Enzyme and Microbial Technology 33 (2003) 845–853
Modeling the effect of free water on enzyme activity in immobilized
lipase-catalyzed reactions in organic solvents
B. Camacho Páez a , A. Robles Medina a,∗ , F. Camacho Rubio b , P. González Moreno a ,
E. Molina Grima a
a
b
Departamento de Ingenierı́a Quı́mica, Universidad de Almerı́a, 04120 Almerı́a, Spain
Departamento de Ingenierı́a Quı́mica, Universidad de Granada, 18120 Granada, Spain
Received 18 January 2003; received in revised form 2 July 2003; accepted 3 July 2003
Abstract
The influence of water content on the lipase-catalyzed acidolysis of triolein (glycerol-trioleate, TO) and caprylic acid (CA) in hexane,
using an immobilized enzyme was studied. An adequate water content (RW ) ranged from near zero to 0.1 g of water/g of dry enzyme.
Over these values there was a decrease in the rate of incorporation of CA into triglyceride. This decrease was attributed to the progressive
flooding of the carrier’s pores, in which the enzyme was immobilized. The flooding reduced the number of the enzyme molecules at the
water–hexane interface and therefore, hindered the accessibility of the hydrophobic substrates (TO and CA) to the enzyme. A simple
physical model based on a characterization of the immobilized enzyme particle by mercury porosimetry was developed. The model agreed
well with both the experimental data and the prior published data. The model may partly explain the observed inhibition when using low
molecular weight alcohols and carboxylic acids in immobilized lipase-catalyzed processes.
© 2003 Elsevier Inc. All rights reserved.
Keywords: Lipase; Water content; Organic-phase biocatalysis; Triolein; Caprylic acid; Acidolysis
1. Introduction
Since the pioneering work of Zaks and Klibanov [1,2],
organic solvents have been used extensively in enzymatic
synthesis. The enzymatic catalysis in non-aqueous media
using lipases is the usual method for synthesizing structured lipids (SLs). SLs are molecules that have been specifically designed to contain long chain functional fatty acids
and short/medium chain fatty acids in the triglyceride backbone. SL of the MLM type are triglycerides containing
medium-chain fatty acids (M) in positions 1 and 3 and an
essential long-chain fatty acid (L) in position 2 of the glycerol backbone.
The simplest and most direct route for the synthesis of
SL of the MLM type is the acidolysis between a long chain
triglyceride (LLL) and medium chain free fatty acids (M)
catalyzed with a 1,3-specific lipase [3–7]. The aim of this
synthesis is to replace the existing long chain acyl groups
(L) in positions 1 and 3, by medium chain fatty acids (M).
The presence of incomplete reaction products of partial esterification, such as di- and monoglycerides, needs to be
∗
Corresponding author. Tel.: +34-950-015065; fax: +34-950-015484.
E-mail address: [email protected] (A.R. Medina).
0141-0229/$ – see front matter © 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0141-0229(03)00219-9
minimized. However, these byproducts form inevitably, because they are essential intermediate in the acidolysis, since
this occurs through the stages of hydrolysis of triglycerides
and esterification of the formed diglycerides with the free
odd fatty acid [8]. Initiating the reaction requires a certain
amount of water; however, an excess of water contributes
to excessive hydrolysis of triglycerides, giving rise to free
fatty acids and partial glycerides (di- and monoglycerides);
a limited water content favors the transesterification reaction. The control of water content in all these systems is of
crucial importance, because all these processes are based on
the manipulation of the chemical equilibrium of thermodynamically reversible reactions in which water participates.
Besides, water is essential for the integrity of the
three-dimensional structure of the enzyme molecule and,
therefore, the lipase activity is a function of the water content. One common hypothesis is that the enzyme molecule
requires a small hydration layer that acts as the primary
component of the enzymatic microenvironment in an organic media. This layer acts as a buffer between the enzyme
surface and the bulk reaction medium [9]. The amount of
water required so that the enzymatic activity be maximal
depend on the enzyme, the support where the enzyme is
immobilized and the organic solvent used.
846
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
The organic solvents vary in their capability to dissolve
water, and therefore, there will be wide variations in the
quantity left in the environment of the biocatalyst. Zaks and
Klibanov [10] demonstrated that much less water was required to reach the maximal activity in hydrophobic solvents
than in their hydrophilic counterparts, and so, at low water
activities, the lower the solvent polarity the higher the enzyme activity. When the catalytic activity was plotted versus
the amount of water bound to the enzyme, a common pattern emerged for different solvents. Thus, one can conclude
that the enzymatic activity in organic media is primary determined not by interactions of the solvent with the enzyme
per se, but by those with the water on the enzyme [10].
