You are on page 1of 12

626

Biotechnol. Prog. 2007, 23, 626637

Modeling Intrinsic Kinetics of Enzymatic Cellulose Hydrolysis


Suma Peri, Srinivas Karra, Y. Y. Lee, and M. Nazmul Karim*
Department of Chemical Engineering, Texas Tech University, Lubbock, Texas, and Department of Chemical Engineering, Auburn University, Auburn, Alabama

A multistep approach was taken to investigate the intrinsic kinetics of the cellulase enzyme complex as observed with hydrolysis of noncrystalline cellulose (NCC). In the first stage, published initial rate mechanistic models were built and critically evaluated for their performance in predicting time-course kinetics, using the data obtained from enzymatic hydrolysis experiments performed on two substrates: NCC and R-cellulose. In the second stage, assessment of the effect of reaction intermediates and products on intrinsic kinetics of enzymatic hydrolysis was performed using NCC hydrolysis experiments, isolating external factors such as mass transfer effects, physical properties of substrate, etc. In the final stage, a comprehensive intrinsic kinetics mechanism was proposed. From batch experiments using NCC, the time-course data on cellulose, cello-oligosaccharides (COS), cellobiose, and glucose were taken and used to estimate the parameters in the kinetic model. The model predictions of NCC, COS, cellobiose, and glucose profiles show a good agreement with experimental data generated from hydrolysis of different initial compositions of substrate (NCC supplemented with COS, cellobiose, and glucose). Finally, sensitivity analysis was performed on each model parameter; this analysis provides some insights into the yield of glucose in the enzymatic hydrolysis. The proposed intrinsic kinetic model parametrized for dilute cellulose systems forms a basis for modeling the complex enzymatic kinetics of cellulose hydrolysis in the presence of limiting factors offered by substrate and enzyme characteristics.

1. Introduction
Kinetics and further modeling studies of hydrolysis are useful in different stages of processing of biomass to fermentable sugars. They span the entire domain of operations, namely, enzyme characterization and modification, substrate preparation, reactor design, and optimization of feeding profiles of substrate in a fed-batch operation. There are two types of modeling approaches, empirical and mechanistic modeling. Empirical models relate the factors using mathematical correlations, without any insight into the underlying mechanisms. These are easy to develop and are useful in enzyme characterization and substrate preparation. Mechanistic models are developed from the reaction mechanisms, mass transfer considerations and other physical parameters that affect the extent of hydrolysis. As these models address the underlying dynamics of the process, they can be extensively used in every stage. Mechanistic models vary in their complexity based on the intended use of the models. These models are quite useful in describing the reaction mechanism between the ligninocellulosic biomass and enzyme. We need to consider many factors that determine the rate and extent of hydrolysis of biomass for developing mechanistic models. There will be many parameters that bear direct or indirect effects on the degradation of cellulose to fermentable sugars in the presence of enzyme, as reported in the literature. Broadly they can be classified as follows: Enzyme characteristics: adsorption of enzyme onto ligninocellulosic biomass prior to reaction; intermediate and endproduct inhibition that is either competitive or noncompetitive;
* To whom correspondence should be addressed. Tel: +1 806 7423553. Fax: +1 806 742-3552. Email: naz.karim@ttu.edu. Auburn University.
10.1021/bp060322s CCC: $37.00

synergy and thermodynamic considerations of the various enzyme compounds; mass transfer limitations affecting the transport of the enzyme to substrate Substrate characteristics: lignin distribution; presence of other components such as hemicellulose, proteins and fats; particle size; crystallinity; degree of polymerization A comprehensive model has to incorporate all of these factors, without constraints on the amount of experimental data and the computation requirements. As discussed earlier the complexity of the mechanistic models varies based on the underlying assumptions made in the model development. To quantify the enzymatic hydrolysis using simplistic models, it can be divided into two stages: initial stage, where the rate of hydrolysis is almost linear, and final stage, where the rate continuously decreases and saturates. The factors affecting the reaction rates in the two stages are distinct in each case: Initial stage: product inhibition is not important; least affected by mass transfer resistances in the case of dilute systems; chemical pretreatment plays an important role in initial rates; pseudo-steady state can be assumed Final stage: rate is higher initially but reduces later due to product inhibition; pseudo-steady state assumptions do not apply as there will be accumulation of intermediates; substrate characteristics change (crystallinity, degree of polymerization, etc.) (Klyosov, 1990; Valjamae et al., 1998; Zhang et al., 1999) In summary, enzymatic hydrolysis of lignocellulosic biomass depends on many factors: physical properties of the substrate (composition, crystallinity, degree of polymerization, etc.), enzyme synergy (origin, composition, etc.), mass transfer (substrate adsorption, bulk and pore diffusion, etc.), and intrinsic kinetics (Zhang and Lynd, 2004; Chang and Holtzapple, 2000). Effect of physical properties and structure of substrate on

2007 American Chemical Society and American Institute of Chemical Engineers Published on Web 04/28/2007

Biotechnol. Prog., 2007, Vol. 23, No. 3

627

enzymatic hydrolysis of biomass has been presented by Perez et al. (2005). Various commercially available enzymes and their synergetic effects on hydrolysis were detailed by Howard et al. (2003). In the past 50 years, there has been a constant influx of research publications addressing the enzymatic kinetics of cellulose degradation (Ghose and Das, 1971; Howell and Stuck, 1975; Haung, 1975; Humphrey et al., 1979; Dwivedi and Ghose 1979; Fan and Lee, 1983; Wald et al., 1984; Holzapple et al., 1984; Gan et al., 2002; Movagharnejad et al., 2002). Despite these many years of research, kinetics of cellulose degradation is still poorly understood because of competing effects such as physical properties of the substrate, enzyme synergy, and mass transfer to the intrinsic kinetics. These effects cannot be distinguished from each other and introduce large bias and variability in estimating kinetic parameters. Accurate estimation of intrinsic kinetics requires a pure form of cellulosic biomass to minimize the mass transfer resistances and effects of physical properties of the substrate and enzyme complex. In the current work, an emphasis is given to the intrinsic kinetics and their controlling factors. In enzyme-catalyzed reaction networks, inhibitory effects of the reaction intermediates and products play an important role. These inhibitors bind to enzyme active sites and reduce their activity. The substrate may act as the inhibitor in some cases. Hence, the intrinsic reaction kinetics of enzymatic cellulose hydrolysis is also subjected to mediation by a host of factors such as inhibitory effects from reaction intermediates and products, enzyme adsorption, etc. Furthermore, the influence of each factor is difficult to quantify in isolation, as many factors are interrelated during the hydrolytic reaction. In this work, a three-step approach was adapted to investigate the intrinsic cellulase kinetics on hydrolysis of noncrystalline cellulose (NCC): Critical evaluation of initial rate mechanistic models for enzymatic hydrolysis Independent inhibition studies with reaction intermediates and products Comprehensive intrinsic kinetic studies of cellulase hydrolysis of NCC To carry out this sequential study, various enzymatic hydrolysis experiments were performed on NCC and R-cellulose.

