Comparison of the performance of commercial immobilized lipases in the synthesis of different flavor esters

Article abstract of DOI:10.1016/j.molcatb.2014.03.021

In this work, it is compared the performance of three commercial lipase preparations (Novozym 435, Lipozyme TL-IM, and Lipozyme RM-IM) in the synthesis of flavor esters obtained by esterification of acetic, propionic, and butyric acids using ethanol, isopropyl alcohol, butanol, or pentanol. A comprehensive comparison was performed verifying activities of these three enzyme preparations versus the different couples of substrates, checking the obtained yields. In general, the longer the acid chain, the higher the reaction yields. Novozym 435 was the most efficient enzyme in most cases, and only Lipozyme RM-IM offered better results than Novozym 435 in the production of ethyl butyrate. Reactions with butyric acid showed the highest conversion rates using all biocatalysts. Using optimal substrates, the reactions catalyzed by the three enzymes were optimized using the response surface methodology, and the catalytic performance of the biocatalysts in repeated batches was assessed. After optimization, yields higher than 90% were obtained for all three enzymes, but Lipozyme TL-IM needed four-times more biocatalyst content than the other two preparations. Novozym 435 kept over 80% of its activity when reused in 9 successive batches, whereas Lipozyme RM-IM can be reused 5 times and Lipozyme TL-IM only 3 times. In general, Novozym 435 showed to be more suitable for these reactions than the other two enzyme preparations.

Full text of DOI:10.1016/j.molcatb.2014.03.021

Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Comparison of the performance of commercial immobilized lipases
in the synthesis of different flavor esters
Andréa B. Martinsa, Alexandre M. da Silvaa, Mirela F. Scheina, Cristina Garcia-Galanb,
Marco A. Záchia Ayuba, Roberto Fernandez-Lafuenteb, Rafael C. Rodriguesa,∗
a Biotechnology, Bioprocess and Biocatalysis Group, Institute of Food Science and Technology, Federal University of Rio Grande do Sul State, Av. Bento
Gonc¸ alves, 9500, PO Box 15090, ZC 91501-970 Porto Alegre, RS, Brazil
b Department of Biocatalysis, ICP-CSIC, Campus UAM-CSIC, Cantoblanco, ZC 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
In this work, it is compared the performance of three commercial lipase preparations (Novozym 435,
Lipozyme TL-IM, and Lipozyme RM-IM) in the synthesis of flavor esters obtained by esterification of acetic,
propionic, and butyric acids using ethanol, isopropyl alcohol, butanol, or pentanol. A comprehensive
comparison was performed verifying activities of these three enzyme preparations versus the different
couples of substrates, checking the obtained yields. In general, the longer the acid chain, the higher the
reaction yields. Novozym 435 was the most efficient enzyme in most cases, and only Lipozyme RM-IM
offered better results than Novozym 435 in the production of ethyl butyrate. Reactions with butyric
acid showed the highest conversion rates using all biocatalysts. Using optimal substrates, the reactions
catalyzed by the three enzymes were optimized using the response surface methodology, and the catalytic
performance of the biocatalysts in repeated batches was assessed. After optimization, yields higher than
90% were obtained for all three enzymes, but Lipozyme TL-IM needed four-times more biocatalyst content
than the other two preparations. Novozym 435 kept over 80% of its activity when reused in 9 successive
batches, whereas Lipozyme RM-IM can be reused 5 times and Lipozyme TL-IM only 3 times. In general,
Novozym 435 showed to be more suitable for these reactions than the other two enzyme preparations.
© 2014 Elsevier B.V. All rights reserved.
Received 18 December 2013
Received in revised form 10 March 2014
Accepted 26 March 2014
Available online 4 April 2014
Keywords:
Esterification
Flavor esters
Lipase
Immobilized enzyme
Enzyme reuse
1. Introduction
and alcoholysis [3–6]. Several researches reported on esterifica-
tions catalyzed by lipases, especially those using a short chain
Many commercially important fruity notes are esters formed
by short chain alcohols and carboxylic acids [1]. The extrac-
tion of these notes from natural products is expensive and, as a
consequence, their chemical synthesis has become the standard
industrial practice, i.e., the reaction using inorganic catalysts and
elevated temperatures (200–250 ◦C) [2], and the obtained esters
are classified as “artificial flavors”. In contrast, the enzymatic syn-
theses of flavor esters catalyzed by lipases allow their classification
as “natural flavors”, with immediate economical advantages [3].
Lipases are well-studied enzymes, showing hydrolytic activities,
but also catalyzing reactions of esterification, transesterification,
alcohol and a long chain fatty acid [2,7–11]. However, reactions
using short carboxylic acids have been the subject of fewer stud-
ies [12]. The enzymatic syntheses of flavor esters require very low
water contents in order to shift the thermodynamic equilibrium
toward synthesis, thus non-aqueous media are required. Solvent-
free systems [13,14], ionic liquids [15,16] or supercritical fluids [17]
have been used as media for lipase catalyzed reactions. However,
still most of the examples that can be found in the literature use
the most traditional organic solvents [13,16,18].
Although this is a very common reaction, and much effort has
been expended to find an ideal biocatalyst, a comparison of the
activity of the most used lipases has not been reported to date.