The three enzymes assayed by these authors needed at least
1000 molecules of water per enzyme molecule, i.e. roughly
a monolayer of water on the enzyme surface. The influence
of the solvent on catalytic activity may be explained by its
ability to affect water–enzyme interactions and also its ability to penetrate into the essential layer of water that stabilizes the enzyme. Water is not displaced from the enzyme
by the solvent because water saturated solvents also influence catalytic activity [11]. The key parameter relating to
non-polar solvents that best fits the catalytic activity, is the
logarithm of the octanol–water partition coefficient (log P)
[11–13]. This parameter is a measure of the behavior of
separated molecules of the solvent dissolved in water and
therefore is a good measure of the ability of the solvent to
penetrate into the enzyme-bound aqueous layer.
Zaks and Klibanov [10] demonstrated in all the cases
they studied, that the enzyme activity increased with water
content, as long as water content remained below saturation
in the organic solvent phase. A similar result was obtained
by Secundo et al. [14] with the transesterification activity of
the crude and purified Candida antarctica lipase B in several
organic solvents.
The objective of this work is not to study the influence
of the water activity on the lipase activity, which has been
broadly studied by the previously mentioned authors. In this
work, we study the addition of an excess of water over a
water activity of one. Robles Medina et al. [15] observed
that the addition of small amounts of water above the water solubility in the organic solvent decreases the esterification activity of the lipase C. antarctica immobilized on a
macroporous acrilic resin (Novozyme 435). At this respect
when the enzyme is suspended (not immobilized) and there
is an excess of water, the enzyme may become soluble in it.
This facilitates its movement to the water–organic phase interface that active the enzyme. This circumstance does not
occur when the enzyme is immobilized on a solid support.
When the enzyme is immobilized, if the amount of water
is sufficiently high, the active sites of the enzyme can be
isolated and far from the water–organic solvent interface.
In this work, we analyze the influence of water content on
the immobilized lipase-catalyzed acidolysis of triolein with
caprylic acid and develop a model that allows an explanation of the influence of water content on enzymatic activity.
The model also explains the inhibition observed when using alcohols and carboxylic acids of low molecular weight
(polar substrates) in lipase-catalyzed processes.
2. Material and methods
Lipase Lipozyme® IM was donated by Novo Nordisk A/S
(Bagsvaerd, Denmark). Lipozyme® is a triacylglycerol hydrolase (E.C. 3.1.1.3) immobilized on a macroporous anion
exchange resin. The water content of the lipase, measured
by the Fisher method, was 3%. This lipase has 1,3-positional
Table 1
Characterization of the immobilized Lipozyme® IM particles by mercury
porosimetry
Applied
pressure
(kPa)
Pore radii,
RP (nm)
Accumulated
volume of
pore (cm3 /g)
Accumulated
surface area
(m2 /g)
2799
3372
7081
8474
9281
10170
12252
13363
14473
15596
16727
20623
22147
25222
26870
28669
32682
34744
36902
41784
44369
46976
52843
59125
62731
66433
70212
74156
83002
92676
103328
109231
142926
150738
159033
176822
196390
207140
218254
348542
366924
262.7
218.1
103.9
86.8
79.2
72.3
60.0
55.0
50.8
47.2
44.0
35.7
33.2
29.2
27.4
25.7
22.5
21.2
19.9
17.6
16.6
15.7
13.9
12.4
11.7
11.1
10.5
9.9
8.9
7.9
7.1
6.7
5.2
4.9
4.6
4.2
3.7
3.6
3.4
2.1
2.0
0.0000
0.0007
0.0027
0.0048
0.0061
0.0082
0.0122
0.0143
0.0170
0.0190
0.0218
0.0361
0.0497
0.0871
0.1034
0.1163
0.1422
0.1558
0.1707
0.2054
0.2231
0.2429
0.3170
0.4068
0.4578
0.4986
0.5299
0.5531
0.5912
0.6184
0.6374
0.6456
0.6687
0.6707
0.6728
0.6748
0.6762
0.6769
0.6776
0.6776
0.6782
0.00
0.01
0.04
0.08
0.11
0.17
0.29
0.36
0.47
0.55
0.67
1.40
2.20
4.60
5.76
6.73
8.88
10.13
11.58
15.29
17.36
19.81
29.95
43.60
52.05
59.20
65.01
69.55
77.66
84.14
89.20
91.56
99.33
100.15
101.01
101.95
102.64
103.00
103.41
103.41
104.08
Volume pore range 3.2–98.5%.
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
specificity. The product is granular with a particle size range
of 0.2–0.6 mm.
