Determination of the enzyme reaction rate in a differential fixed-bed reactor: a case study

DETERMINATION OF THE ENZYME REACTION RATE IN A DIFFERENTIAL FIXED-BED REACTOR: A CASE STUDY

E.A.Baruque Filho, M.G.A.Baruque and G.L.Sant’Anna Jr.^*

Universidade Federal do Rio de Janeiro, COPPE, P.O.Box 68502, 21945-970,

Rio de Janeiro - RJ, Brazil

(Received: January 16, 1998 ; Accepted: November 3, 1999)

Abstract - The reaction rate of starch hydrolysis catalyzed by a glucoamylase covalently bound to chitin particles was measured in a Differential Fixed-Bed Reactor (DFBR). Under selected test conditions the initial reaction rate may represent biocatalyst activity. Some aspects which influence measurement of the initial reaction rate of an immobilized enzyme were studied: the amount of desorbed enzyme and its hydrolytic activity, the extent of pore blockage of the biocatalyst caused by substrate solution impurities and the internal and external diffusional mass transfer effects. The results showed that the enzyme glucoamylase was firmly bound to the support, as indicated by the very low amount of desorbed protein found in the recirculating liquid. Although this protein was very active, its contribution to the overall reaction rate was negligible. It was observed that the biocatalyst pores were susceptible to being blocked by the impurities of the starch solution. This latter effect was accumulative, increasing with the number of sequential experiments carried out. When the substrate solution was filtered before use, very reliable determinations of immobilized enzyme reaction rates could be performed in the DFBR. External and internal diffusional resistences usually play a significant role in fixed-bed reactors. However, for the experimental system studied, internal mass transfer effects were not significant, and it was possible to select an operational condition (recirculation flow rate value) that minimized the external diffusional limitations.

Keywords: immobilized enzyme activity, bioreactor, enzyme desorption, enzyme activity, enzyme reactor.

INTRODUCTION

The experimental set-up often used to perform kinetic studies with immobilized enzymes is the Differential Fixed-Bed Reactor (DFBR) (Ford et al., 1972), which is usually connected to a stirred tank and operated in batch mode as described by Vallat and Monsant (1986). This system is particularly well suited to the evaluation of biocatalysts which will be used in industrial scale fixed-bed reactors (Pitcher, 1975).

Determination of immobilized enzyme-inherent parameters is often a challenging task in view of the complex aspects that influence heterogeneous enzymatic catalysis, such as enzyme desorption and consequent mixed kinetics (both immobilized and free enzymes are active), internal and external diffusional effects and enzyme deactivation. Determination of the reaction rate of an immobilized enzyme under standard conditions is a necessary previous step for the study of enzyme kinetics and mass transfer in catalytic beds.

Although the DFBR set-up has been extensively studied and often employed to determine biocatalyst activity, few works analyze in depth the phenomenon of enzyme desorption and its influence on the experimental determination of the initial reaction rates used to express immobilized enzyme activity. In order to perform standard determinations of biocatalyst activity and to establish an experimental procedure leading to the evaluation of kinetic parameters, the magnitude of enzyme desorption and mixed kinetics should be precisely measured.

This article deals with the determination of the reaction rate of an immobilized enzyme. To carry out this work the reaction of starch hydrolysis catalyzed by a covalently bound glucoamylase was selected in view of its industrial importance (Roig et al., 1995; Zanin and De Moraes, 1996) and our previous experience with this biocatalyst (Freire and Sant’Anna Jr., 1990a).

MATERIALS AND METHODS

Enzyme

Aspergillus niger glucoamylase (E. C. 3. 2. 1. 3) from Novo Nordisk, Denmark (300 AGU/ml) was immobilized (without previous purification) onto chitin particles activated with hexamethylenediamine and glutaraldehyde as described by Bon et al. (1984).

Support

The chitin particles used for enzyme immobilization were obtained from crab shells, which were submitted to the following treatment processes: cleaning and washing, acid treatment for calcium carbonate removal, washing, drying, crushing and size separation. The average size of the selected particles was in the range of 0.177 to 0.420 mm.

Substrate

Soluble starch AG from Reagen was used to prepare substrate solutions of 0.2%, 1%, 5% and 10% (w/v). The starch solutions were prepared with distilled water and the pH value was adjusted to 3.8. Unless otherwise indicated in this text, these solutions were not filtered.