2. Materials and Methods


2.1. Enzyme. Cellulase enzyme, Spezyme CP (Genencor, lot no. 301-00348-257), was obtained from NREL and had an average activity of 31.2 filter paper units (FPU)/mL. The enzyme solutions were diluted to 1, 3, and 15 FPU/mL by adding appropriate amounts of buffer solutions. 2.2. Substrate. R-Cellulose and NCC were used as substrates in multiple enzymatic hydrolysis experiments to investigate the intrinsic enzymatic kinetics. R-Cellulose: R-Cellulose is a pure form of cellulose. It was purchased from Sigma, St. Louis, MO (no. C8802). The composition as determined by NREL standard procedure (Sluiter et al., 2005 and 2006) was 76.58% glucan, 21.81% xylan, and traces of arabinan and mannan. NCC: Cotton (supplied by Buckeye Tech, Memphis, TN) and R-cellulose, which are pure forms of cellulose, are used as raw materials for producing noncrystalline cellulose (NCC). NCC was prepared in the Biofuels/Biomass Engineering Laboratory of Auburn University by a patent pending procedure (U.S. Provisional Patent Application 60/762,439). NCC contains 87.27% glucan, 10.51% xylan, 1.7% ash, and a small amount mannan.

2.3. Enzymatic Hydrolysis. The hydrolysis of cellulose was performed in 250 mL shake flasks with a 100 mL working volume. One gram of glucan (1% w/v dry basis) was taken as the basis for each flask, and 0.4 mL of tetracycline (10 mg/mL in 70% ethanol) and 0.3 mL of cyclohexamide (10 mg/mL in distilled water) were added as antibiotics to prevent any type of growth. Sodium citrate buffer (0.05 M) was used to bring the final working volume to 100 mL. The pH of 4.5 was assumed to be maintained throughout the reaction because of the buffer addition. All components were assumed to have a density of 1 g/mL in the flask. Substrate blanks and enzyme blanks were run to account for any glucose contribution from the samples and any protein from enzymes. The flasks were heated for 1 h at 50 C before the addition of the 1 mL of cellulase enzyme, Spezyme CP. A New Brunswick Scientific (Edison, NJ) Series 25 incubator shaker was used for agitation and temperature control. The flasks were maintained in an incubated shaker at 50 C and 150 rpm throughout the experiments. Samples were taken at 0, 0.25, 0.5, 0.75,1, 2, 3, 4, 5, 6, 12, 24, 36, 48, 72, and 96 h and boiled for 5 min to inactivate the enzyme, thus confirming the cessation of the reaction. Then, the samples were centrifuged and analyzed for glucose, cellobiose, and higher cello-dextrins using HPLC. The samples, after the carbohydrate analysis and the enzyme hydrolysis, were analyzed for sugars using HPLC equipped with RI detector and Bio-Rads Aminex HPX-87P column maintained at 85 C with DI water as the mobile phase. Glucose and cellobiose were taken as the standards for measuring glucose and cellobiose concentrations, respectively, whereas all of the other peaks that have lower retention times as compared to cellobiose and glucose were lumped together as cello-oligosaccharides and their concentration was measured against glucose standard. Remaining cellulose was calculated by subtracting glucose, cellobiose, and cello-oligosaccharides concentration from initial cellulose loading. 2.4. Critical Evaluation of Initial Rate Mechanistic Models for Enzymatic Hydrolysis. In this work, a comparative study was carried out using different mechanistic models available in the literature to capture the initial stages of hydrolysis using R-cellulose as the substrate for enzymatic hydrolysis. The six classes of initial rate mechanistic models tested are tabulated in Table 1. These six categories of the models cover the gamut of mechanistic models of cellulose hydrolysis. Although the model structures were same, in some cases as the underlying assumptions were different in deriving them, the constants were interpreted differently. In other cases, the models were applied multiple times to each enzyme and substrate component (Brown, 1990). A critical evaluation of these models was done to test their effectiveness in predicting the long-range kinetics of R-cellulose hydrolysis. 2.5. Independent Inhibition Studies with Reaction Intermediates and Products. Enzymatic hydrolysis of lignocellulosic biomass depends on competing effects such as physical properties of the substrate, enzyme synergy, mass transfer, and intrinsic kinetics. These effects cannot be distinguished from each other. Accurate estimation of intrinsic kinetics requires a pure form of cellulosic biomass to minimize the mass transfer resistances and effects of physical properties of the substrate and enzyme complex. For this purpose, enzymatic hydrolysis was performed on amorphous noncrystalline cellulose. In this part of the study, inhibitory effects of soluble cello-oligosaccharides, cellobiose, and glucose on enzymatic hydrolysis of NCC were quantified independently. These reaction intermediates and products were externally added to the substrate (NCC) initially. Later, the