2.1. Characterization of the immobilized enzyme
particles
The density and porosity of the solid particles, the pore
size distribution, the specific pore volume, and the specific
surface area, were determined using a mercury porosimeter
(Quantachrome Autoscan 60), which attained a pressure of
4.2 × 105 kPa. This is capable of determining the volume
of mercury which can penetrate into the particle pore with
a pore diameters over 3.7 nm. Table 1 shows the increasing
accumulated volume of the pores and the increasing surface
area of the particle of carrier as a function of the applied pressure, corresponding to the volume pore range 3.2–98.5%.
Fig. 1 shows the pore radii distribution, d(accumulated volume)/d(log P) and d(accumulated area)/d(log P) versus pore
radii. Note that both cumulative distributions are identical
and that a bimodal pore size distribution, with peaks at
12.5 and 30 nm, was obtained, these were the more frequent
pore sizes in the Lipozyme® IM preparation. Additional
data on the immobilized enzyme characteristics are given in
Table 2.
2.2. Acidolysis reaction
log(dV/dlog(P)), cm 3/g and log(dA/dlog(P) 10-2), m2/g
A typical enzymatic reaction used triolein (TO), 100 mg
(0.113 mmol); caprylic acid (CA), 65.2 mg (0.453 mmol);
hexane, 3 ml; lipase 25.2 mg (with a water content of 3%
dry weight) and added water. The amount of water added
to the different experiments varied from 0 to 100 mg. Two
additional experiments were carried out: one with the en-
847
Table 2
Characteristics of the immobilized enzyme particles
Bulk density, ρ (g/cm3 )
Specific pore volume, Vv (cm3 /g)
Solid density, ρs (g/cm3 )
Specific surface area, Sg (m2 /g)
Average particle diameter, dp (mma )
Average pore diameter, DP = 4 Vv /Sg (nm)
Porosity, ε = Vv ρ
a
Data provided by Novo Nordisk Bioindustrial.
zyme completely soaked in water and the other experiment
with 4 Å molecular sieves added to the reaction mixture. In
the former the enzyme was submerged completely in water
during 24 h, the excess of water was separated and the enzyme added to the reaction mixture. This one was placed
in 50-ml Erlenmeyer flasks with silicone-capped stoppers.
The mixture was incubated at 50 ◦ C and agitated in an orbital shaking air-bath at 200 rpm. The reaction was stopped
at different times by separation of lipase from the reaction
mixture by filtration. The filtrate was stored at −20 ◦ C until analysis. Analytical-grade TO, CA, molecular sieves and
hexane were obtained from Sigma–Aldrich (St. Louis, MO).
Each reaction was carried out in triplicate at the same experimental conditions and the deviations of the GC analysis
always were less than 8%.
2.3. Identification of reaction products and estimation
of the molar fraction of caprylic acid incorporated
into triglycerides
Hexane was removed from the reaction mixture in a vacuum evaporator. Then the glycerides were purified by three
1
0
-1
Intruded volume
Surface area
-2
0
5
10
15
0.6677
0.6789
1.2213
104.7
0.4
26.0
0.453
20
25
30
Pore radius, nm
Fig. 1. Pore size distribution for Lipozyme® IM.
35
40
848
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
extractions with 3 ml hexane + 2 ml of 0.5N KOH (20%
ethanol solution) for each 70 mg of the initial reaction mixture. This treatment removed the free fatty acids. These glycerides were identified by thin-layer chromatography (TLC)
followed by quantitative gas chromatography (GC). The
TLC analysis has been described elsewhere [15]. Fractions
corresponding to each glyceride type were scraped from
the plates and methylated by direct transesterification with
acetyl chloride/methanol (1:20) according to the method of
Lepage and Roy [16]. These methyl esters were analyzed
by capillary GC following the procedure described in [17].
3. Results and discussion
Table 3 shows the molar fraction of caprylic acid incorporated into the triglycerides as a function of the amount of
water added and the reaction time. For this reaction and at
the same CA/TO molar ratio (m0 = 4), the CA incorporated
into the triglyceride at equilibrium was FMT = 0.563 [18].
Note that for the experiments carried out at RW = 0.031 g
of water/g of dry lipase (without adding water), the incorporation was 29% after 1 h and the equilibrium of CA incorporation was reached at 18 h. However, for RW values
of 0.44 and higher the reaction rates decreased significantly.
So, for example, in the experiment with 100 mg of added
water (RW = 4.122), the incorporation remained at 0.35
after 32 h and in the experiment with the enzyme particles
completely soaked, the incorporation after 1 h, was only 1%
(mol/mol). Therefore, the data in Table 3 imply that the initial water content does not influence the exchange equilibrium between CA and the oleic acid, however, this water
content influence the reaction rate [18].