Immobilized Enzyme Activity Determination

The activity of biocatalyst samples (20mg) was determined in a small basket reactor in the absence of diffusional effects by measuring the initial rate of hydrolysis of a soluble starch solution (5% w/v), as proposed by Freire and Sant’Anna Jr. (1990a).

One unit of activity (U) was defined as the amount of enzyme which catalyzes the liberation of reducing sugars equivalent to 1mmol of glucose per minute from the soluble starch solution at pH 3.8 and 45°C.

DFBR Operation

The DFBR used in this work was based on the systems proposed by Ford et al. (1972)and Vallat and Monsant (1986). Figure 1 shows the experimental set-up that was composed of the reactor, a stirred tank and a peristaltic pump to promote the complete recirculation of the substrate solution.

The glass reactor had an external jacket connected to a water bath to keep the reaction temperature constant at 45°C. The biocatalyst was confined inside the reactor in the thin space bounded by two porous plates. A bed volume of 7.3ml (diameter = 2.9cm, height = 1.1cm.) contained about 0.8 g of biocatalyst. The total liquid volume in the experimental set-up (DFDR, tubes and stirred tank) was about 600 ml.

The substrate solution was recirculated at a constant flow rate in the range of 3.5 to 350 ml/min. Liquid samples were withdrawn from the stirred tank and immediately analyzed. The stirred tank temperature was also maintained constant at 45°C.

The experimental set-up was designed to perform successive experimental runs with the same biocatalyst sample. When the run was finished, the liquid in the reactor was drained and a buffer solution was recirculated through the catalytic bed to remove glucose containing liquid fractions entrapped in the bed voids.

Determination of Reaction Velocities

Two reaction velocities were considered in this study:

[ v_i ] corresponds to the reaction rate, which is practically constant in short-term runs when the reaction conversion is maintained below 5%.

[ v_s ] is the instantaneous reaction rate catalyzed by the enzyme desorbed from the support, which is free in the recirculating liquid, as discussed below.

In the short-term experimental runs the conversion was below 5%, and a linear relationship between product concentration (P) and reaction time (t) was then observed. Thus [ v_i ] is slope of the straight line relating (P) and (t).

Some long-term experimental runs were conducted to study the variation in enzyme desorption with time. In these experiments the reaction rate [ v_i ] was calculated from the slope of the reaction progress curve at any given time.

[ v_s ] was determined for short and long-term experimental runs by the following procedure: a 3 ml sample was withdrawn from the stirred tank at a particular time of reaction (t). A fraction of this sample was immediately placed in a boiling water bath to deactivate the free enzyme in solution, enabling the determination of its original glucose content. The remaining sample volume was incubated at 45°C and the variation in its glucose content was monitored. About five glucose determinations were made in a 24 h period of incubation to obtain a progress curve, which was a straight line for all the experiments performed. [ v_s ] corresponds to the slope of this straight line. Several determinations of [ v_s ] were made during an experimental run in order to report its variation in the course of each experiment.

Evaluation of External and Internal Mass Transfer Limitations

In order to study the magnitude of the internal diffusional effects, experiments were carried out in the DFBR by varying particle size within the range of 0.177 to 0.420 mm. The substrate concentration was 5% (w/v) and the recirculation flow rate was fixed at 250 ml/min. The reaction progress curve obtained for each short-term experiment led to the determination of the initial reaction rate.

External mass transfer limitations were evaluated in the DFBR by varying the recirculation flow rate within the range of 3.5 to 350 ml/min. Three different substrate concentrations were tested: 0.2%, 1% and 5% (w/v). From the experimental data it was possible to calculate the external effectiveness factor (h), using the following equation:

where v_obs is the observed reaction rate and v_cin is the experimental reaction rate measured in the absence of external diffusional effects.

To express some results, superficial velocity (U_L) was used instead of the recirculation flow rate. That parameter is defined as follows:

where q_r is the flow rate crossing the catalytic bed and A is the cross-sectional flow area of the reactor.

Analytical Methods

Glucose was determined by the glucose-oxydase/peroxydase test using enzymes supplied by Sigma Chemical Company.

Protein was determined by the classic method proposed by Bradford (1976), using BSA as the standard.