This may be interesting to understand the behavior and specifici-
ties on the synthesis of flavor esters, and, for first time, this has
been performed in this manuscript. Thus, the main objective of
this work was to evaluate the performance of three of the most
used commercially available biocatalysts (Novozym 435; Lipozyme
TL-IM; and Lipozyme RM-IM) for the synthesis of esters of dif-
ferent short chain acids (acetic, propionic, and butyric acids), and
∗
Corresponding authors at: Federal University of Rio Grande do Sul State, Biotech-
nology, Bioprocess and Biocatalysis Group, Institute of Food Science and Technology,
Av. Bento Gonc¸ alves, 9500, PO Box 15090, 91501-970 Porto Alegre, RS, Brazil.
Tel.: +55 51 3308 7793; fax: +55 51 3308 7048.
E-mail addresses: rfl@icp.csic.es (R. Fernandez-Lafuente),
rafaelcrodrigues@ufrgs.br (R.C. Rodrigues).
URL: www.ufrgs.br/bbb (R.C. Rodrigues).
http://dx.doi.org/10.1016/j.molcatb.2014.03.021
1381-1177/© 2014 Elsevier B.V. All rights reserved.

A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
19
Table 1
alcohols (ethanol, butanol, and pentanol) on organic solvents to
Experimental design and results of CCD.
comprehend the enzymes activities and specificities against these
substrates. A secondary alcohol (2-propanol) has also been tested
in order to allow us a better understanding of the enzymes speci-
ficities. The resulting esters are among the most interesting fruit
flavors, such as pineapple, banana, peach, apricot, mango, among
others.
Novozym 435 is possibly the most used biocatalyst [19]. This
immobilized preparation of lipase B from Candida antarctica (CALB)
is obtained by immobilization via interfacial activation of the
enzyme on a moderately hydrophobic resin, Lewatit VP OC 1600
[20].
The second used immobilized lipase biocatalyst was that from
Thermomyces lanuginosus (TLL). It is immobilized on a cationic sili-
cate via anion exchange (Lipozyme TL IM®) [21]. This enzyme has
been used in multiple reactions [22] and is produced by a geneti-
cally modified strain of Aspergillus oryzae [23]. Its structure is also
well known and the enzyme has a large lid, able to fully seclude the
active center from the reaction medium in the closed form [24].
The third biocatalyst used in this study is the immobilized
lipase from Rhizomucor miehei (RML), commercially available from
Novozymes as immobilized form (Lipozyme RM-IM). The support
of this immobilized enzyme is Duolite ES 562, a weak anion-
exchange resin based on phenol–formaldehyde copolymers [25].
The enzyme has been used on many processes [26,27] and its struc-
ture shows a typical lid that isolates the active center of the enzyme
from the medium when closed [28]. Therefore, these three bio-
catalysts differ not only about the enzyme source, but also in the
nature of the matrix and on the mechanism of immobilization of the
enzyme and significant differences on enzyme activity, specificity,
and stability may be expected [29].
Run
X1
X2
X3
X4
Novozym
435 (%)
Lipozyme
RM-IL (%)
Lipozyme
TL-IM (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
−1
−1
−1
−1
−1
−1
−1
−1
1
1
1
1
1
−1
−1
−1
−1
1
1
1
1
−1
−1
−1
−1
1
1
1
1
0
−1
−1
1
−1
1
25.67
25.19
54.43
59.71
11.51
19.94
53.81
47.58
32.09
50.17
78.17
75.37
43.58
45.13
69.64
67.94
14.92
59.68
45.57
52.47
25.32
73.23
49.57
38.09
48.60
36.44
47.94
44.13
30.17
23.48
60.00
62.18
29.17
33.87
51.45
42.06
48.75
53.49
72.29
67.47
29.76
39.32
78.47
74.62
35.92
1.25
44.35
40.17
16.13
81.25
62.77
53.97
43.86
46.67
55.29
48.62
43.24
48.38
77.62
81.71
13.14
12.10
44.67
30.29
0.48
24.06
17.14
24.48
28.86
4.19
45.62
37.62
42.52
18.19
68.86
0.76
−1
1
1
−1
−1
1
−1
1
−1
1
1
−1
−1
1
−1
1
−1
1
1
−1
−1
1
1
0
0
0
0
−1
1
1
1
1
−1
1
−2
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
−2
2
0
0
0
0
0
0
0
0
−2
3.71
2
87.90
55.52
46.86
60.10
55.00
58.76
48.38
0
0
0
0
0
0
−2
2
0
0
0
0
0
0
X1: temperature; X2: substrate molar ratio; X3: biocatalyst content; X4: added
water.
reaction temperature (30–60 ◦C), substrate molar ratio (1:1–5:1
alcohol:acid molar ratio), amount of added water (0–1%) and bio-
catalyst content (% considering in the calculations the mass of both
substrates in stoichiometric ratio). The range for biocatalyst content
was different for each enzyme according to their activity: 1–10% for
Novozym 435 and Lipozyme RM-IM; 5–45% for Lipozyme TL-IM.
Reactions were carried out by mixing 0.1 M of acid with different
concentrations of alcohol into 50 mL Erlenmeyer flasks (working
volume of 10 mL), followed by the addition of various amounts
of water, immobilized enzyme (dried as described above), and n-
hexane as solvent. The mixtures of acids, alcohol, and enzymes were
incubated in an orbital shaker (200 rpm) at various reaction tem-
peratures for 2 h for Novozym 435 and Lipozyme RM-IM, and 5 h for
Lipozyme TL-IM. The conditions for each reaction were determined
by the CCD, which is presented in Table 1.
We also studied the enzyme performance on some of the ester-
ification reactions using the response surface methodology, and
tested the biocatalysts activities during reuse in several batches.