These results agree with those reported by Akoh and
Huang [6]. These authors studied the influence of water on
the acidolysis of TO and CA in the range 0–0.62 (g of water/g of dried enzyme), with the reaction times of up to 24 h.
Under these conditions they also achieved a constant degree
of incorporation of CA for a similar RW interval as used in
this work.
To analyze the influence of water concentration on the reaction rate we calculated the relative enzyme activity of the
different experiments (Table 3). For this calculation we can
use the FMT values far from the equilibrium (e.g. at 8 h). In
addition, because of the large variation of RW (g water/g dry
lipase), we can use log RW for this study. To calculate this
relative enzyme activity, the lipase activity was assumed to
be at its maximum in the following cases: the experiment
with no added water, the experiment with added molecular
sieves and in the experiment with 1 mg of water added (i.e.
for the experiments in which RW ≤ 0.072). The FMT values
reached at 8 h were very similar in these three series of experiment (average FMT = 0.476). The relative enzyme activity
was calculated by dividing the FMT value at 8 h by 0.476.
These relative enzyme activities can also be obtained by
using the initial reaction rates calculated in a previous work
[18] by fitting the experimental results to a kinetic model
that assumes that the odd acid only incorporate in positions 1
and 3 of the glyceride backbone and that both these positions
are equivalents. These initial reaction rates can be calculated
by the equation
r0 = k[TG]0 [M]0 = k m0 [TG]0 2
(1)
where [TG]0 and [M]0 are the initial concentrations of triolein and caprylic acid, respectively, and m0 is the caprylic
acid/triolein molar ratio. Table 3 also shows the initial
Table 3
Acidolysis of TO and CA: influence of the initial water amount and the reaction time on the fraction of CA incorporated into the triglyceride, FMT , on
the initial reaction ratesa and on the relative enzymatic activityb
Added water (mg)
Rw c (g/g)
Reaction time (h)
FMT
Initial reaction rates, r0 a
Relative enzyme activityb
No, 300 mg 4 Å molecular sieves
≈0
1.0
0.031
17.1 × 10−4
1.0
1.0
0.072
15.3 × 10−4
1.0
10.0
0.440
100.0
4.122
0.476
0.519
0.547
0.290
0.417
0.482
0.562
0.469
0.512
0.544
0.336
0.525
0.118
0.349
0.010
15.9 × 10−4
No
8
18
32
1.00
6.00
8.00
18.00
8
18
32
8
18
8
32
1
Enzyme completely soaked
9.1 × 10−4
0.706
2.3 × 10−4
0.248
Experimental conditions: lipase/TO mass ratio: 0.244; CA/TO molar ratio: 4.0.
a Calculated by Eq. (1) [18].
b Relative enzyme activity: F
MT at 8 h/FMT at 8 h in the experiment carried out with RW values of 0, 0.031 and 0.072 = FMT at 8 h/0.476.
c Water/dry lipase mass ratio.
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
reaction rates, r0 , calculated by using the kinetic constants
that best fit the experimental results [18]. The calculation of
the relative enzyme activity by dividing the r0 values by the
average initial reaction rate for the experiments in which
RW ≤ 0.072 drives at similar results as when the relative
enzyme activity was calculated using data at 8 h.
Fig. 2 shows the variation of the experimental relative
lipase activity, calculated using the conversions at 8 h and
the initial reaction rates (Table 3), versus log RW (a RW
value of 0.001 was considered for experiments carried out
with molecules sieves (log RW = −3)). It is seen that the
relative lipase activity does not vary for RW values below
0.1 g of water/g of dry lipase (log RW = −1); however, for
water contents above Rw = 0.1, the reaction rate decreased
significantly.
There are several papers on the study of the influence of
water and solvent on lipase-catalyzed reactions in organic
media [1,2,10,12,19,20]. In general, the optimal content is
found at water/dry lipase mass ratio (RW ) range 0–0.1 (g
of water/g of dry lipase) and in many cases these values
concur with the water content that remains in the enzyme
in the immobilization process (RW between 0.01 and 0.05;
1–5% dry weight). A previous work on the synthesis of
triglycerides with the lipase Novozym® 435 also verified
that the enzyme water contents of 2–3% dry weight were
sufficient for an effective lipase-catalyzed reaction [21].
However, there must be some reason for the significant
decrease in the reaction rate when the water concentration
exceeds the values pointed out.