RESULTS AND DISCUSSION

Determination of Reaction Rates

Our original intent was to conduct a series of short-term experimental runs (low conversion, constant [ v_i ]) in the DFBR. Thus, a reliable mean initial reaction rate should be calculated from several experimental determinations. However, as shown in Figure 2, an unexpected result was observed: the [ v_i ] values quickly decreased when determined in successive experimental runs. This result did not allow the calculation of an average [ v_i ], corresponding to the mean activity of the biocatalyst. Thus the question arose of how to explain this unexpected catalytic behavior if all the experimental runs were carried out under the same test conditions. The only operation carried out between two successive runs was reactor washing with 500 ml of sodium acetate buffer solution (pH=3.8) to eliminate residual glucose which would interfere with the results of the next experiment. The observed decrease in [ v_i ] may not be attributed to a conventional deactivation of the immobilized enzyme because strong evidence in the literature indicates that the covalent binding of glucoamylase is sufficiently stable (Freire and Sant’Anna Jr., 1990b; Lobavarewski et al., 1984; Cabral, 1982; Roig et al., 1995). Furthermore, in the course of each experimental run no detectable enzyme deactivation was observed, as confirmed by the strict linear relationship between glucose produced and reaction time.

Some hypotheses were formulated to explain the decrease in [ v_i ] shown in Figure 2:

(a) intense flow through the catalytic bed causes enzyme desorption from the support. This desorbed enzyme may be lost during reactor washing between two successive experimental runs.

(b) mechanical stresses caused by variation in flow and air flow through the bed (during reactor start-up and shut-off periods) promote enzyme deactivation.

(c) pore blockage of the biocatalyst is occurring. This accumulative effect, which may increase with the number of experimental runs, may be caused by some component of the substrate solution used to feed the reactor.

To check these hypotheses, a set of tests was carried out and their results are described as follows.

Desorbed Protein Evaluation

To measure the protein content in the recirculating liquid, the Bradford method was used instead of the classic Lowry method, as suggested by Baruque Filho (1991), who observed that the bifunctional agent used for enzyme immobilization (glutaraldehyde) interferes in the analysis of protein by the Lowry method. Furthermore, the Bradford method had been shown to be precise in detecting protein concentrations in the range of 3 to 6 mg/ml, like those found in this work.

It was observed that, in successive experimental runs, protein concentration fell within the above range. This was observed even when more than 12 successive runs were performed (the protein content never surpassed 6 mg/ml). This protein concentration was very close to that corresponding to the natural protein found in the starch used to prepare the substrate solution. Thus, it can be stated that a very small amount of protein was desorbed from the catalytic bed. The variation in recirculating protein content between two successive runs was so small that the Bradford method was not precise enough to detect it. However, determination of the hydrolytic activity of this protein seems to be a sensible way to quantify such a variation, indicating how much the desorbed enzyme contributes to the overall reaction rate. [ v_s ] is the parameter which represents the activity of the desorbed protein, which was always below 4% of [ v_i ] in all experiments, showing that for short-term runs the contribution of the desorbed enzyme to the overall kinetics is negligible. Thus, reaction rate was determined by the immobilized enzyme itself.

Figure 3 illustrates the profile of desorbed enzyme activity [ v_s ] versus time of reaction for two successive experimental runs. These two profiles confirm that enzyme desorption occurs and there was a decrease in [ v_s ] from the first to the second run, although the recirculating protein content was practically the same in the two experiments.

The results of a long-term experimental run, when the DFBR was operated during 480 minutes and the conversion of starch to glucose almost attained 80%, also indicate that the recirculating protein concentration was in the range of 3 to 6 mg/ml and the [ v_s/v_i ] ratios were also very small (lower than 10%). In contrast with the results shown in Figure 3, the activity of the desorbed protein (expressed as [ v_s ]) showed a decreasing profile in relation to time of reaction as illustrated in Figure 4. Considering the [ v_s ] profiles shown in Figures 3 and 4, it may be concluded that activity of the recirculating enzyme (desorbed) increased rapidly only during the first minutes of operation and then fell showing a profile which presents a decreasingly oscillatory shape. The decay in activity of the desorbed enzyme observed in long-term runs reflects an intrinsic deactivation caused by the mechanical stresses associated with the high recirculation flow. The free desorbed enzyme seems to be more sensitive to operational stresses than the immobilized enzyme in a DFBR.

These experiments have shown that a very small amount of protein was desorbed from the catalytic bed of the DFBR. This desorbed protein had a hydrolytic activity, but the contribution of the active recycling protein to the overall rate of reaction was negligible. Thus, the decrease in [ v_i ] observed in successive experimental runs (Figure 2) may not be attributed to the amount of enzyme desorbed in the course of each experiment.