2. Materials and methods
2.1. Chemicals
The enzymes used in this work were Candida antarctica lipase
B immobilized in a macroporous resin (Novozym 435); T. lanugi-
nosus lipase immobilized in a silicate support (Lipozyme TL-IM);
and R. miehei lipase immobilized in an anion-exchange resin
(Lipozyme RM-IM), all of them kindly donated by Novozymes
(Spain). Acetic, propionic, and butyric acids; ethanol, 2-propanol, 1-
butanol, 1-pentanol, and other chemicals were of analytical grade
and purchased from Sigma-Aldrich (Sigma, St. Louis, USA).
2.3. Hydrolytic activity
2.2. Esterification reactions
The hydrolytic activity of the immobilized enzymes was mea-
sured according to methodology previously developed [33]. A
volume of 5 mmol of soybean oil was added into 50 mL Erlenmeyer
flasks, followed by addition of 60 mmol of water (12:1 water:oil
molar ratio), and 10% of biocatalyst (by oil mass). The mixtures of
soybean oil, water, and lipases were stirred in an orbital shaker
(200 rpm) for 1 h at 40 ◦C. The progress of hydrolysis was moni-
tored by determination of the free fatty acid released by titration
of 0.3 g samples using 0.01 M NaOH using phenolphthalein as pH
indicator and 5 mL of ethanol as quenching agent. One unit (U) was
defined as the amount of enzyme that releases 1 ␮mol of fatty acid
per minute at the experimental conditions.
2.2.1. Screening experiments
Esterification reactions were carried out using 0.1 M of each
different acid and alcohol (1:1 alcohol:acid molar ratio) in the pres-
ence of n-hexane as solvent into 50 mL Erlenmeyer flasks (working
volume of 10 mL), under agitation in an orbital shaker (200 rpm)
for 2 h. Reaction temperature and the amount of each biocatalyst
were set according to previous studies [30–32]: 10% (mass frac-
tion of substrate) for Novozym 435 and Lipozyme RM-IM; 30% for
Lipozyme TL-IM; 40 ◦C for Novozym 435 and Lipozyme RM-IM; and
50 ◦C for Lipozyme TL-IM.
2.2.2. Reaction optimization
From the preliminary screening, one combination of acid and
alcohol was chosen for each enzyme. The optimization was per-
formed by central composite design (CCD) and response surface
methodology (RSM). The variables studied on the CCD were
2.4. Reaction analysis
The progress of esterification was monitored following the
determination of the residual acid content by titration of 0.5 mL

20
A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
samples using 0.01 N NaOH, phenolphthalein as pH indicator,
and 5 mL of ethanol as quenching agent. The amount of ester
was calculated as being equivalent to the amount of acid con-
sumed. The molar conversion was determined from the values
obtained for the blank and the test samples. The accuracy of
this method was also tested by determination of ester concentra-
tion on gas chromatograph (SHIMADZU, GC-2010 plus), equipped
with a flame ionization detector (FID) and an AT.FFAP column
(30 m × 0.32 mm × 0.25 ␮m). The carrier gas was nitrogen. The tem-
peratures of the injector and detector were both set to 250 ◦C, and
the split ratio was 1:10. The oven temperature program was: start
at 60 ◦C, 10 ◦C min−1 to 90 ◦C, 30 ◦C min−1 to 240 ◦C, and then was
held at 240 ◦C for 2.5 min.
3. Results and discussion
3.1. Screening of alcohols and acids
Fig. 1 shows the behavior of the three commercial immobilized
enzymes Novozym 435, Lipozyme RM-IM, and Lipozyme TL-IM
in the esterification of three short chain acids (acetic, propionic
and butyric), and the four selected alcohols (ethanol, 2-propanol,
1-butanol, and 1-pentanol). Novozym 435 showed the highest
activity when used with most of the substrates, except for butyric
acid (Fig. 1c), which produced good and similar conversions with
all lipases. In the production of ethyl butyrate, Lipozyme RM-IM
offered better results than Novozym 435. Lipozyme TL-IM was the
biocatalyst generally showing the lowest yields. The only secondary
alcohol assayed, 2-propanol, showed to be the poorest substrate for
esterification, independent of the acids and enzymes used, except
for the combination Lipozyme TL-IM and acetic acid, when this
enzyme presented highest activity with 2-propanol (Fig. 1a). The
selectivity of Novozym 435 versus different fatty acids (C4–C16)
and propanol or 2-propanol was studied by Arsan and Parkin [35].
The authors observed that Novozym 435 presented an initial reac-
tion rate higher with propanol than 2-propanol. Moreover, using
hexanoic acid, Novozym 435 showed lower activity with other
secondary alcohols such as 2-butanol, phenol, cyclohexanol, and
(−)menthol [35]. Although lipases can be reactive versus secondary
alcohols [36] and secondary esters [37], their reaction rates with
primary alcohols and esters are markedly faster.
As a general trend, it can be established that the longer the acid
chain, the higher the yield. This may be exemplified by Novozym
435 as biocatalyst and pentanol as alcohol, producing conver-
sions of 56, 72, and 83% for acetic, propionic, and butyric acids,
respectively. Using Lipozyme RM-IM, regardless the alcohol used,
increasing the acid chain, the yield was improved. When using
Novozym 435 and Lipozyme TL-IM as biocatalysts, it produced
lower yields using propionic acid, except when using pentanol.
Although the observed yield using propionic acid was lower than
those using the other two acids for Novozym 435, with butanol and
pentanol it was reached yields near to 70% in 2 h. Kuperkar et al.
