Biotechnol. Prog.

2003, 19, 803−807 803

Covalent Coupling Method for Lipase Immobilization on Controlled

Pore Silica in the Presence of Nonenzymatic Proteins

Cleide M. F Soares,†,‡ M. Helena A. Santana,‡ Gisella M. Zanin,§ and

Heizir F. de Castro*,†
Department of Chemical Engineering, Faculty of Chemical Engineering of Lorena, PO Box 116,
12606-970, Lorena, SP, Brazil, Faculty of Chemical Engineering, UNICAMP, PO Box 6066,
13081 970, Campinas, SP, Brazil, and Department of Chemical Engineering, State University of Maringa,
Campus Universitário, 87020-900, Maringa, PR, Brazil

Candida rugosa lipase was covalently immobilized on silanized controlled pore silica
previously activated with glutaraldehyde in the presence of nonenzymatic proteins.
This strategy is suggested to protect the enzyme from aggregation effects or
denaturation that occurs as a result of the presence of silane precursors used in the
formation of the silica matrix. The immobilization yield was evaluated as a function
of the lipase loading and the additive type (albumin and lecithin) using statistical
concepts. In agreement with the mathematical model, the maximum coupling yield
(32.2%) can be achieved working at high lipase loading (450 units‚g-1 support) using
albumin as an additive. In these conditions, the resulting immobilized lipase exhibits
high hydrolytic (153.2 U‚mg-1) and esterification (337.6 mmol‚g-1‚min) activities. The
enhanced activity of the final lipase derivative is the sum of the benefits of the
immobilization (that prevents enzyme aggregation) and the lipase coating by additives
that increases the accessibility of active sites to the substrate.

Introduction facilitate mass transport when additives are used to-

Lipases have become a powerful catalyst tool in a wide
variety of chemical processes. For many applications
lipases are preferably used in an immobilized state in
order to easily separate the catalyst from the product
stream. One of the main issues concerning enzyme
immobilization is maintaining or eventually enhancing
the structural stability of the catalytic macromolecules
in view of long-term applications. The biocatalyst behav-
ior may indicate the efficiency of the immobilization
procedure and may also reflect the balance between the
acquired conformational stability created around the
biocatalyst. A promising approach to lipase immobiliza-
tion makes use of inorganic and hydrophobic matrixes,
such as controlled pore silica (CPS). In accordance with
this methodology silica is functionalized via silane cou-
pling with γ-(aminopropyl)triethoxysilane followed by
glutaraldeyde activation (Soares et al., 1999). The surface
activation with silane precursors seems to reduce the
original activity on the carrier, and the addition of
stabilizing additives is usually recommended for optimum
lipase performance (Reetz et al., 1996).
The effect of additives on the activity of lipase prepara-
tion is not well understood yet. Probably they act by a
combination of various effects, including (i) enzyme
protection from inactivation during the immobilization
step, (ii) retention of water layer around the catalyst, and
(ii) dispersing effects of the enzyme molecules that
* To whom correspondence should be addressed. Tel: + 55-12-
31595063. Fax: +55-12-553-3224. E-mail: heizir@dequi.faenquil.br.
† Faculty of Chemical Engineering of Lorena.
‡ UNICAMP.
§ State University of Maringa.
10.1021/bp025779q CCC: $25.00 © 2003 American Chemical Society and American Institute of Chemical Engineers
Published on Web 03/29/2003
gether with immobilizing matrixes (Rocha et al., 1998).
The kind of additive, its concentration, and the contact
time are critical parameters that have to be optimized
(Wehtje et al., 1993; Triantafyllou et al., 1995).
In the specific case of lipases, which demand an
interface for their total catalytic activity, the use of
macromolecular additives such as proteins, poly(ethylene
glycol), and poly(vinyl alcohol) (Reetz et al., 1996) have
shown stabilizing effects on the activity of the enzyme,
avoiding changes of protein structure. On the other hand,
the use of low molecular weight compounds such as
monomeric carbohydrates (sorbitol, aribitol) and polysac-
charides (dextran, starch) had no effect on the enzyme
activity (Wehtje et al., 1993; Bosley, 1991; Reetz et al.,
1996). Sometimes, the role of these additives is masked
by inert impurities included in commercial preparations
(Rocha et al., 1998).
In the present study, different immobilization proce-
dures were evaluated: a direct lipase binding on deriva-
tized silica and the simultaneous addition of lipase and
stabilizing additive on the same support. The effect of
nonenzymatic proteins (albumin and lecithin) was in-
vestigated for different lipase loading according to a 22
full factorial design with center face (Box et al., 1978).
These additives were chosen on the basis of their
compatibility with lipase and support as reported in the
literature (Reetz et al., 1996). The immobilized systems
on CPS in the presence of the additive were used both in
the hydrolysis of olive oil and in the synthesis of butyl
butyrate. Data were compared with those attained for
CPS-immobilized lipase lacking additive under the same
operational conditions.
804 Biotechnol. Prog., 2003, Vol. 19, No. 3