628

Biotechnol. Prog., 2007, Vol. 23, No. 3

Table 1. Initial Rate Mechanistic Models Tested with the Estimated Model Parameter Values for r-Cellulose Hydrolysis and R2 for the Initial Stage of Hydrolysis model (A) Michales-Menton (MM kinetics with competitive/noncompetitive inhibition, with/without quasi-steady state approach (Ghose and Das, 1971; Dwivedi and Ghose, 1979; Howell and Stuck, 1975; Gan et al., 2003) dS -kSE ) dt Km + S (B) shrinking site hydrolysis model with Langmuir-type adsorption isotherm (Humphrey, 1979; Movagharnejad et al., 2003) dS -kS4/3E ) dt R+E (C) two phases of cellulose: amorphous and crystalline (Wald et al., 1984) dS - kSE ) dt R + E (D) hydrolysis of both cellulose and cellobiose: (Fan and Lee, 1983) dS KSE ) -k dt R+S (E) MM kinetics with competitive inhibition and Langmuir adsorption (Huang, 1975) dS -kSE ) dt R + S + E (F) MM kinetics with noncompetitive inhibition and Langmuir adsorption (Holtzapple et al., 1984; Brown and Holtzapple et al., 1990) (S - E - R) + dS -kSE where ) ) dt R + S + E (R + E - S)2 + 4RS 2S k ) 0.076 g/L R ) 39.67 g/(gh) 0.9824 K ) 0.0356 (L/g)(1/3)h-1 R ) 39.76 (g/L) 0.9846 estimated model parameter value k ) 0.0136 g/(gh) Km ) 43.12 g/L R2 0.9696

k ) 0.094 g/(Lh) K ) 0.077 g/(gh) R ) 83.12 g/L K ) 0.0786 g/(gh) R ) 41.25 g/L ) 0.95 g/g K ) 0.0726 g/(gh) R ) 37.97 g/L ) 0.898 g/g

0.9915

0.9919

0.9931

cellulose concentration profiles were studied after introducing the enzyme complex to the substrate solution. The series of experiments were designed in such a way that the initial concentration of one of the components (cello-oligosaccharides, cellobiose, or glucose) supplemented with NCC was varied (0%, 5%, and 10% w/w) while keeping the others constant. Idea behind this strategy was that the resultant variations in the hydrolysis rates may be solely attributed to the constituent whose composition was varied. 2.6. Comprehensive Intrinsic Kinetic Model of Cellulase Hydrolysis of NCC. The focus of this study was not to propose altogether a new phenomenological reaction mechanism, but rather to qualitatively and quantitatively analyze the underlying steps in enzymatic hydrolysis and to develop a well-grounded understanding of the controlling factors of intrinsic kinetics. The analysis presented in this work highlights the dynamically changing reaction rates, inhibitory effects of reaction intermediates and products (cello-oligosaccharides, cellobiose, and glucose), and variability in available active enzyme. The saturating kinetics in a finite batch time was also considered. A simplified mechanism of the hydrolysis of cellulose (NCC) is given by the schematic shown in Figure 1A. Cellulose is broken down to smaller chain length cello-oligosaccharides (insoluble ) degree of polymerization >7, soluble ) degree of polymerization <7; Gray et al., 2003) by the action of endo-glucanases. Further breakdown of insoluble cello-oligomers to glucose-dimer (cellobiose) is catalyzed by exo-glucanases. -Glucosidases act on both soluble oligosaccharides and cellobiose and convert them to fermentable sugar (glucose) (Beldman et al., 1985; Henriksson, 1996; Valijamae et al., 1998). The following assumptions were made to simplify the mechanism and derive the pertinent mathematical model: The cellulase system (E) of endo-glucanase (E1), exoglucanase (E2), and -glucosidase(E3) is considered as having

a constant composition for the given complex. These individual enzymes may be independently inhibited by the reaction intermediates and products. The reducing sugars inhibit the enzyme in a reversible and competitive/noncompetitive manner (Gusakov and Sinitsyn, 1992; Holtzapple et al., 1990). Cellulase adsorption to the substrate surface is reversible and is governed by a simple Langmuir-type adsorption isotherm (Huang, 1975; Lee and Fan, 1982; Mandels et al., 1971; Moloney and Coughlan, 1983). Cellulose and insoluble cello-oligosaccharides possess similar inhibitory effects on enzymes, and also their hydrolysis kinetics is assumed to be similar.

Figure 1. (A) Schematic showing the simplified mechanism of the enzymatic hydrolysis of NCC. C ) noncrystalline cellulose, S ) insoluble cello-oligosaccharides, O ) soluble cello-oligosaccharides, B ) cellobiose, G ) glucose, E1 ) endo-glucanases, E2 ) exoglucanases, E3 ) -glucosidases. (B) Schematic of NCC hydrolysis mechanism after further simplification

Biotechnol. Prog., 2007, Vol. 23, No. 3

629

Resistances offered by the crystallinity and composition variation with respect to the degree of hydrolysis were neglected, as these studies were carried out on noncrystalline cellulose. As the time scale of hydrolysis is much larger than the time scale of bulk diffusion of enzyme, mass transfer resistances were considered to be negligible (Fan et al., 1981; Fan and Lee, 1983; Lynd et al., 2002). Along with the simplifying assumptions that were stated earlier, it was also assumed that the enzymes catalyzing each reaction step are inhibited by different reaction intermediates and products, as they are distinct in their action and behavior, even though they were considered to be a single complex quantitatively. The following are the detailed inhibitory mechanisms proposed: Enzymes E1 and E2 were subjected to noncompetitive inhibition by soluble cello-oligosaccharides, cellobiose, and glucose. It was observed that as the glucose concentration increases, the inhibition rates of E1 and E2 increase exponentially. From the numerical experiments it was found that the probability of glucose (inhibitor) binding to the enzyme is three times higher than the probability of substrate binding. Enzyme, E3 was solely inhibited by glucose in competitive inhibition. At this juncture, the mechanism may be redrawn as shown in Figure 1B. As the properties of insoluble cello-oligosaccharides and cellulose are assumed to be the same, conversion of cellulose to cellobiose was lumped into a single step. From the aforementioned discussions, the rate of each reaction can be written as follows:

dC dO dB ) -r1 - r3; ) r 3 - r4 ; ) r1 - r2; dt dt dt dG ) r2 + r4 (2) dt


The detailed development is given in the Appendix. Model parameters were estimated using a nonlinear trajectory optimization as explained in the following section. 2.7. Model Parameter Estimation using Nonlinear Trajectory Optimization. The model represented earlier can be written as

dx ) f(x,) dt

(3)