[38] under their optimal conditions obtained 90% in the synthesis
of isobutyl propionate catalyzed by Novozym 435 in solvent-free
system after 10 h.
In general, concerning the effect of the alcohol used in the
reaction, pentanol, and especially butanol, showed to be the best
substrates for most enzymes and acids combinations. Ethanol also
produced good conversions, especially in combination with butyric
acid. Moreover, using acetic acid as substrate and Novozym 435 as
biocatalyst, ethanol showed to be a good acyl acceptor.
Results shown in Fig. 1 demonstrate the different behaviors
of these three commercial biocatalysts during esterification, with
significant interactions between the used acyl donors and the
nucleophiles, strongly affecting the enzymes activities.
It is worthwhile noting that these reactions were carried out
using fixed conditions (see Section 2.2.1), with further influences in
the reaction expected caused by the immobilization protocol used
[31,39], and therefore these results must be taken as an indicative
trend.
2.5. SDS-PAGE
SDS-polyacrylamide gel electrophoresis was performed accord-
ing to Laemmli [34] using a SE 250-Mighty small II electrophoretic
unit (Hoefer Co.); 15% (concentration fraction) running gel in a sep-
aration zone of 9 × 6 cm, and a stacking gel of 5% (concentration
fraction) polyacrylamide. Samples of 100 mg of the commercial
preparations of the immobilized enzyme were re-suspended in
1 mL of rupture SDS-buffer 2% (mass fraction), 10% (volume frac-
tion) mercaptoethanol, and 1 M NaCl, boiled for 5 min. A volume
of 20 ␮L aliquots of the supernatant was loaded onto gels. After
running, gels were stained with Coomassie brilliant blue. Low
molecular weight markers (GE Healthcare) were used as reference
(14,000–97,000 Da).
2.6. Statistical analysis
The conversions obtained in the esterification reactions were
calculated. CCD and analysis of results were carried out using Sta-
tistica 7.0 (Statsoft, USA). The statistical analysis of the model was
performed as analysis of variance (ANOVA). The significance of
the regression coefficients and the associated probabilities, p(t),
were determined by Student’s t-test; the second order model
equation significance was determined by Fisher’s F-test. The vari-
ance explained by model was given by the multiple determination
coefficients, R2. For each variable, the quadratic models were
represented as contour plots (2D). The second-order polynomial
equation for the variables was as follows:
ꢀ
ꢀ
ꢀ
Y = ˇ0
+
ˇiXi +
ˇijXiXj +
ˇiiXi2
(1)
where Y is the response variable, ˇ0 the constant, ˇi, ˇii, ˇij were
the coefficients for the linear, quadratic, and for the interaction
effects, respectively, and Xi and Xj the coded level of variables xi and
xj. The above quadratic equation was used to plot surfaces for all
variables.
2.7. Enzyme reuse
3.2. SDS-PAGE of protein loading of the different biocatalysts
After the esterification reaction, the immobilized enzyme was
separated from the reaction medium by vacuum filtration using
a sintered glass funnel. In some cases, in order to remove any
adsorbed substance from the support, the recovered biocatalysts
were washed with 10 volumes of n-hexane, dried for 24 h at 25 ◦C
and reused in a new fresh reaction [32]. Hexane is highly volatile,
providing a dry biocatalyst after 24 h. In other cases, the reuse was
performed after vacuum filtration without further treatments.
Fig. 2 shows the result of SDS-PAGE, which measured (semi-
quantitatively) the amount of lipase immobilized in each support.
This experiment helps to better understand the relation between
the biocatalysts activities and enzyme load. The gel shows that
Lipozyme TL-IM has the highest protein load, followed by Novozym
435, and Lipozyme RM-IM, with the lowest load. Because Lipozyme
TL-IM produced the lowest esterification activities (Fig. 1) among

A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
21
a
b
c
Fig. 1. Esterification of short chain acids catalyzed by the three lipases. (a) Acetic acid; (b) propionic acid; (c) butyric acid. Bars: black, ethanol; dark gray, 2-propanol; light
gray, butanol; white, pentanol.
all three preparations, its high protein load suggests a significant
lower specific activity for these reactions.
reported to alter the properties of the same lipase in one spe-
cific reaction [42], as it has been shown for esterification reactions
[39,43]. The nature of the support can affect the enzyme confor-
mation, as well as the partition of substrates and products from
the enzyme environment, which might difficult the access of the
substrate to the enzyme active site. The immobilization protocol
is another important aspect that has been shown to alter enzyme
properties even when immobilized on the same support [44,45].
Finally, the reaction pH and water content will influence enzyme
properties and reaction rates.
The nature of the immobilization supports of the three biocat-
alysts used in this work was different. Lewatit, which was used to
immobilize CALB to produce Novozym 435, has a fairly hydropho-
bic surface [46,47], while a very hydrophilic silicate was used to
immobilize TLL to produce Lipozyme TL-IM [21] and Duolite ES
562, a weak anion-exchange resin based on phenol–formaldehyde
copolymers was used to immobilized RML [25]. Concerning the
Nevertheless, when we compared the hydrolytic activity, the
Lipozyme TL-IM showed the highest activity (16.6 U g−1 of biocat-
alyst), followed by Lipozyme RM-IM (13.6 U g−1 of biocatalyst) and
Novozym 435 (3.6 U g−1 of biocatalyst). It is important to note that
the best biocatalyst in the esterification reaction is the worst in
hydrolysis, confirming the complexity of lipase catalyzed reactions.