Materials and Methods Table 1. Factor Levels According to Experimental

Design
Materials. Commercial Candida rugosa lipase (Type levels
VII) was purchased from Sigma Chemical Co. (St. Louis,
MO). The lipase was a crude preparation with a nominal variables low (-) center (0) high (+)
specific activity of 1500 units‚mg-1 solid and 10% protein additive (x1) lecithin albumin
based on Bradford’s method for protein assay (Bradford, lipase loading (U‚g-1 support, x2) 150 300 450
1976). Controlled pore silica (CPS), supplied by Corning
Glass Works-USA, had the following characteristics: Table 2. Experimental Design and Coupling Yields
According to 22 Factorial Design
average particle porosity () 0.566; particle matrix density
(Fs) 2.178 g‚cm-3; particle density (dry) (Fp) g‚cm-3; lipase loading activity ηa
particle size 30 mesh (0.59 mm) containing pores of 375 run x1 x2 additive (U‚g-1 support) (U‚mg-1 support) (%)
Å (Soares et al., 1999). The silane γ-(aminopropyl)- 1 -1 -1 lecithin 150 26.7 18.0
triethoxysilane and glutaraldehyde (25% solution) came 2 +1 -1 albumin 150 38.3 23.6
from Sigma Chemical Co. Bovine albumin BSA (Sigma) 3 -1 +1 lecithin 450 46.1 10.8
and soy lecithin (Synth) were used as stabilizing agents. 4 +1 +1 albumin 450 153.2 32.2
5 -1 0 lecithin 300 51.3 15.4
Olive oil (low acidity) was purchased at a local market. 6 +1 0 albumin 300 79.5 25.9
Substrates for esterification reactions (n-butanol and 7 0 -1 absent 150 17.7 10.1
butyric acid from Merck) were dehydrated, with 0.32-cm 8 0 +1 absent 450 57.6 10.3
molecular sieves (aluminum sodium silicate, type 13 9 0 0 absent 300 50.3 17.3
X-BHD Chemicals, Toronto, Canada). Solvents were a Calculated according to eq 1.
standard laboratory grade and other reagents were
purchased either from Aldrich Chemical Co. (Milwaukee, coupling reaction mixture and the amounts of protein in
WI) or from Sigma Chemical Co. the filtrate and in the washing solutions.
Immobilization of Lipase on CPS. Lipase was Esterification Reactions. Reaction systems con-
immobilized by being covalently bound to CPS previously sisted of heptane (10 mL), n-butanol (250 mM), butyric
treated with γ-(aminopropyl)triethoxysilane followed by acid (250 mM), and immobilized lipase (0.85 g, dry
the reaction of the pretreated beads with glutaraldehyde weight). The mixture was incubated at 37 °C for 24 h
solution, according to the procedure previously described with continuous shaking at 150 rpm. The remaining
(Soares et al., 1999). Suitable amounts of enzyme (0.1- butanol and the product formed were determined by gas
0.3 g) were dissolved in 10 mL of distilled water and chromatography using a 6-ft 5% DEGS on Chromosorb
mixed with the support (1 g, dry weight) under low WHP, 80/10 mesh column (Hewlett-Packard, Palo Alto,
stirring for 2 h at room temperature. Albumin and CA) and hexanol as the internal standard. Esterification
lecithin were added together with the enzyme solution activity was expressed as micromoles of butyl butyrate
at a fixed amount (5 mg‚g-1 support, 200 µL of aqueous formed the per minute per gram of dry support.
solution containing 50 mg additive‚mL-1), 10 mL of Experimental Design and Statistical Analyses. A
hexane was added, and the enzyme-support-additive 22 full factorial central composite design leading to nine
mixture was incubated overnight at 4 °C. The im- sets of experiments was performed. The range and levels
mobilized lipase was filtered (Whatman filter paper 41) of the variables investigated in this study are given in
and thoroughly rinsed with hexane. Analyses of hydro- Table 1. The coupling yield was taken as the response of
lytic activities carried out on the lipase loading solution the design experiment. Statistica (version 5.0) software
and immobilized preparations were used to determine the was used for regression and graphical analyses of the
coupling yield (η%) according to eq 1: data obtained. The statistical significance of the regres-
sion coefficients was determined by the Student T test,
Us the second-order model equation was determined by
η (%) ) × 100 (1) Fischer’s test, and the proportion of variance explained
Uo by the model obtained was given by the multiple coef-
ficient of determination, R2.
where Us ) total activity recovered on the support and
Uo ) units offered for immobilization. Results and Discussion
Hydrolytic Activities. Hydrolytic activities were Effect of Additives on Lipase Immobilized De-
assayed by the olive oil emulsion method. The substrate rivatives. Candida rugosa lipase was covalently im-
was prepared by mixing 50 mL of olive oil with 50 mL of mobilized on silanized controlled pore silica (CPS) pre-
arabic gum solution (7% w‚v-1). The reaction mixture viously activated with glutaraldehyde at different loadings
consisting of 5 mL of emulsion, 2 mL of 100 mM sodium in the presence and absence of nonenzymatic proteins
phosphate buffer (pH 7.0), and immobilized (100-250 (lecithin or albumin). The aim was to observe if covalent
mg) lipase was incubated for 5 min at 37 °C (Soares et immobilization to a solid support and the presence of
al., 1999). The reaction was stopped by the addition of these proteins promote stability by inhibiting protein
10 mL of acetone-ethanol solution (1:1). The liberated fluctuations leading to unfolding. Albumin and lecithin
fatty acid was titrated with 25 mM potassium hydroxide are expected to exert the same behavior depicted by