The Influence of Operational Procedure on Determination of the Reaction Rate

To verify if the immobilized enzyme was being deactivated by the mechanical stresses associated with the start-up and operation shut-off procedures, two sets of experiments were conducted. First, the reactor was operated smoothly, to avoid an interruption in flow between sequential experimental runs. In a second set of experiments, the reactor was submitted to some drastic operational conditions to evaluate their effect on immobilized enzyme activity.

To avoid an interruption in flow during the experiments, the DFBR was carefully operated in a such way that at the end of each experimental run the flow rate of the substrate solution was reduced to a minimum value and the tube line was immediately connected to the buffer solution flask. After reactor washing, a similar reverse operation was carried out to connect the tube line with the stirred tank containing the substrate. This careful operation maintained a flow of liquid through the catalytic bed, to avoid its contact with air during the draining period before washing, avoiding air hold-up in the tubeline, air flow through the bed and air entrapment in the voids during start-up. Permanent contact between the biocatalyst and the substrate is considered by many authors to be an important factor in avoiding its deactivation (Ulbrich et al., 1986; Sungur and Akbulut, 1994; Roig et al., 1995). Bubble shocks with the biocatalyst particles may cause enzyme deactivation, as suggested by Freire and Sant’Anna Jr. (1990b). The use of deaeration devices to eliminate air bubbles from substrate solutions was recommended by industrial reactor manufacturer (Daniels, 1985).

Figure 5 illustrates the behavior of the biocatalyst, expressed in terms of initial rate [ v_i ], in a series of sequential experimental runs. In comparison with the results shown in Figure 2, it should be pointed out that the activity of the biocatalyst was practically the same in seven successive runs. If we had not tried to do further experiments, we could have concluded that the deactivation profile observed in Figure 2 was associated with the interruption in flow between sequential experimental runs. However, we insisted on performing more experiments, and Figure 5 shows that after the seventh run a pronounced decrease in [ v_i ] was observed. In some runs we tried to increase the rate of flow through the catalytic bed (run numbers 12 and 15) and this resulted in an increase in [ v_i ], but this change in flow could not restore the biocatalyst activity to the level observed in the former experimental runs. The curve shown in Figure 5 suggests that an irreversible deactivation phenomenon was occurring in the DFBR.

Figure 5

Biocatalyst Pore Blockage

Diffusional Effects

To study the external diffusional effects, the recirculation flow rate in the DFBR and the substrate concentration were increased. So, the thickness of the stagnant film around the biocatalyst particle should be reduced to a minimum value no longer affecting the reaction rate.

The experimental results shown in Figure 7, obtained with a catalytic bed composed of 0.250 mm particles, indicate that the external diffusional effects are significant for the three substrate concentrations tested. However, the reaction rate [ v_i ] reaches a limit value when the recirculation rate is increased. In all cases, above a flow rate of 250 ml/min (U_L higher than 38 cm/min) the external diffusional resistence is reduced to a minimum value.

Substrate concentration directly affects the substrate gradient in the stagnant film around the particle. Thus, the increase in substrate concentration leads to a higher mass tranfer rate and consequently a higher [ vi ].

Figure 8 illustrates both effects (flow rate and substrate concentration) on the reaction rate, expressed as effectiveness factor, as previously defined. It may be observed that for a low substrate concentration (0.2%), h increases slowly with the recirculation flow rate or superficial velocity (U_L), attaining an upper limit value when high U_L values (above 38 cm/min) are used. These results were expected, because it is well known that external diffusional effects are reduced for higher substrate concentrations because the substrate gradient is increased. Working with high substrate concentations (5 to 30%), Vallat and Monsant (1986) and Weetal and Hovewala (1972) observed that the external diffusional effects were negligible in their systems under variable flow conditions when substrate concentration was above 10%.

The magnitude of the internal difffusional effects was evaluated in the DFBR operated at a flow rate of 250 ml/min and a substrate concentration of 5%. Under these conditions the external diffusional resistences were minimized, as already shown in Figure 8. A known technique for determining internal mass tranfer limitations by varying particle size was used in this work.

The results shown in Table 1 indicate that the reaction rate [ v_i ] was not affected by particle size in the range studied. The immobilization procedure and the characteristics of the support material, which has a low internal porosity, suggest that the reaction effectively occurs on the external particle surface without internal gradients of concentration. Thus, internal diffusional limitations were not significant in the DFBR operated under the above conditions.

Furthermore, it should be mentioned that the particle diameter range studied is adequate for immobilized bed reactors. A reduction in particle diameter will lead to an appreciable increase in the pressure drop along the reactor bed, affecting process viability.