849
3.1. The model
drophilic support (anionic resin) in which the lipase is
adsorbed can completely submerge the enzyme to decrease
or annul its enzymatic activity. Even for non-hydrophilic
supports, covered by hydrophilic molecules (proteins), the
water sequestered from the organic phase will accumulate
on the surface of the lipase. Therefore, water content of
the reaction mixture must be limited. Above a certain water content, the substrates will no longer reach the surface
bound enzyme. These ideas are represented in Fig. 3, which
combines the basic principles of the model. The model
is, therefore, based on a purely physical phenomenon: the
observed decline in the reaction rate may be caused by the
physical separation between the reactants and the lipase
active sites. Next few paragraphs attempt to demonstrate
this rationale.
According to the immobilized lipase characteristics shown
in Table 2, the average pore diameter is 26 nm, although the
majority of the pore radii are around 30 and 12.5 nm in diameter (Fig. 1). Because the enzyme diameter is approximately
7.5 nm [22,23], this average pore diameter would allow an
adsorbed enzyme monolayer, covered by water molecules,
and may even provide a clearance channel that would be occupied by the organic phase, allowing substrates accessibility to the enzyme (Fig. 3a). Note that mercury porosimetry
was performed on commercially produced immobilized lipase, i.e. the volume and the pore diameter were determined
with the lipase already immobilized, and thus the whole pore
diameter of Lipozyme® IM is available to both the aqueous
phase surrounding the lipase and organic phase channel. The
amount of water per gram of lipase necessary to fill all the
particle pores is
It is generally accepted that lipases act at the water–organic
phase interface. An excess of water adsorbed on the hy-
ε(π/6)dp 3 ρW
g water
0.453 × 1
= 0.68
=
3
0.6677
g lipase
(π/6)dp ρ
Relative lipase activity
2
1
0
Model
Conversion at 8 h (Table 3)
Initial reaction rates, r0 (Table 3)
-4
-3
-2
-1
0
1
log(Rw)
Fig. 2. Influence of the water/lipase ratio (RW ) on the decrease in the enzymatic activity: comparison between the experimental data and the prediction
of the proposed model.
850
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
Fig. 3. Simplified scheme of an average pore of the biocatalyst (carrier + enzyme). The inner walls are covered by enzyme molecules (approximately
7.5 nm in diameter) which are bound to the carrier and surrounded by water. The number of water monolayers around the enzyme affects the proximity
of the enzyme active sites to the water–organic phase interface, and this in turn affects substrate access and therefore the reaction rate. If the number
of water monolayers is sufficiently reduced, the enzyme active sites then have access to the interface and substrate allowing the reaction to take place.
When the water activity is below 1, the water of the reaction mixture will be adsorbed by the enzymatic film (a). For water activities equals to 1, the
excess of water penetrates into the pore particles by capillary adsorption, flooding firstly, the narrower pores and then the larger pores. For RW = Vv ,
all the pores will be flooded. At this point, the water will start to accumulate on the external surface. This is, likely to give rise to particle aggregation
and cessation of enzymatic activity. Some of the intermediate situations of the inner state of the pore through this process are also shown (b and c).
Thus, when the RW = 0.4Vv (b), the 40% of the particle pore volume will be flooded and those enzyme molecules covered by the flooding, will be
unavailable for biocatalysis (the substrate molecules dissolved in hexane will not reach the enzyme). For RW = 0.8Vv (c), the number of inaccessible
enzyme molecules will have increased significantly.
where ε is the porosity of the catalyst particle, dp is the
average particle diameter, ρW is the density of water and ρ
is the bulk density of the particle (Table 2). It is also possible
to calculate the amount of water required to develop a layer
with a thickness greater than the lipase diameter, for example
10 nm, on the enzyme’s external surface, and thus submerge
the lipase molecules adsorbed on the external surface. This
amount would be
πdp 2 10 nm ρW
g water
= 2.25 × 10−4
3
g lipase
(π/6)dp ρ
which is a negligible amount compared to the amount of
water required to completely fill the pores of a particle
(0.68 g/g). Therefore, with this last amount of water the substrates are unable to reach the lipase molecules and the lipase active sites are unable to access the interface. Note that
RW = 0.68 is 35% higher than RW = 0.44 (i.e. when 10 mg
of water were added to 25.2 mg of lipase). In this last experiment, the relative enzyme activity was still appreciable
at 0.706 (Table 3).