Here, x is the concentration vector that encompasses cellulose, soluble cello-oligosaccharides, cellobiose, and glucose; represents the vector of model parameters. Integrating this differential equation(s) yield time-course data of concentrations. Hence, the predicted concentration vector of the enzymatic hydrolysis reaction components can be represented as

x(t) )

0t f(x,)

(4)

where, x(t) is the predicted concentration vector of cellulose, soluble cello-oligosaccharides, cellobiose, and glucose. The objective function that is to be minimized to solve for the model parameters is given as
tbatch

Min

[x(t) - x(t)]2 t)0 (5)

r1 )

(KC + C) 1 +

( (

k1CE B O G + + KBI KOI K3


3

GIn

) )

subjected to L e e H

(E + KE)

r2 )

( (

k2BE G KB 1 + + B (E + KE) KGI

) )

In this work, integration of differential equations representing the model equations is performed using the ode45 routine in MATLAB. On the outer frame work, the nonlinear constrained optimization is performed using the fmincon routine. After this algorithm converged for each data set, optimal set of parameters that yielded the predictions closest to the experimental values were obtained.

r3 )

k3CE

3. Results and Discussions


3.1. Evaluation of Initial Rate Mechanistic Models. Enzymatic hydrolysis data of R-cellulose was used to estimate the model parameters of the initial rate mechanistic models. On performing trajectory optimization using nonlinear constrained optimization on the hydrolysis data of different substrates at three different enzyme loadings, model parameters were obtained. These are tabulated in Table 1 along with the R2 values of each model for initial range predictions. Except for the MM model, all of the other five models represented the initial hydrolysis rates effectively, which is evident from the circled portion of Figure 2. Even though these models can explain the initial progression of the enzymatic hydrolysis very well, they fail to predict the later stages of the hydrolysis. This can be seen in Figure 2; the model predictions represented by solid or dashed lines show larger deviations from the experimental results shown by triangular markers. Initial rate mechanistic models assume 100% hydrolysis, as they do not consider the decelerating reaction rate due to the increasing enzymatic inhibition with the increase in hydrolysis time. Errors result from neglecting the effects of hydrolysis intermediates such as cellooligosaccharides and cellobiose on the enzymatic kinetics.

(KC + C) 1 +

B O G3 + + 3 (E + KE) KBI KOI K


GIn

r4 )

( (

k4OE G KO 1 + + O (E + KE) KGI

) )

(1)

In these rate equations, ki (i ) 1, 2, 3, 4) are the individual reaction rate constants (g/gmin); KC, KB, and KO are cellulose saturation constant (g/L), cellobiose saturation constant (g/L), and soluble cello-oligosaccharides saturation constant (g/L), respectively; KBI, KOI, and KGI are inhibition constants (g/L) of cellobiose, soluble cello-oligosaccharides and glucose, respectively; KE is the desorption equilibrium constant (g/L) for cellulases, from the NCC surface; and C, O, B, and G are concentrations (g/L) of NCC, soluble cello-oligosaccharides, cellobiose, and glucose, respectively. Further, the accumulation rates of cellulose, soluble cellooligosaccharides, cellobiose, and glucose are written as

630

Biotechnol. Prog., 2007, Vol. 23, No. 3

Figure 2. Performance of mechanistic models under study for long-range enzymatic hydrolysis of R-cellulose at 1 FPU/g-glucan enzyme loading. Experimental data is represented by triangles, and various model predictions are represented by continuous lines. Circled portion of the plots is the initial stage of the hydrolysis (0-6 h).

Figure 3. Enzymatic hydrolysis of NCC substrate with initial addition of (A) soluble cello-oligosaccharides (COS), (B) cellobiose (B), and (C) glucose (G) in different proportions with an enzyme loading of 1 FPU/g-glucan.

Hence, there is a need of accommodating the increasing inhibition rates, while modeling the kinetics of enzymatic hydrolysis. 3.2. Independent Inhibition Studies. To study the individual contributions of cello-oligosaccharides, cellobiose, and glucose on the inhibition of cellulase, a fixed set of enzymatic hydrolysis experiments were performed on NCC supplemented with cellooligosaccharides, cellobiose, or glucose as substrate. Figure 3A shows cellulose hydrolysis in three different cases: with pure NCC, addition of 5% (w/w) cello-oligosaccharides, and addition of 10% (w/w) cello-oligosaccharides at 1 FPU/g-glucan enzyme loading. It is evident from Figure 3A that at higher initial oligomer concentrations the initial reaction hydrolysis rate lasted for a relatively smaller time duration and also resulted in reduced extent of hydrolysis after 96 h of hydrolysis reaction. It can be

inferred from these observations that oligomers strongly inhibit the hydrolysis rates. Figure 3B and C were plotted for different initial compositions of substrate (NCC) with cellobiose and glucose, respectively with enzyme loadings of 1 FPU/g-glucan in each case. In both cases, the effect of external additions was not distinguishable as the rate and extent of hydrolysis with the increase in either cellobiose or glucose additions are found to be somewhat random. At 5% (w/w) addition of cellobiose, the higher initial rate prolonged for longer time and the extent of hydrolysis was also higher. However, further increase in external cellobiose addition to 10% (w/w) resulted in reduction in the final hydrolysis extent compared to the pure NCC case. In case of 5% (w/w) glucose addition, the hydrolysis extent is slightly higher than for pure NCC hydrolysis, whereas 10% (w/w)

Biotechnol. Prog., 2007, Vol. 23, No. 3


Table 2. Parameter Values of Proposed Comprehensive Kinetic Model for Cellulase Hydrolysis of NCC no. 1 2 3 4 5 6 7 8 9 10 11 12 parameter k1 (g/gmin) k2 (g/gmin) k3 (g/gmin) k4 (g/gminn) KC (g/L) KB (g/L) KO (g/L) KOI (g/L) KBI (g/L) KGI (g/L) KE (g/L) KGIn (g/L) description rate constant rate constant rate constant rate constant saturation constant for NCC saturation constant for cellobiose saturation constant for COS inhibition constant for COS inhibition constant for cellobiose inhibition (competitive) constant for glucose enzyme desorption constant inhibition (noncompetitive) constant for glucose value 38.29 32.92 20.62 14.84 9.35 13.40 14.28 8.69 5.20 0.08 0.04 0.43

