D.; Buess, C. M.; Carper, W. R. Biochim. Biophys. Res. Commun. 1978, 84 (2), 322. (9) Grossman, S. H.; Axelrod, B. J. Biol. Chem. 1973, 248, 4846. (10) Hucho, F., Wallenfels, K. Biochim. Biophys. Acta 1972, 276, 176. (11) Brodie, A. F.; Lipmann, F. J. Biol. Chem. 1955, 212, 677. (12) Jermyn, M. A. Biochim. Biophys. Acta 1960, 37, 78. (13) (a) Hanazato, Y.; Inatomi, K.; Nakako, M.; Shiono, S.; Maeda, M. Anal. Chim. Acta 1988, 212, 49. (b) Hanazato, Y.; Shiono, S.; Maeda, M. Anal. Chim. Acta 1990, 231, 213. (c) Miwa, Y.; Nishizawa, M.; Matsue, T.; Uchida, I. Bull. Chem. Soc. Jpn. 1994, 67, 2864. (14) Sato-Matsuo, E.; Tanaka, T. J. Chem. Phys. 1988, 89, 1695. (15) (a) Kabra, B. G.; Gehrke, S. H. Polym. Commun. 1991, 32, 322. (b) Wu, X. S.; Hoffman, A. S. J. Polym. Sci., Part A 1992, 30, 2121. (16) (a) Kaneko, Y.; Sakai, Y.; Kikuchi, A.; Yoshida, R.; Sakurai, Y.; Okano, T. Macromolecules 1995, 28, 7717. (b) Yoshida, R.; Uchida, K.; Kaneko, Y.; Sakai, K.; Kikuchi, A.; Sakurai, Y.; Okano, T. Nature 1995, 374, 240. (17) (a) Kokufuta, E.; Matsukawa, S. Ber. Bunsen-Ges. Phys. Chem. 1996, 100, 1073. (b) Ogawa, K.; Wang, B.; Kokufuta, E. Langmuir 2001, 17, 4704. (18) Matsuda, K.; Orii, H.; Hirata, M.; Kokufuta, E. Polym. Gels Networks 1994, 2, 299. (19) Kokufuta, E.; Aman, Y. Polym. Gels Networks 1997, 5, 439. (20) Kokufuta, E.; Suzuki, H.; Yoshida, R.; Yamada, K.; Hirata, M.; Kaneko, F. Langmuir 1998, 14, 788. (21) Ilmain, F.; Tanaka, T.; Kokufuta, E. Nature 1991, 349, 400. (22) Suzuki, H.; Wang, B.;. Yoshida, R.; Kokufuta, E. Langmuir 1999, 15, 4283. (23) The pH-activity curves are close to those previously reported: for GOD (A. niger source), Bright, H. J.; Appleby, M. J. Biol. Chem. 1971, 246, 2734; and for GL (P. fluorescens source) see ref 12. (24) Tanaka, T.; Fillmore, D. J. J. Chem. Phys. 1979, 70, 1214. (25) Tanaka, T.; Sato, E.; Hirokawa, Y.; Hirotsu, S.; Peetermans, J. Phys. ReV. Lett. 1985, 55, 2455. (26) When comparing the present and the previous study, we should call attention to the difference in the overall GOD concentration in a measuring cell; the concentration in ref 18 was 10 times that of the present study because of the difference in the volume ratio of external solution to a gel sample.

BM025512F