Table 1 shows the pore size distribution, given by the
mercury porosimeter, which allows a rigorous assessment
of the physical model. Because of the hydrophilic nature
of the support surface, as water content increases, the
smallest pores will be the first to be flooded, followed by
the larger pores. Then for the experimental data shown in
Table 1, one can, for example, calculate the amount of water (per gram of lipase) necessary for complete flooding of
all the pores with diameter less than 17.6 nm, this amount
is:
ρW (Vv − 0.2054) = 1 × (0.6789–0.2054) = 0.4735
g water
g lipase
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
Vv is the specific pore volume given in Table 2. If at the
same time we assume that the residual lipase activity is the
surface fraction that is not flooded, for a water content of
0.4735 g water/g lipase, the residual lipase activity would
be
Residual lipase activity (model) =
15.29 m2 /g
15.29 m2 /g
=
= 0.146
Sg
104.7 m2 /g
where 104.7 m2 /g is the specific surface area of the catalyst
(Table 2) and 15.29 m2 /g is the accumulated surface area
of all the pores with diameter less than 17.6 nm (Table 1),
which are flooded with the amount of water previously calculated (0.4735 g water/g lipase). The residual lipase activities calculated with all the experimental data collected by
mercury porosimetry are also represented in Fig. 2, along
with the experimental results. It can be seen that, according
to the proposed model, lipase activity decreases with the increase of water content. Similar shapes for the three curves
are observed, although the predicted decrease in activity is
greater than that obtained with the experimental values. The
difference observed may be due to the experimental procedure: water is added to the organic solvent containing the
substrate, if in excess, tiny drops in the organic phase are
observed. When the immobilized lipase is added to start the
experiment, the tiny water drops are adsorbed, or vice versa:
the particles move toward to the tiny water drops, demonstrating the hydrophilic nature of the carrier. However, under
these conditions, it is almost impossible to obtain a homogeneous distribution of the water over the particle surface,
because, inevitably, some particles will take in more water
851
than others. This may give rise to little spots on the carrier
where the added water does not flood the pores, leading
to a lower decrease in activity than if the water had been
homogeneously distributed and the whole particle had been
flooded. The above hypothesis is verified by the experiments
in which the enzyme had been previously fully soaked because under these circumstances the water had to fill all the
pores. In these experiments, after 1 h, only 1% of incorporation of caprylic acid was observed (Table 3). This small
activity and the activity observed when 100 mg of water was
added may be due to a non-homogeneous distribution of
water on the particle surface. In addition, the above rationale
assumes a constant pore size, pore size distribution and the
particle size. However, the support matrix may swell with
the amount of moisture present and therefore pore sizes,
size distribution and particle sizes are subjected to change.
This swelling promotes a pore contraction that may intensify the physical separation between the adsorbed lipase and
the substrates.
The model proposed in this work can also explain the
results of Chulalaksananakul et al. [20] describing the influence of water concentration on the esterification of oleic
acid with ethanol in hexane. These authors determined an
optimal water/enzyme mass ratio of 0.1. Above this value
there was a decline in the lipase activity. For values in the
range 0.5–0.7, the activity became nil (Fig. 3 of their paper).
Using the results presented in their figure we calculated the
relative enzymatic activities by dividing the average initial
reaction rate by the maximum initial reaction rate at the different temperatures assayed. The values that we obtained are
represented in Fig. 4 along with data from Fig. 2 (model).
Clearly, there is a coincidence between the way in which the
1
Relative lipase activity
T = 27 ºC
T = 40 ºC
T = 55ºC
Model
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Rw , g water/(g dried lipase)
Fig. 4. Variation of the lipase relative activity with water content. Data from Chulalaksananakul et al. [20]. The continuous line represents the lipase
activity predicted by the model.
852
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
enzymatic activity decreases for both the experimental data
and the model. It must be kept in mind that, even though the
lipase used was the same brand from the same manufacturer,
10 years have passed since the Chulalaksananakul’s paper
and during this time both methodology for enzyme production and methodology for immobilization have changed
considerably. This, in our opinion, reinforces the idea that
the phenomenon, which promotes the decrease of the catalytic activity, is due to the pore structure and the pore
size distribution of the carrier. It must be remarked that the
way in which Chulalaksananakul et al. [20] introduced water into the immobilized enzyme was via one of the substrates: ethanol. This substrate facilitated the homogeneous
distribution of water. However, as previously mentioned, the
method by which we introduced water, did not facilitate its
homogeneous distribution. On the other hand, it is likely that
ethanol inhibition observed by these authors was related to
the water solubility in the ethanol, which determines a partitioning of the ethanol between the water and the organic
solvent. When the ethanol concentration increased, the aqueous phase volume surrounding the enzyme also increased.
Therefore, the fraction of enzyme molecules with their active sites near the water–organic solvent interface decreases
also.