631
Table 3. Experimental Planning for Carrying out the Comprehensive Intrinsic Kinetic Studies of Cellulase Hydrolysis of NCC experiments set 1 2 3 4 5 with 1 FPU/g-glucan enzyme pure NCC NCC + 5% w/w COS NCC + 10% w/w COS NCC + 5% w/w cellobiose NCC + 5% w/w glucose with 3 FPU/g-glucan enzyme pure NCC NCC + 5% w/w COS NCC + 10% w/w COS NCC + 5% w/w cellobiose NCC + 5% w/w glucose

glucose addition resulted in reduction in the extent of hydrolysis. This may be due to the following reasons: Along with the cellulose decomposition, these additions may be triggering some intermediate steps in the cellulose hydrolysis to form glucose. Due the accumulation of cello-oligomers, cellobiose, and glucose, the effect of initial addition of either cellobiose or glucose was not distinct.

Initial high concentrations of glucose/cellobiose may inhibit the cellulose flux toward soluble cello-oligomers and hence may reduce the inhibition on cellulases, resulting in higher extent of hydrolysis. This study on external additions of cello-oligosaccharides, cellobiose, and glucose exemplified the need to develop a comprehensive method to investigate the inhibitory effects of all these compounds together, as these interactions are interdependent in nature. 3.3. Comprehensive Intrinsic Kinetic Studies of Cellulase Hydrolysis of NCC. The reaction kinetics of enzymatic cellulose hydrolysis is subjected to mediation by a host of factors such as intermediate and product inhibitions, enzyme-substrate

Figure 4. Experimental and predicted concentration profiles of cellulose, soluble cello-oligosaccharides, cellobiose, and glucose in the enzymatic hydrolysis of (a) pure noncrystalline cellulose (R2 ) 0.992), (b) noncrystalline cellulose with 5% (w/w) cello-oligosaccharides (R2 ) 0.9893), (c) noncrystalline cellulose with 5% (w/w) glucose (R2 ) 0.9799), (d) noncrystalline cellulose with 5% (w/w) cellobiose with 1 FPU/g-glucan enzyme loading (R2 ) 0.9848). Experimental data is shown by discrete points, whereas predictions are shown by continuous or dashed lines.

632

Biotechnol. Prog., 2007, Vol. 23, No. 3

Figure 5. Experimental and predicted concentration profiles of cellulose, soluble cello-oligosaccharides, cellobiose, and glucose in the enzymatic hydrolysis of (a) pure noncrystalline cellulose (R2 ) 0.9884), (b) noncrystalline cellulose with 5% (w/w) cello-oligosaccharides (R2 ) 0.9908), (c) noncrystalline cellulose with 5% (w/w) glucose (R2 ) 0.9760), (d) noncrystalline cellulose with 5% (w/w) cellobiose (R2 ) 0.9821) with 3 FPU/ g-glucan enzyme loading. Experimental data is shown by discrete points whereas predictions are shown by continuous or dashed lines.

adsorption, etc. The influence of each factor is difficult to quantify in isolation, as many factors are interrelated during the hydrolytic reaction. As discussed in section 2.6, in synthesizing a mathematical representation of the hydrolytic reaction kinetics, a strategy was adopted to incorporate vital information with respect to the reaction mechanism but without unnecessarily complicating the model equations for all the interwoven events in the complex heterogeneous reaction. The goal was to obtain parsimony model equations, which describe the overall kinetic behavior of cellulose hydrolysis. Initially, the parameters for the proposed model (section 2.6) were identified on the basis of the data obtained from cellulase hydrolysis of pure NCC at two enzyme loadings (1 FPU/gglucan and 3 FPU/g-glucan) using nonlinear constrained trajectory optimization (section 2.7). These model parameters are listed in Table 2. Later, the proposed methodology was validated against different sets of experimental data. The experimental design is presented in Table 3. From Figures 4 and 5, it is evident that for all of the cases the proposed model effectively predicted the concentration profiles of cellulose, cello-oligosaccharides, cellobiose, and glucose which were in close agreement with experimental data. The R2 value in each case is provided with Figures 4 and 5. From these high R2 values, we believe

that this model can explain the entire enzymatic hydrolysis batch with high precision. 3.4. Sensitivity Analysis. Further investigation was carried out to explore the controlling factors of the enzymatic hydrolysis of cellulose by performing sensitivity analysis on the developed mathematical model. This study was performed to formulate the theoretical directives for enzyme modifications and substrate preparation to increase the extent of hydrolysis. Every kinetic parameter was varied between two limits ((100% of their nominal values), and the variation in the final extent of hydrolysis to glucose was observed. Figure 6 shows the varying final glucose concentration (g/L) in the medium, which was initially loaded with 11 g/L of NCC, with respect to the change in each kinetic parameter of the proposed model. Vertical lines represent the nominal model parameter values. It can be observed that the increase in reaction rate constants k1 and k3 leads to the higher overall yield of glucose until a critical value, and then the extent of hydrolysis is either constant or it decreases. As k1 and k3 are the rates corresponding to the reactions cellulose to cellobiose and cellulose to soluble cellooligosaccharides, respectively, these parameters determine the activity of the endo and exo-glucanases directly. As seen from the graphs, a 4-fold increase in k1 (from 38.29 to 153.16) results

Biotechnol. Prog., 2007, Vol. 23, No. 3

633

Figure 6. Effect of kinetic parameter variation on ultimate glucose yield for an NCC loading of 11 g/L.