Novozym 435 is 4-times more active than Lipozyme TL-IM in syn-
thesis, and Lipozyme TL-IM was 5-times more active than Novozym
435 in hydrolysis. Lipozyme RM-IM showed good activity in both
reactions.
Several aspects might influence the activities of the biocatalysts.
Enzyme sources (microorganism) should naturally be expected to
show structural differences, having strong influence on the biocat-
alysts properties and activities, even in similar reaction systems
[40,41]. The nature of the immobilization support has also been

22
A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
Table 2
Linear effects of the variables in the reactions catalyzed by the three enzymes used
in this work.
Variable
Novozym
435
Lipozyme
RM-IM
Lipozyme
TL-IM
X1
X2
X3
X4
Temperature
21.14*
−2.32
29.09*
−0.06
5.07
−3.95
29.23*
−1.76
−18.03*
−19.73*
29.42*
−2.11
Substrate molar ratio
Biocatalyst content
Added water
*
Statistically significant at 95% of confidence level.
presenting F-values of 13.06, 2.63, and 6.77 for reactions catalyzed
by Novozym 435, Lipozyme RM-IM, and Lipozyme TL-IM, respec-
tively, all being statistically significant (p-values: <0.0001, 0.0450,
and 0.0007, respectively). The determination coefficients (R2) and
the correlation coefficients (R) were higher than 0.8 for all models,
meaning that more than 80% of the variation for the syntheses are
attributed to the independent variables, and could be explained by
their respective models, while the high correlation coefficients sug-
gest a satisfactory representation of the process models and good
correlations between experimental results and theoretical values
predicted by the models equations. Therefore, the second-order
polynomial models are given by
Y1 = 44.27 + 10.57X1 − 1.30X1X1 − 1.16X2 + 1.62X2X2
+ 14.54X3 + 1.69X3X3 − 0.03X4 + 0.32X4X4
+ 1.41X1X2 − 0.81X1X3 − 0.98X1X4 + 0.50X2X3
Fig. 2. SDS-PAGE of the immobilized lipases. Lane 1: molecular weight markers
(kDa); lane 2: Novozym 435; lane 3: Lipozyme RM-IM; lane 4: Lipozyme TL-IM.
− 1.12X2X4 − 2.06X3X4
(2)
(3)
(4)
immobilization protocols, Novozym 435 is prepared via interfacial
activation of the enzyme on the hydrophobic surface of the sup-
port [46,47], whereas the other two enzymes are immobilized via
anionic exchange [21,25]. All these differences of the biocatalysts
make more complex the understanding of the interactions of acids
and alcohols on the enzymes activities.
Based on the results presented in Figs. 1 and 2, and consider-
ing all discussed complexities that influence enzyme reactions, it
was chosen one combination of acid and alcohol for each immobi-
lized enzyme in order to perform the optimization of the reaction
parameters using RSM.
Y2 = 48.60 + 2.53X1 − 6.28X1X1 − 1.97X2 − 0.27X2X2
+ 14.61X3 + 1.33X3X3 − 0.88X4 + 3.75X4X4
− 0.03X1X2 + 1.40X1X3 + 0.52X1X4 + 0.92X2X3
+ 0.350X2X4 − 1.76X3X4
Y3 = 55.55 − 9.01X1 − 7.55X1X1 − 9.86X2 − 6.37X2X2
+ 14.71X3 − 3.62X3X3 − 1.05X4 − 2.28X4X4
+ 12.55X1X2 − 3.13X1X3 + 0.94X1X4 + 0.27X2X3
− 5.51X2X4 − 0.86X3X4
3.3. Experimental design for flavor esters syntheses
The esters chosen for each enzyme were: pentyl acetate (banana
and apple flavors) for Novozym 435; pentyl butyrate (pear and apri-
cot flavors) for Lipozyme RM-IM; and ethyl butyrate (pineapple and
mango flavors) for Lipozyme TL-IM. These combinations were cho-
sen because they present some room to improvement and due to
their importance as flavor for food industry. Reaction temperature,
substrate molar ratio, biocatalyst content and added water were
evaluated and the results for flavors esters synthesis are presented
in Table 1. The reaction times considered for analyses were 2 h for
Novozym 435 and Lipozyme RM-IM, and 5 h for Lipozyme TL-IM,
which were chosen in order to obtain relative high yields, but not as
high as to make it difficult to detect improvements in activity when
changing reaction conditions. The range of the variables studied in
the CCD were the same, the only exception being the biocatalyst
content for Lipozyme TL-IM that, due to its low activity, was var-
ied from 5 to 45%, while for the other 2 biocatalysts it was from
1 to 10%. The highest yields obtained were around 80% for each
enzyme (Table 1). Conversions varied widely, with yields lower
than 10% to higher than 80%, clearly showing the importance of
reaction optimization.
where Y1, Y2, and Y3 are the percentage yields for pentyl
acetate catalyzed by Novozym 435, pentyl butyrate catalyzed by
Lipozyme RM-IM, and ethyl butyrate catalyzed by Lipozyme TL-IM,
respectively, and X1, X2, X3, and X4 are the coded values of tem-
perature, substrate molar ratio, enzyme content, and added water,
respectively.
3.4. Effects of process variables
The relation between the process variables can be observed
in the contour plots of Fig. 3. For each enzyme, it was plotted
the curves for two variables according to their linear effects as
presented in Table 2: biocatalyst content and temperature for
Novozym 435 and Lipozyme RM-IM (Fig. 3a and b), and substrate
molar ratio (alcohol:acid) and temperature for Lipozyme TL-IM
(Fig. 3c). Positive values mean that the response is improved when
increasing the variable from level (−1) to level (1), and inversely
for negative values. The response in this case is the reaction yield.