A similar explanation can be made for the methanol inhibition observed by Ramamurti and McCurdi [24]. These
authors studied the lipase-catalyzed esterification of the
oleic acid and methanol in hexane. The observed inhibition was stronger than that previously reported for oleic
acid–ethanol system [20], which could be due to the fact
that methanol, a more polar solvent, is more hydrophilic
than ethanol, favoring the partitioning of the methanol
toward the aqueous phase of the binary water-hexane system. In addition, the hypothesis proposed in our model is
also supported by the fact that, when using high molecular
weight alcohols (i.e. alcohols that are sparingly soluble in
water), alcohol inhibition was not observed [23]. A similar
inhibition was also observed using short chain carboxylic
acids [25]. In addition, the decrease in the observed enzymatic activity may be due, in part, to the increase of the
aqueous phase volume as consequence of the increase of
the solubility of the carboxylic acid within it, although in
this case the aqueous phase is also acidified and this fact
may be more a more pertinent explanation for the inhibition
observed.
It must also be taken into account that quantitative results
will depend on the characteristics of the carrier (particle
size distribution, specific surface area and its hydrophobic
or hydrophilic feature) and the characteristics of the enzyme/carrier ratio (enzyme distribution on the support, type
of enzyme bound, etc.). This may explain the conflicting
results published by others on the influence of water on the
rate of the lipase-catalyzed reactions. In our opinion this
influence is due, mainly, to a physical phenomenon: the
proximity of the enzyme’s active site to the interface and
therefore its access to the substrates.
4. Conclusions
It seems clear that when an immobilized enzyme is dispersed in an organic phase containing lipidic substrates, a
small increase in the water concentration, measured as RW,
gives rise to the water initially forming a thin layer on the
enzyme molecules and catalytic activity either increases or
remains constant. Under these conditions, water content is
low and possibly the activity of the water is still below 1.
This first step is clearly observed in the results published
by Chulalaksananakul et al. [20], Valivety et al. [26] and
in our results (Table 3) and corresponds to a RW range
of 0–0.1. Throughout this step, water does not flood the
carrier’s pores but it covers the pore’s inner walls as a thin
layer on the adsorbed enzyme molecules (Fig. 3a). This
first step has not been considered in the calculations shown
in Fig. 2 (model). When the water content increases (and
possibly the water activity reaches a value of 1), the water
begins to flood the carrier’s pores by capillary adsorption,
starting with the narrower pores, and the catalytic activity
begins to decline (Fig. 3b). Catalytic activity continues decreasing as RW increases, until the pores are flooded and the
enzyme molecules are isolated from the substrates (Fig. 3c).
The catalytic activity becomes nil when water floods all the
pores, which corresponds to a RW value equal to the specific pore volume Vv . This fully explains the results depicted
in Fig. 4. The deviation between the model and the experimental data for the RW range of 0–0.1 may be because the
model does not consider the first step (for water activities
below 1) and deviation for the RW range 0.5–0.7 is likely
to be because the specific pore volume of the Lipozyme®
used by Chulalaksananakul et al. [20] was slightly less than
0.5–0.6 cm3 /g.
Acknowledgments
This research was supported by grants from the Dirección General de Enseñanza Superior e Investigación
Cientı́fica (Ministerio de Educación y Cultura, Spain),
Project 1FD97-0731 and Plan Andaluz de Investigación,
CVI 0173.
References
[1] Zaks A, Klibanov AM. Enzymatic catalysis in organic media at 100
◦ C. Science 1984;224:1249–51.
[2] Zaks A, Klibanov AM. Enzyme-catalyzed processes in organic
solvents. Proc Natl Acad Sci USA 1985;82:3192–6.
[3] Shimada Y, Sugihara A, Nakano H, Yokota T, Nagao T, Komemushi
S, et al. Production of structured lipids containing essential fatty
acids by immobilized Rhizopus delemar lipase. J Am Oil Chem Soc
1996;73:1415–20.
[4] Shimada Y, Sugihara A, Maruyama K, Nagao T, Nakayama
S, Nakano H, et al. Production of structured lipid containing
B.C. Páez et al. / Enzyme and Microbial Technology 33 (2003) 845–853
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
docosahexaenoic and caprylic acids using immobilized Rhizopus
delemar lipase. J Ferment Bioeng 1996;81:299–303.
Shimada Y, Suenaga M, Sugihara A, Nakai S, Tominaga Y.
Continuous production of structured lipid containing ␥-linolenic and
caprylic acids by immobilized Rhizopus delemar lipase. J Am Oil
Chem Soc 1999;76:189–93.
Akoh CC, Huang KH. Enzymatic synthesis of structured lipids:
transesterification of triolein and caprylic acid. J Food Lipids 1995;2:
219–30.