in approximately 30% increase in the extent of hydrolysis, whereas similar change in k3 (from 32.92 to 128.7) results in an increase of 15%. Increase or decrease in k2 and k4 from nominal parametric values (32.92 and 14.83, respectively) results in reduction in the extent of overall cellulose hydrolysis. If the reaction rate decreases, rate of formation of glucose also decreases. Increase in the reaction rate constants results in a higher amount of glucose formed, resulting in higher inhibition rates and hence reduced extent of hydrolysis. As these rate constants correspond to the reactions catalyzed by -glucosidases, it may be concluded that further enzymatic modifications are not required to improve the catalytic activity of -glucosidases. However, as seen from the final glucose concentration versus KGI plot, if KGI is increased from the nominal value, the yield of glucose on the hydrolysis of cellulose increases in an logarithmic manner. This indicates that there is a scope for improvement in -glucosidases; making them inaccessible for glucose will ensure higher extent of hydrolysis. From a plot of final glucose concentration versus enzyme desorption constant KE, it is evident that as the value of KE decreases, final glucose concentration increases, denoting a higher extent of hydrolysis. Hence, if any surface-modifying agents such as surfactants are added to the reaction medium, they lessen the desorption constant and make the reaction

proceed further to yield higher final glucose concentration. Plots of KC, KB, KO, KOI, and KBI show that nominal values of these parameters lie at the optima, and any change in these parameter values results in reduction in the final extent of hydrolysis. Parametric sensitivity analysis studies how a variation in the model output can be apportioned to the variation of the different parameters. Let y ) f(x,) is the equation relating output y with an input vector x. is the parameter vector. Sensitivity index of the ith parameter can be given as

S )

dy y di

The individual parameter sensitivity indices over different time points of the hydrolysis reaction are estimated. A subset of these parameters, which showed noticeable patterns in their sensitivity indices over the reaction time, is plotted in Figure 7. k3 and KC were also presented to draw a comparison with the normal behavior of the parameters, which is characterized by a sudden drop in the absolute value of the sensitivity index in the initial stage of hydrolysis. The sensitivity of KGI dropped to 30% of its initial value over the reaction time. k2 and KB showed an interesting behavior in their sensitivity indices. Surprisingly, as time of hydrolysis increased, the sensitivity

634

Biotechnol. Prog., 2007, Vol. 23, No. 3

Figure 7. Variation in parametric sensitivity with batch time

Figure 8. Ultimate glucose yield at different enzyme compositions (x1 ) mole fraction of endo-glucanases; x2 ) mole fraction of exo-glucanses; x3 ) 1 - x1 - x2 is the mole fraction of -glucosidases) for the cellulose substrate loading of 11.49 g/L.

indices of some parameters changed signs. At the initial stages of the reaction, increase in the reaction rate constant of reaction 2 increases the glucose formation, whereas at the end of the batch a similar increase in k2 results in a decrease of the extent of hydrolysis. Similar behavior is observed with KB but in the opposite direction. These sign changes in the sensitivities were observed at 40 h of reaction time. From the experimental observations, it was found that after this time the concentration of cello-oligosaccharides was almost constant with no further change and cellobiose reached a minimal level. It is logical that after this time further increase in the reaction rate constant of B to G will not be effective as already the substrate for this reaction (cellobiose) is in negligible concentrations. This observation suggests that increase of the cellulose flux toward glucose through reaction 4 (O to G) in the later stages of hydrolysis, instead through reaction 2, will be beneficial. 3.5. Chemical Modifications To Improve Glucose Yield. There are two ways to improve the glucose yield in enzymatic hydrolysis of cellulose. The first one is chemical or physical modifications that can be done readily, and the second one is enzymatic modifications. In this work, the effect of change in enzyme complex composition and addition of surfactant was studied, which falls under the first category. Let us consider that the mole fractions of endo-glucanses, exo-glucanases, and -glucosidases are x1, x2, and x3, respec-

tively. For known composition of enzyme complex, the model equations presented earlier can be modified as

r1 )

(KC + C) 1 +

k1CE B O G3 + + 3 KBI KOI K k2BE

GIn

)(

E+

K1E (x1 + x2)

r2 )

((

KB 1 +

K2E G +B E+ KGI x3 k3CE

) )(

)
K3E x2

r3 )

(KC + C) 1 +

O G3 B + + 3 KBI KOI K k4OE

GIn

)(

E+

)
(6)

r4 )

((

K4E G KO 1 + +O E+ KGI x3

) )(

where K1E ) 0.97KE, K2E ) 0.03KE, K3E ) 0.8KE, and K4E ) 0.03KE. The enzyme complex composition used for the experiments to generate model parameter analyses are x1 ) 0.17, x2

Biotechnol. Prog., 2007, Vol. 23, No. 3

635

Figure 9. Effect of addition of surfactant (T-20) to the medium on enzymatic hydrolysis of NCC: (A) enzymatic hydrolysis of NCC, (B) enzymatic hydrolysis of sonicated NCC, and (C) enzymatic hydrolysis of sonicated NCC in the presence of surfactant.

) 0.8, and x3 ) 0.03. By varying the composition of the enzyme, the yield of glucose at 98 h of reaction time for a cellulose substrate loading of 11 g-glucan/L with an enzyme loading of 3 FPU/g-glucan is plotted in Figure 8. The regions separated by contours shown in the figure denote the enzyme complex compositions that give the same extent of hydrolysis. As seen from the figure, the current enzyme complex lies in region A; this is characterized as the region of enzyme complex mixture that yields higher hydrolysis extent. This shows that further chemical composition modifications in the enzyme complex may not result in any significant increase in the glucose yield. As discussed earlier in section 3.4, reduction of the desorption constant of the enzyme-substrate complex greatly affects the glucose yield. To validate this hypothesis experimentally, enzymatic hydrolysis was carried out in the presence of surfactant Tween-20 (1% w/w) along with sonication. Figure 9 shows the progress of enzymatic hydrolysis of NCC, sonicated NCC, and sonicated NCC with a small addition of Tween-20. The results of hydrolysis of sonicated NCC indicated very high initial hydrolysis rates when compared to the original NCC. The total digestibility for 96 h was reported to be 20% higher than that of unsonicated NCC. This indicates that breaking the particle size and making a homogeneous reaction mixture would favorably affect the enzyme-substrate reaction rate. The accumulation of intermediates was less when surfactant was added and most of the cellulose flux was routed to glucose even though the total hydrolysis remained the same. These results were in agreement with Eriksson et al. (2002) and Karr and Holtzapple (1998).