solution using phenolphthalein as an indicator. One unit stabilizing agents along with the effects encountered in
(U) of enzyme activity was defined as the amount of chemical cross-linking of enzymes.
enzyme that produced 1 µmol of free fatty acid per min The experimental design and results are given in Table
under the assay conditions. 2. When no additive was used (runs 7-9), the coupling
Protein Assay. Protein was determined according to yields varied from 10.1% to 17.3% and the highest value
Bradford’s method (Bradford, 1976) using bovine serum was attained for lipase loading at 300 units‚g-1 support.
albumin (BSA) as a standard. The amount of bound The results suggest that for lipase loadings over 300
protein was determined indirectly by the difference units‚g-1 support, instead of obtaining the desired crowded
between the amount of protein introduced into the upright binding of enzyme onto the solid surface, mul-
Biotechnol. Prog., 2003, Vol. 19, No. 3
Figure 1. Comparative profile for the hydrolytic activities and
coupling yields for CPS immobilized lipase without additive (a)
and with additives (b,c) at different lipase loadings.
tilayer adsorption may occur, which could effectively
block or inhibit the access to enzyme active sites. Low
yields are typical of covalent immobilizations, as de-
scribed in several papers, including those published by
our research team (Triantafyllou et al., 1995; Gonçalves
et al., 1999; Soares et al., 1999).
The influence of nonenzymatic proteins on the lipase
immobilization can be better evaluated in Figure 1a-c,
805
in which values for lipase loadings were plotted against
the hydrolytic activities and efficiency (activity/loading).
By comparing the profiles for CPS-lipase (Figure 1a) and
CPS-additive-lipase derivatives (Figure 1b,c), one can
observe that albumin showed a stabilizing effect greater
than that of lecithin. While lecithin was efficient only for
the lowest lipase loading (150 units.g-1 support), albumin
had a positive effect for all tested lipase loadings (150-
450 units‚g-1 support). This effect was more pronounced
for the highest level (32.2%) than for the other two levels
(23.6% and 25.9%), with a maximum hydrolytic activity
of 153.2 U‚mg-1 support, which represents activity three
times higher than that obtained with the control.
In the case of lecithin, when the lipase loading was
increased from the low level (150 units‚g-1 support) to
the high level (450 units‚g-1 support), a decrease in the
coupling yield from 18.0% to 10.8% was verified, resulting
in hydrolytic activities similar to the one obtained
without additive (runs 7-9). The decrease in activity was
interpreted as a result of the properties of lecithin, which
is a highly viscous material and may restrict lipase
contact with the support surface (Rocha, 2000). It appears
that the internal diffusional limitations become profound
for higher lipase loadings due to the clumping of the final
derivative. Thus, the negative effect of lecithin on the
mechanical properties of the immobilized preparation
overwhelms the positive effect of this additive.
It is worth noticing that the lipase activity profile
determined titrimetrically against oil emulsions (as
described in Materials and Methods) is subject to many
interference factors, since the lipase activity is dependent
on the way the substrate is presented to the enzyme
(Rocha, 2000). A small percentage of surface-active
contaminants or hydrolysis products could markedly alter
the properties of the surface phase. As lecithin is a
zwitterionic surfactant, it is possible that the derivative
coated with lecithin might inhibit the lipase activity
mainly by changing the emulsion surface structure. This
would explain the low coupling yields observed when
lecithin was used as stabilizing additive. Similar results
were found by Bosley and Moore (1994) in a patent
description for immobilization of lipase on hydrophobic
support, such as Acurel, in the presence of several
surfactants. According to this invention, only nonionic
surfactants are suitable to achieve high activity and high
activity recovery of the immobilized derivatives. Applica-
tion of surfactants based on nonionic ethers or on anionic
or cationic surfactants led to a loaded immobilized
enzyme without a significant esterification activity.
To verify whether the same behavior can be found in
nonaqueous medium, the lipase derivatives prepared
from CPS at constant lipase loading (300 units‚g-1
support) in the presence and absence of additives were
subjected to esterification reactions. The esterification
activity was then measured as previously described, and
data are shown in Figure 2.
With the control (CPS immobilized lipase without
additive) the esterification activity was 161 µmol‚g-1‚min
(Figure 2a) but when the lipase derivatives were pre-
pared in the presence of lecithin or albumin the esteri-
fication activities increased about 2-fold (Figure 2b and
2c). The effect of additives on the immobilized derivatives
upon the esterification reaction was exceptionally large,
as compared to the effects exhibited by the additives on
the hydrolysis. Both additives exerted a positive influence
on the esterification activities by increasing up to 2.7-
fold the activity attained with the control.
This enhancement could be attributed to distinct
effects of additives. It appears that a certain change of
806 Biotechnol. Prog., 2003, Vol. 19, No. 3