It can be seen in Table 2, that the content of the biocatalyst was
the variable presenting the highest reaction effect and that, inde-
pendently of the enzyme and acid/alcohol combination, the effect
In order to check the models fitness, Fisher’s statistical test for
analysis of variance (ANOVA) was performed for the 3 reactions,

A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
23
Fig. 3. Contour plots for esterification reactions catalyzed by the three biocatalysts used in this work. (a) Pentyl acetate, Novozym 435; (b) pentyl butyrate, Lipozyme RM-IM;
(c) ethyl butyrate, Lipozyme TL-IM.
of biocatalyst content was very similar. This high and positive
effect was expected because the amount of enzyme should directly
relate to the reaction rate. More complex and interesting results
were observed for the effects of temperature and substrate molar
ratio on the yields of reactions. It is expected that temperature,
should positively affect reaction rate up to a point where enzyme
inactivates, which was observed for Novozym 435 and Lipozyme
RM-IM. Observing the contour plots of Fig. 3a, increasing the reac-
tion temperature and the biocatalyst content, the reaction yield
was improved, which was expected because both biocatalyst con-
tent and temperature should increase reaction rate. Concerning
Lipozyme RM-IM (Fig. 3b), the temperature had a positive effect
around the central point, meaning that low and high temperatures
will negatively affect the enzyme activity.
analyzing data presented in Table 1 and Fig. 3c. At high substrate
molar ratios, low yields were obtained, independent of the tem-
perature, but fixing the substrate molar ratio at the lower level and
increasing the temperature, the yield increased.
3.5. Optimal reaction conditions and model validation
The optimal conditions for the syntheses of the different esters
using the three biocatalysts were determined by the response desir-
ability profile, calculated using the Statistica 7.0 software, where
the optimal values for each variable were obtained for the desired
response, which in this work was the maximal yields after 2 h or
5 h of reaction, depending on the enzyme. The optimal conditions
for each enzyme are shown in the Table 3.
Although biocatalyst content presented the highest effect for
Lipozyme TL-IM, temperature and substrate molar ratio showed
important effects on the reaction. Somewhat surprisingly, tem-
perature showed a significant and negative effect on the reaction
catalyzed by Lipozyme TL-IM, since the enzyme in this biocatalyst
is known to be thermophilic [22]. One possible explanation could
be the combination effect of substrate molar ratio and temperature.
This interaction effect was highly significant when using Lipozyme
TL-IM (X1X2 = 25.11, p = 0.0024), and can be best understood
In order to validate the prediction models, experiments were
carried out at the optimal conditions and the time-courses are pre-
sented in Fig. 4. Novozym 435 was slightly faster in the synthesis of
pentyl acetate than the other lipases in the synthesis of their respec-
tive esters. Moreover, it is important to remark that Lipozyme TL-IM
needs four-times more biocatalyst in order to show similar reaction
rates. The synthesis of ethyl butyrate was also studied by Guillén
et al. [48], and the authors needed two-times more enzyme for
the reaction catalyzed by Rhizopus oryzae, immobilized on Accurel

24
A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
Table 3
Optimal conditions for the esterification reactions catalyzed by the three enzymes
used in this work.
Enzyme
Novozym
435
Lipozyme
RM-IM
Lipozyme
TL-IM
Temperature (◦C)
Substrate molar ratioa
Biocatalyst contentb (%)
Waterb (%)
50
47.5
2.4:1
10
32.5
1.7:1
40
3:1
7.75
0.25
0.12
0.38
a
(alcohol:acid).
(% by mass of substrate).
b
Fig. 5. Stability of the biocatalysts over repeated batches under optimal conditions.
(ꢀ) Novozym 435, (ꢁ) Lipozyme RM-IM and (ꢀ) Lipozyme TL-IM.
batches, and the results are presented in Fig. 5. Novozym 435
showed the best performance, being possible to use the biocat-
alyst for 10 successive batches retaining more than 80% of its
initial activity. Similar stability was obtained by Kuperkar et al.
[38], who observed that there were marginal changes in percent-
age conversion after 7 cycles of reuse using Novozym 435 in the
synthesis of isobutyl propionate in solvent free system. Lipozyme
TL-IM, however, was fully inactivated after five batches, while
Lipozyme RM-IM was after seven batches. These differences could
be explained by diverse enzymes properties, such as CALB being
more stable than the other two lipases under the conditions of reac-
tions. Another hypothesis is the difference in the immobilization
protocol and nature of supports, as it has been pointed out before.
Finally, the different substrates in each reaction might also have
affected the enzyme stability, as they can have a lower or higher
partition effect from the enzyme environment.
Fig. 4. Time course of esterification reactions under optimal conditions catalyzed
by (ꢀ) Novozym 435, (ꢁ) Lipozyme RM-IM and (ꢀ) Lipozyme TL-IM.