4. Conclusions
Existing kinetic models may explain the progression of enzymatic hydrolysis of cellulose in the initial stages of hydrolysis, but they fail to model later stages of the hydrolysis process. From a series of controlled experiments, it was found that independently quantifying the inhibitory effects of hydroly-

sis intermediates (soluble cello-oligosaccharides and cellobiose) and product (glucose) on cellulases, is not possible, as these effects are simultaneous in nature and cannot be separated from each other. This investigation coupling experimental methods with mathematical modeling and simulation analysis highlights the changing reaction dynamics of batch cellulose hydrolysis, which is influenced by substrate binding of enzyme and competitive/ noncompetitive product inhibition. As a result of revisiting the extensively studied subject with new analysis and experiments, some new and additional understanding of the enzymatic reaction kinetics are offered: Cellulose breaks down to smaller chain length cellooligosaccharides (insoluble ) degree of polymerization > 7, soluble ) degree of polymerization < 7) by the action of endoglucanases. Further breakdown of insoluble cello-oligomers to glucose-dimer (cellobiose) is catalyzed by exo-glucanases. -Glucosidases act on both soluble oligosaccharides and cellobiose and converts them to fermentable sugar (glucose). Cellulase adsorption to the substrate surface is reversible and is governed by a simple Langmuir-type adsorption isotherm. The reducing sugars inhibit the enzyme in a reversible and competitive/noncompetitive manner Enzymes endo-glucanases and exo-glucanases were subjected to noncompetitive inhibition by soluble cello-oligosaccharides, cellobiose, and glucose. As the glucose concentration increases, the inhibition rate of endo-glucanases and exo-glucanases increases. The probability of glucose (inhibitor) binding to the enzyme is three times higher than the probability of substrate binding. Enzymes, -glucosidase were solely inhibited by glucose through a competitive inhibition mechanism. With NCC as substrate, accumulation of significant amounts of cellobiose and soluble cello-oligosaccharides were observed. By sensitivity analysis, it was found that the enzyme complex is already at optimal composition for most cases, and further

636

Biotechnol. Prog., 2007, Vol. 23, No. 3

increase in the glucose yield by altering its composition may not be feasible. Adding surface active agents to the medium improved the enzyme-substrate adsorption, which resulted in lower accumulation rates of intermediates and higher yield of glucose. Enzyme modifications using protein engineering principles to enhance their activity by altering the properties, such as substrate binding at active site, access to the inhibitors, and allosterism, are out of the scope of current work. However, the understanding of the underlying mechanism and the nature of enzyme-substrate interaction at the macroscopic level is as a starting point for such future studies.

r2 )

( (

k2BE ; G KB 1 + + B (E + KE) KGI

) )

r4 )

( (

k4OE (A4) G KO 1 + + O (E + KE) KGI

) )

Finally, the accumulation rates of all reaction components can be written by

Appendix
The reactions involved in the enzymatic hydrolysis of cellulose are plotted in Figure 1. They can be listed as follows: 1. Cellulose (C) to cellobiose (B) 2. Cellobiose (B) to glucose (G) 3. Cellulose (C) to soluble cello-oligosaccharides (O) 4. Soluble cello-oligosaccharides (O) to glucose (G) Assuming Michael-Menton-type kinetics for each reaction step, the rates of these reactions may be given by

dC ) -r1 - r3; dt

dO ) r3 - r4; dt

dB ) r1 - r2; dt dG ) r2 + r4 (A5) dt

Acknowledgment
This research was partially funded by DOE (Project DE-PS3600GO10482, a subcontract through Dartmouth College).

References and Notes


Beldman, G.; Searle-van Leeuwen, M. F.; Rombouts, F. M.; Voragen, F. G. J. The cellulase of Trichoderma viride: purification, characterization and comparison of all detectable endoglucanases, exoglucanases, and h-glucosidase. Eur. J. Biochem. 1985, 146, 301- 308. Brown, R. F.; Holtzapple, M. T. A comparison of the MichaelisMenten and HCH-1 Models. Biotechnol. Bioeng. 1990, 36, 11511154. Brown, R. F. 1990, Kinetics of the enzymatic hydrolysis of AFEXtreated wheat straw. Ph.D. Thesis, Texas A&M University, College Station, TX. Chang, V. S.; Holtzapple, M. T. Fundamental factors affecting biomass enzymatic reactivity. Appl. Biochem. Biotechnol. 2000, 84/86, 5-37. Dwivedi, C. P.; Ghose, T. K. A model on hydrolysis of bagasse cellulose by enzyme from Trichoderma reesei QM 9414. J. Ferment. Technol. 1979, 57, 15. Eriksson, T.; Borjesson, J.; Tjerneld, F. Mechanism of surfactant effect in enzymatic hydrolysis of lignocellulose. Enzyme Microb. Technol. 2002, 31, 353-364. Fan, L. T.; Lee, Y. H. Kinetic studies of enzymatic hydrolysis of insoluble cellulose: Derivation of a mechanistic kinetic model. Biotechnol. Bioeng. 1983, 27, 1328. Fan, L. T.; Lee, Y. H.; Beardmore, D. R. The influence of major structural features of cellulose on rate of enzymatic hydrolysis. Biotechnol. Bioeng. 1981, 23, 419-424. Gan, Q.; Allen, S. J.; Taylor, G. Kinetic dynamics in heterogeneous enzymatic hydrolysis of cellulose: an overview, an experimental study and mathematical modeling. Process Biochem. 2003, 38, 1003-1018. Ghose, T. K.; Das, K. 1971, A simplified kinetics approach to cellulosecellulase system. In AdVances in Biochemical Engineering; Ghose, T. K., Fiechter, A., Eds.; Springer-Verlag: New York, 1971; No. 1. Gray, M. C.; Converse, A. O.; Wyman, C. E. Sugar monomer and oligomer solubility. Appl. Biochem. Biotechnol. 2003, 105-108, 179-193. Gusakov, A. V.; Sinitsyn, A. P. A theoretical analysis of cellulase product inhibition: effect of cellulase binding constant, enzyme/ substrate ratio, and h-glucosidase activity on the inhibition pattern. Biotechnol. Bioeng. 1992, 40, 663-671. Henriksson, H.; Stahlberg, J.; Isaksson, R.; Pettersson, G. The active site of cellulases are involved in chiral recognition: a comparison of cellobiohydrolase I and endoglucanase I. FEBS Lett. 1996, 390, 339-344. Holtzapple, M. T.; Caram, H. S.; Humphrey, A. E. The HCH-1 model of enzymatic cellulose hydrolysis. Biotechnol. Bioeng. 1984, 26, 775-780. Holtzapple, M. T.; Cognata, M.; Shu, Y.; Hendrickson, C. Inhibition of Trichoderma reesei cellulase by sugars and solvents. Biotechnol. Bioeng. 1990, 36, 275-287. Howell, J. A.; Stuck, J. D. Kinetics of solka floc cellulose hydrolysis. Biotechnol. Bioeng. 1975, 17, 873.