Table 3. Estimated Effects, Standard Errors, and

Student t test for Coupling Yield According to
Experimental Design
source effect standard error t values
mean 13.91 (2.31 6.01
x1 (additive) 12.50 (2.53 4.93a
x12 (additive) 16.86 (4.39 3.84a
x2 (lipase loading) 0.49 (2.53 0.19
x22 (lipase loading) -4.04 (4.39 0.91
x1x2 7.90 (3.10 2.54b
a p < 0.05. b p < 0.01

Table 4. Analysis of Variance (ANOVA) for Model

Regression Representing Yields for Lipase on CPS in the
Presence of Proteins According to Experimental Designa
sum of degree of mean
source square freedom square F value p value
x1 234.50 1 234.50 24.34 0.0159
x12 142.18 1 142.18 14.76 0.0311
x2 0.37 1 0.37 0.038 0.8571
x22 8.14 1 8.14 0.85 0.4256
x1x2 62.49 1 62.49 6.48 0.0841
error 28.90 3 9.63
total 476.60 8
a R2 ) 0.94.

the enzyme. In addition, by controlling the water system,

a shift of the thermodynamic equilibrium toward the
esterification reactions also could be obtained. This
performance was twice superior to that attained with
lipase immobilized on CPS without additive, minimizing
in this way the lipase denaturation during its fixation
onto solid supports. Therefore, when the derivatives were
used in nonaqueous media, their performance was found
to be independent of the kind of additive.
Determination of Significant Paramceters by
Analysis of Variance. The effect of additive and lipase
loading on the activity recovery of immobilized CPS-
lipase was also statistically analyzed (Table 3). The
Student t-test values show a significant effect for the
variable additive (x1) and its quadratic effect at 95%
confidence level but not for lipase loading and its interac-
tion with the additive. In the experimental range studied,
the best stabilizing effect was observed when albumin
was used as the additive. Regardless of the lipase loading,
runs using lecithin gave much lower coupling yields.
Based on the response evaluated, a mathematical
model representing the process in the experimental range
studied (eq 2) was developed. Table 4 shows the analysis
of variance for the model used to estimate the coupling
yield as a function of the experimental factor (x1) and the
interaction (x1x2). The regression is statistically signifi-
cant at 5% probability level and also presents a good
determination coefficient (R2 ) 0.94).