EP100, compared to the results obtained using Lipozyme TL-IM in
the present work. Moreover, the authors obtained 90% of conver-
sion in 8 h [48]. Nevertheless, in our study for all 3 biocatalysts,
yields over 90% were obtained in less than 6 h of reaction, and the
results were better than others found in the literature. Milasˇinovic´
et al. under their optimized conditions for the synthesis of isobu-
tyrate catalyzed by Candida rugosa lipase obtained over 90% of ester
yield after 48 h [6]. Our productivity was 10-fold higher (around
50 mmol L−1 h−1 against 5 mmol L−1 h−1 [6]). Using the lipase from
Staphylococcus simulans, Ghamgui et al. achieved 64% of conversion
after 8 h for the synthesis of isoamyl acetate in a solvent-free sys-
tem [13]. In another study, de Paula et al. [40] tested seven lipases
in the synthesis of butyl butyrate. The authors found that C. rugosa
lipase was the most feasible to ester synthesis obtaining around
75% of conversion after 24 h.
4. Conclusion
A comprehensive comparison of three of the most commonly
used commercial immobilized biocatalysts was performed as bio-
catalyst of esterification reactions, analyzing the effects of alcohol
and acid natures in the reaction rate. The results suggested
that Novozym 435 is a better biocatalyst than Lipozyme RM-IM
and Lipozyme TL-IM. However, their performances were highly
dependent on the alcohols and acids used, and Lipozyme RM-IM
was found to be more active for the production of ethyl butyrate.
Moreover, although Lipozyme TL-IM presents a higher protein load,
it needs four-fold more mass of biocatalyst to give the same reaction
rates than the other two biocatalysts. Nevertheless, after optimiza-
tion, the 3 biocatalysts showed good activity for specific flavor
esters, and in less than 6 h of reaction high yields could be obtained.
Results have shown that the enzymatic reactions are depend-
ent upon several factors, varying from the more predictable, such
as temperature, to the more complex ones such as substrates com-
bination, and types of biocatalysts. It may be expected that using
other immobilization protocols, the performance of all the enzymes
may be tuned and the results presented in this paper may be
modified.
3.6. Enzyme reuse
One of the most important properties of immobilized enzymes
is the possibility of recovering and reusing them, which are impor-
tant economic and environmental aspects and could define their
future industrial applications. As previously described [32], wash-
ing the biocatalyst with n-hexane between reaction batches helps
to remove any excess of substrates or products from inside the
porous of the biocatalysts, which is the main cause for decreasing
enzyme activity. Therefore, we decided to test the reuse of the three
immobilized lipases, performing a wash with n-hexane between

A.B. Martins et al. / Journal of Molecular Catalysis B: Enzymatic 105 (2014) 18–25
25
Acknowledgments
[21] L. Peng, X. Xu, H. Mu, C.E. Høy, J. Adler-Nissen, Enzyme Microb. Technol. 31
(2002) 523–532.
[22] R. Fernandez-Lafuente, J. Mol. Catal. B: Enzym. 62 (2010) 197–212.
[23] W. Prathumpai, S.J. Flitter, M. McIntyre, J. Nielsen, Appl. Microbiol. Biotechnol.
65 (2004) 714–719.
[24] U. Derewenda, L. Swenson, R. Green, Y. Wei, S. Yamaguchi, R. Joerger, M.J. Haas,
Z.S. Derewenda, Protein Eng. 7 (1994) 551–557.
[25] B. Huge-Jensen, D.R. Galluzzo, R.G. Jensen, Lipids 22 (1987) 559–565.
[26] R.C. Rodrigues, R. Fernandez-Lafuente, J. Mol. Catal. B: Enzym. 64 (2010) 1–22.
[27] R.C. Rodrigues, R. Fernandez-Lafuente, J. Mol. Catal. B: Enzym. 66 (2010)
15–32.
[28] Z.S. Derewenda, U. Derewenda, G.G. Dodson, J. Mol. Biol. 227 (1992) 818–839.
[29] R.C. Rodrigues, C. Ortiz, A. Berenguer-Murcia, R. Torres, R. Fernandez-Lafuente,
Chem. Soc. Rev. 42 (2013) 6290–6307.
This work was supported by grants and scholarships (first and
second authors) from FAPERGS (Fundac¸ ão de Amparo a Pesquisa
do Rio Grande do Sul process 11/0485-0 ARD/2011), CNPq (Brazil-
ian Bureau of Science and Technology process 478169/2011-6
Universal 2011), and CTQ2009-07568 (Ministerio de Ciencia y Inno-
vación, Spain). The authors wish to thank Mr. Ramiro Martínez
(Novozymes, Spain) for kindly supplying the enzymes used in this
research.
[30] A.S.G. Lorenzoni, N.G. Graebin, A.B. Martins, R. Fernandez-Lafuente, M.A. Záchia
Ayub, R.C. Rodrigues, Flavour Fragrance J. 27 (2012) 196–200.
[31] A.B. Martins, J.L.R. Friedrich, J.C. Cavalheiro, C. Garcia-Galan, O. Barbosa, M.A.Z.
Ayub, R. Fernandez-Lafuente, R.C. Rodrigues, Bioresour. Technol. 134 (2013)
417–422.
[32] A.B. Martins, N.G. Graebin, A.S.G. Lorenzoni, R. Fernandez-Lafuente, M.A.Z.
Ayub, R.C. Rodrigues, Process Biochem. 46 (2011) 2311–2316.
[33] J.S. Alves, N.S. Vieira, A.S. Cunha, A.M. Silva, M.A. Záchia Ayub, R. Fernandez-
Lafuente, R.C. Rodrigues, RSC Adv. 4 (2014) 6863–6868.
References
[1] B. Gillies, H. Yamazaki, D.W. Armstrong, Biotechnol. Lett. 9 (1987) 709–714.
[2] A.R.M. Yahya, W.A. Anderson, M. Moo-Young, Enzyme Microb. Technol. 23
(1998) 438–450.