CONCLUSIONS

Although this experimental work was carried out with a specific biocatalyst (glucoamylase immobilized on chitin particles), we believe that other covalently bounded enzymes may present similar results in a DFBR. The results related to protein desorption and enzyme stability under mechanical stresses associated with the reactor start-up and cleaning periods obtained in a DFBR may supply useful information for the design and operation of industrial enzymatic fixed-bed reactors.

For the biocatalyst (glucoamylase covalently bound to chitin particles) used in the DFBR, the desorption of enzyme was extremely low. The activity of the desorbed enzyme detected in the recirculating liquid increased in the first 2 minutes of reactor operation, remained stable during 10 or 15 minutes, and then fell showing a decreasing oscillatory profile. The protein concentration in the recirculating liquid was in the range of 3 to 6 mg/ml, which was very close to the original protein content of the starch solution (1% w/v) fed to the reactor.

The contribution of the desorbed enzyme to the overall reaction rate was negligible. Even in short-term experiments, when the activity of that enzyme was higher, its hydrolytic rate was lower than 4% of the overall reaction rate. Thus, for the enzyme reactor system studied, the reaction was exclusively catalyzed by the immobilized enzyme; in other words, mixed kinetics was not observed.

Impurities found at low levels in the starch used to prepare the substrate solutions were able to block the catalyst pores, preventing reproducible determinations of reaction rate in the DFBR. A simple filtration of the substrate solution was effective, removing these impurities, and very reproducible measurements of reaction rate could be performed in more than ten sequential experimental runs.

The external effectiveness factor was close to one when the reactor was operated with a superficial velocity higher than 38 cm/min. Variation in particle size did not affect the reaction rate, suggesting that the reaction occurs predominantly at the particle surface in the system studied.

• Baruque Filho, E. A., M. Sc. Thesis, COPPE/ Universidade Federal do Rio de Janeiro (1991).
• Bon, E., Freire, D.G., Mendes, M.F., Moreira, C.P. and Soares, V.F., Biotechnol. Bioeng. Symp., 14, 485 (1984).
• Bradford, M. M., Analytical Biochemistry, 22, 248 (1976).
• Cabral. J.M.S., Ph.D. diss., Universidade Técnica de Lisboa (1982).
• Daniels, M.J., Industrial operation of immobilized enzymes, in Hidrolise Enzimática de Biomassas, F.F. Moraes and G..M. Zanin (Eds.), Fundaçăo Universidade Estadual de Maringá, Brazil, 167-177 (1985).
• Ford, J.R., Lambert, A.M., Cohen, W. and Chambers, R.P., Biotechnol. Bioeng. Symp., 3, 267 (1972).
• Freire, D.G. and SantAnna Jr., G.L., Biomass, 23, 71 (1990a).
• Freire, D.G. and SantAnna Jr., G.L., Appl. Biochem. Biotechnol., 26(1), 23 (1990b).
• Illanes, A., Zuniga, M.E., Chamy, R. and Marchese, M.P., In Bioreactor Immobilized Enzymes and Cells Fundamentals and Applications, Moo-Young, M. (Ed.) (1988).
• Lobavarewski, J., Paszczynski, A., Wolski, T. and Fiedureck, J., Biotechnol. Biophys. Res. Commun., 121, 220 (1984).
• Pitcher, W., Design and analysis of immobilized enzyme reactor, In Immobilized enzymes for industrial reactors. R. Messing (Ed.), Academic Press, 151-199 (1975).
• Roig, M.G., Slade, A., Kennedy, J.F., Taylor, D.W. and Garaita, M.G., Appl. Biochem. Biotechnol., 50, 11 (1995).
• Scott, T.C., Ph.D. diss., University of Wisconsin (1985).
• Sungur, S. and Akbulut, U., J. Chem. Tech. Biotechnol., 59, 303 (1994).
• Ulbrich, R., Schallenberger, A. and Dareman, W., Biotechnol. Bioeng., 28, 511-22 (1986).
• Vallat, I. and Monsan, P., Biotechnol. Bioeng., 28, 151 (1986).
• Weetall, H.H. and Havewala, N.B., Biotechnol. Bioeng. Symp., 3, 241-266 (1972).
• Zanin, G.M. and De Moraes, F.F., Appl. Biochem. Biotechnol., 57/58, 617 (1996).

*

To whom correspondence should be addressed

Publication Dates

• Publication in this collection
  25 May 2001
• Date of issue
  Mar 2001

History

• Received
  16 Jan 1998
• Accepted
  03 Nov 1999