ŷ ) 13.91 + 6.25x1 + 8.43x21+ 3.95x1x2 (2)

where ŷ is the value predicted for the coupling yield; x1

Figure 2. Synthesis of butyl butyrate (3) from butanol (b) and
butyric acid (9) using lipase immobilized on CPS without
additive (a) and with additives (b,c). Procedure conditions as
described in Materials and Methods.
the protein conformational took place when the additive
bound to the biocatalyst or the retention of the water
shell around the catalyst gave an additional stability to
and x2 are the coded values for additive and lipase
loading, respectively.
In agreement with the mathematical model (eq 2),
maximum coupling yield (32.2%) can be achieved working
at high lipase loading (450 units‚g-1 support) and using
albumin as an additive. In these conditions, the resulting
immobilized lipase exhibits high hydrolytic (153.2 U‚mg-1)
and esterification (337.6 µmol‚g-1‚min) activities.
Conclusions
To improve the activity and stability of lipase im-
mobilized on controlled pore silica, the effect of the
Biotechnol. Prog., 2003, Vol. 19, No. 3
addition of nonenzymatic proteins during the immobili-
zation procedure was studied. The immobilized biocata-
lysts were used in the hydrolysis of olive oil and synthesis
of butyl butyrate. The data were compared with those
obtained with the immobilized biocatalyst in the absence
of additive under the same operational conditions.
In aqueous media the catalytic activity of immobilized
CPS-albumin-lipase is greater than that observed for
CPS-lecithin-lipase, probably as a result of the limitations
imposed by the physical properties of lecithin. However,
in nonaqueous media both additives enhanced the activ-
ity of the final lipase derivative. The improved lipase
activity was considered to be the sum of the benefits of
the immobilization (which prevents enzyme aggregation)
and the enzyme coating with the additive, which in-
creases the accessibility of active sites to the substrate.
References and Notes
Bosley, J. A. Supported Enzymes. Patent Application EQP
424130 A1, 1991.
Bosley, J. A.; Moore, S. R. Immobilized Lipases. Patent Applica-
tion WO 94/28118, 1994.
Box, G. E. P.; Hunter, W. G.; Hunter, J. S. Statistics for
Experimenters: An Introduction to Design, Data Analysis and
Model Building; Wiley & Sons Inc.: New York, 1978; p 653.
Bradford, M. M. A. Rapid and Sensitive Method for the
Quantification of Microgram Quantities of Protein Utilizing
the Principle of Protein-Dye Binding. Anal. Biochem. 1976,
72, 248.
807
Gonçalves, A. M.; Schucht, E.; Matthijs, G.; Aires Barros, M.
R.; Cabral, J. M. S.; Gil, M. H. Stability Studies of a
Recombinant Cutinase Immobilized to Dextran and Deriva-
tized Silica Supports. Enzyme Microb, Technol. 1999, 24, 60-
66.
Reetz, M. T.; Zonta, A.; Simpelkamp, J. Efficient Immobilization
of Lipases by Entrapment in Hydrophobic Sol-Gel Materials.
Biotechnol. Bioeng. 1996, 49, 527.
Rocha, J. M. S.; Gil, M. H.; Garcia, F. A. P. Effects of Additives
on the Activity of a Covalently Immobilised Lipase in Organic
Media. J. Biotechnol. 1998, 66, 61.
Rocha, J. M. S. Aplicações de Agentes Tensioactivos em Biotec-
nologia. Bol. Biotecnol. 2000, 5, 25.
Soares, C. M. F.; de Castro, H. F.; Moraes, F. F.; Zanin, G. M.
Characterization and Utilization of Candida rugosa Lipase
Immobilized on Controlled Pore Silica. Appl. Biochem. Bio-
technol. 1990, 77, 745.
Triantafyllou, A. O.; Wehtje, E.; Adlercreutz, P.; Mattiasson, B.
Effects of Sorbitol Addition on the Action of Free and
Immobilized Hydrolytic Enzymes in Organic Media. Biotech-
nol. Bioeng. 1995, 45, 406.
Wehtje, E.; Adlercreutz, P.; Mattiasson, B. Improved Activity
Retention of Enzymes Deposited on Solid Supports. Biotech-
nol. Bioeng. 1993, 41, 171.
Accepted for publication March 3, 2003.
BP025779Q