[3] F. Hasan, A.A. Shah, A. Hameed, Enzyme Microb. Technol. 39 (2006) 235–251.
[4] D.M. Chesterfield, P.L. Rogers, E.O. Al-Zaini, A.A. Adesina, Chem. Eng. J. 207–208
(2012) 701–710.
[5] D. Goswami, J.K. Basu, S. De, Chem. Eng. J. 181–182 (2012) 542–548.
[6] N. Milasˇinovic´, Z. Knezˇevic´-Jugovic´, Z. Jakovljevic´, J. Filipovic´, M. Kalagasidis
Krusˇic´, Chem. Eng. J. 181–182 (2012) 614–623.
[7] R. Lortie, Biotechnol. Adv. 15 (1997) 1–15.
[8] A. Rajendran, A. Palanisamy, V. Thangavelu, Braz. Arch. Biol. Technol. 52 (2009)
207–219.
[9] I. Aissa, M. Sellami, A. Kamoun, Y. Gargouri, N. Miled, Curr. Chem. Biol. 6 (2012)
77–85.
[10] H. Horchani, A. Bouaziz, Y. Gargouri, A. Sayari, J. Mol. Catal. B: Enzym. 75 (2012)
35–42.
[11] H. Horchani, I. Aissa, S. Ouertani, Z. Zarai, Y. Gargouri, A. Sayari, J. Mol. Catal. B:
Enzym. 76 (2012) 125–132.
[12] E. Alvarez-Macarie, J. Baratti, J. Mol. Catal. B: Enzym. 10 (2000) 377–383.
[13] H. Ghamgui, M. Karra-Chaâbouni, S. Bezzine, N. Miled, Y. Gargouri, Enzyme
Microb. Technol. 38 (2006) 788–794.
[14] A. Kumari, P. Mahapatra, V.K. Garlapati, R. Banerjee, S. Dasgupta, Food Technol.
Biotechnol. 47 (2009) 13–18.
[15] A. Rios, F.J. Hernandez-Fernandez, F. Tomas-Alonso, D. Gomez, G. Villora,
Flavour Fragrance J. 23 (2008) 319–322.
[16] T. De Diego, P. Lozano, M.A. Abad, K. Steffensky, M. Vaultier, J.L. Iborra, J. Bio-
technol. 140 (2009) 234–241.
[34] U.K. Laemmli, Nature 227 (1970) 680–685.
[35] J. Arsan, K.L. Parkin, J. Agric. Food Chem. 48 (2000) 3738–3743.
[36] D. Rotticci, F. Hæffner, C. Orrenius, T. Norin, K. Hult, J. Mol. Catal. B: Enzym. 5
(1998) 267–272.
[37] H. Horchani, N.B. Salem, A. Chaari, A. Sayari, Y. Gargouri, R. Verger, J. Colloid
Interface Sci. 347 (2010) 301–308.
[38] V.V. Kuperkar, V.G. Lade, A. Prakash, V.K. Rathod, J. Mol. Catal. B: Enzym. 99
(2014) 143–149.
[39] N.G. Graebin, A.B. Martins, A.S.G. Lorenzoni, C. Garcia-Galan, R. Fernandez-
Lafuente, M.A.Z. Ayub, R.C. Rodrigues, Biotechnol. Prog. 28 (2012) 406–412.
[40] A.V. de Paula, G.F.M. Nunes, J.D.L. Silva, H.F. De Castro, J.C. Dos Santos, Appl.
Biochem. Biotechnol. 160 (2010) 1146–1156.
ˇ
[41] R.C. Rodrigues, G. Volpato, K. Wada, M.A.Z. Ayub, J. Am. Oil Chem. Soc. 85 (2008)
925–930.
[42] G. Fernandez-Lorente, Z. Cabrera, C. Godoy, R. Fernandez-Lafuente, J.M. Palomo,
J.M. Guisan, Process Biochem. 43 (2008) 1061–1067.
[43] J.L.R. Friedrich, F.P. Pen˜a, C. Garcia-Galan, R. Fernandez-Lafuente, M.A.Z. Ayub,
R.C. Rodrigues, J. Chem. Technol. Biotechnol. 88 (2013) 1089–1095.
[44] O. Barbosa, C. Ortiz, R. Torres, R. Fernandez-Lafuente, J. Mol. Catal. B: Enzym.
71 (2011) 124–132.
[45] G. Fernández-Lorente, L. Betancor, A.V. Carrascosa, J.M. Palomo, J.M. Guisan, J.
Am. Oil Chem. Soc. 89 (2012) 97–102.
[46] B. Chen, J. Hu, E.M. Miller, W. Xie, M. Cai, R.A. Gross, Biomacromolecules 9 (2008)
463–471.
[47] B. Chen, M.E. Miller, R.A. Gross, Langmuir 23 (2007) 6467–6474.
[48] M. Guillén, M.D. Benaiges, F. Valero, Biochem. Eng. J. 65 (2012) 1–9.
[17] M.N. Varma, G. Madras, J. Chem. Technol. Biotechnol. 83 (2008) 1135–1144.
[18] L. Gubicza, A. Kabiri-Badr, E. Keoves, K. Belafi-Bako, J. Biotechnol. 84 (2000)
193–196.
[19] E.M. Anderson, K.M. Larsson, O. Kirk, Biocatal. Biotransform. 16 (1998) 181–204.
[20] J. Uppenberg, S. Patkar, T. Bergfors, T.A. Jones, J. Mol. Biol. 235 (1994) 790–792.