r1 )

k1CEeff (KC + C)

r2 )

k2BEeff (KB + B)

r3 )

; (KC + C) k4OEeff r4 ) (A1) (KO + O)

k3CEeff

Here, Eeff is the effective enzyme concentration that is bound to the substrate. Assuming simple Langmuir-type adsorption isotherms for enzyme substrate adsorption, effective enzyme concentration in terms of total enzyme concentration is given by

Eeff )

E (E + KE)

(A2)

Enzyme endo-glucanases and exo-glucanases were subjected to noncompetitive inhibition by soluble cello-oligosaccharides, cellobiose, and glucose. In addition, it was also found that glucose binding probability to the enzyme is three times higher than that of substrate binding. The presence of the noncompetitive inhibitors in the reaction medium lessens the rate constant of the reaction catalyzed by the enzyme (Shuler and Kargi, 2002). These effects can be incorporated into reaction rates r1 and r3 as shown by

r1 )

B O G3 (KC + C) 1 + + + 3 (E + KE) KBI KOI K


GIn

k1CE

r3 )

(KC + C) 1 +

k3CE

B O G3 + + 3 (E + KE) KBI KOI K


GIn

(A3)

Enzyme -glucosidases were solely inhibited by glucose through a competitive inhibition mechanism. Competitive inhibition alters the enzymatic reaction kinetics by increasing the enzyme saturation constant by a factor that is proportional to the inhibitor concentration (Shuler and Kargi, 2002). As reactions 2 and 4 are catalyzed by -glucosidases, these reaction rates can be modified to account for the competitive inhibition by glucose as shown by

Biotechnol. Prog., 2007, Vol. 23, No. 3


Howard, R. L.; Abotsi, E.; Rensburg, J. V. E. L.; Howard, S. Lignocellulose biotechnology: issues of bioconversion and enzyme production. Afr. J. Biotechnology 2003, 2 (12), 602-619. Huang, A. A. Kinetic Studies on Insoluble Cellulose-Cellulase System. Biotechnol. Bioeng. 1975, 17, 1421-1433. Humphrey, A. E. The hydrolysis of cellulosic materials to useful products. In Hydrolysis of Cellulose: Mechanisms of Enzymatic and Acid Catalysis; Brown, R. D., Jurasek, L., Eds.; Advances in chemistry series no. 181; American Chemical Society: Washington, DC, 1979. Klyosov, A. A. Trends in biochemistry and enzymology of cellulose degradation. Biochemistry 1990, 29, 10577-10585. Kaar, W. E.; Holtzapple, M. T. Benefits from Tween during enzymic hydrolysis of corn stover. Biotechnol. Bioeng. 1998, 59 (4), 419427. Lee, Y. H.; Fan, L. T. Kinetic studies of enzymatic hydrolysis of insoluble cellulose: analysis of the initial rates. Biotechnol. Bioeng. 1982, 24, 2383-2406. Lynd, L. R.; Weimer, P. J.; van Zyl, W. H.; Pretorius, I. S. Microbial cellulose utilization: fundamentals and biotechnology. Microbiol. Mol. Biol. ReV. 2002, 66, 506-577. Mandels, M.; Kostick, J.; Parizek, R. The use of adsorbed cellulase in the continuous conversion of cellulose to glucose. J. Polym. Sci. 1971, 36, 445-459. Moloney, A.; Coughlan, M. P. Sorption of Talaromyces-emersion II cellulase on cellulosic substrates. Biotechnol. Bioeng. 1983, 25, 271280. Movagharnejad, K.; Sohrabi, M. A. model for the rate of enzymatic hydrolysis of some cellulosic waste materials in heterogeneous solidliquid systems. Biochem. Eng. J. 2003, 14, 1-8.

637
Perez, L. L.; Teymouri, F.; Alizadeh, H.; Dale, B. E. Understanding factors that limit enzymatic hydrolysis of biomass. Appl. Biochem. Biotechnol. 2005, 121-124, 1081-1099. Shuler, M. L.; Kargi, F. Bioprocess Engineering-Basic Concepts, 2nd ed.; Prentice Hall Inc.: New Jersey, 2002. Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D.; Crocker, D. NREL Analytical Procedure, Determination of Structural Carbohydrates and Lignin in Biomass; National Renewable Energy Laboratory: Golden, CO, 2006 (Website: http:// devafdc.nrel.gov/pdfs/9572.pdf). Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D. NREL Analytical Procedure, Determination of Ash in Biomass; National Renewable Energy Laboratory: Golden, CO, 2005 (Website: http://devafdc.nrel.gov/pdfs/9338.pdf). Valjamae, P.; Sild, V.; Pettersson, G.; Johansson, G. The initial kinetics of hydrolysis by cellobiohydrolases I and II is consistent with a cellulose surface-erosion model, Eur. J. Biochem. 1998, 253, 469475. Wald, S.; Wilke, C. R.; Blanch, H. W. Kinetics of the enzymatic hydrolysis of cellulose. Biotechnol. Bioeng. 1984, 26, 221-230. Zhang, S.; Wolfgang, D. E.; Wilson, D. B. Substrate heterogeneity causes the nonlinear kinetics of insoluble cellulose hydrolysis. Biotechnol. Bioeng. 1999, 66, 35-41. Zhang, Y. H. P.; Lynd, L. R. Towards an aggregated understanding of enzymatic hydrolysis of cellulose: Noncomplexed cellulase systems, Biotechnol. Bioeng. 2004, 88 (7), 797-824. Received October 24, 2006. Accepted April 6, 2007. BP060322S

You might also like