Xu X, Balchen S, Hoy CE, Adler-Nissen J. Production of specificstructured lipids by enzymatic interesterification in a Pilot continuous
enzyme bed reactor. J Am Oil Chem Soc 1988;75:1573–9.
Ainsworth S, Versteeg C, Plamer M, Millikan MB. Enzymatic
interesterification of fats. Milkfat Update Conference, 27–28 February
1996, Werribee, Victoria (Australia). Austr J Dairy Technol 1996;51:
105–7.
Dordick JS. Enzymatic catalysis in monophasic organic solvent.
Enzyme Microb Technol 1989;11:194–210.
Zaks A, Klibanov AM. Enzymatic catalysis in nonaqueous solvents.
J Biol Chem 1988;263:3194–201.
Valiveti RH, Johnston GA, Suckling CJ, Halling PJ. Solvent effect
on biocatalysis in organic systems: equilibrium position and rates
of lipase catalysed esterification. Biotechnol Bioeng 1991;38:1137–
43.
Laane C, Boeren S, Vos K, Veeger C. Rules for optimization of
biocatalysis in organic solvents. Biotechnol Bioeng 1987;30:81–7.
Janssen AEM, Hadini M, Boer NW, Walinga R, Van der Padt A,
Van Sonsbeek HM, Van’t Riet K. The effect of organic solvents on
enzymatic esterification of polyols. In: Tramper J, editor. Biocatalysis
in non-conventional media. Amsterdam: Elsevier; 1992. p. 155–
61.
Secundo F, Carrea G, Soregaroli C, Varinelli D, Morrone R. Activity
of different Candida antarctica lipase B formulations in organic
solvents. Biotechnol Bioeng 2001;73:157–63.
Robles Medina A, Esteban Cerdán L, Giménez Giménez A, Camacho
Páez B, Ibáñez González MJ, Molina Grima E. Lipase-catalyzed
esterification of glycerol and polyunsaturated fatty acids from fish
and microalgae oils. J Biotechnol 1999;70:379–91.
853
[16] Lepage G, Roy C. Improved recovery of fatty acid through direct
transesterification without prior extraction or purification. J Lipid
Res 1984;25:1391–6.
[17] Camacho Páez B, Robles Medina A, Camacho Rubio F, González
Moreno P, Molina Grima E. Production of structured triglycerides
rich in n-3 polyunsaturated fatty acids by the acidolysis of cod liver
oil and caprylic acid in a packed-bed reactor: equilibrium and kinetic.
Chem Eng Sci 2002;57:1237–49.
[18] Camacho Páez B, Robles Medina A, Camacho Rubio F,
Esteban Cerdán L, Molina Grima E. Kinetics of lipase-catalysed
interesterification of triolein and caprylic acid to produce structured
lipids. J Chem Tech Biotechnol 2003;78:461–70.
[19] Goderis HL, Ampe G, Feyten MP, Fouwé BL, Guffens WM, Van
Cauwenbergh SM, et al. Lipase-catalyzed ester exchange reactions
in organic media with controlled humidity. Biotechnol Bioeng
1987;30:258–66.
[20] Chulalaksananakul W, Condoret JS, Delorme P, Willemont RM.
Kinetic study of esterification by immobilized lipase in n-hexane.
FEBS Lett 1990;276:181–4.
[21] Esteban Cerdán LE, Robles Medina A, Giménez Giménez A, Ibáñez
González MJ, Molina Grima E. Synthesis of polyunsaturated fatty
acid-enriched triglycerides by lipase-catalyzed esterification. J Am
Oil Chem Soc 1998;75:1329–37.
[22] Ison AP, Macrae AR, Smith CG, Bosley J. Mass transfer effects in
solvent-free fat interesterification reactions: influences on catalysis
design. Biotechnol Bioeng 1994;43:122–30.
[23] Yong YP, Al Duri B. Kinetic studies on immobilised lipase
esterification of oleic acid and octanol. J Chem Tech Biotecnol
1996;65:239–48.
[24] Ramamurti S, McCurdi AR. Lipase-catalyzed esterification of oleic
acid and methanol in hexane–a kinetic study. J Am Oil Chem Soc
1994;71:927–30.
[25] Manjón A, Iborra JL, Arocas A. Short-chain flavour ester synthesis by
immobilized lipase in organic media. Biotechnol Lett 1991;13:339–
44.
[26] Valivety RH, Halling PJ, Peilow AD, Macrae AR. Lipases from
different sources vary widely in dependence of catalytic activity on
water activity. Biochim Biophys Acta 1992;1122:143–6.