Towards a green enantiomeric esterification of R/S-ketoprofen: A theoretical and experimental investigation

Article abstract of DOI:10.1016/j.molcatb.2015.05.003

Methanol, ethanol, 1- and 2-propanol were used as reactants and solvents in the esterification of R/S-ketoprofen catalyzed with Novozym 435. The interaction of the alcohols with Novozym 435 was studied at a molecular level through various spectroscopic techniques and molecular modeling. The results evidenced the dissolution of the polymeric support, loss of active protein, strong adsorption of the alcohols, modification of the secondary structure of the protein and smoothing of the inner structure of the biocatalyst's beads upon extended contact with the alcohols. Nevertheless, none of those drawbacks influences the specific activity and enantiomeric excess toward S(+)-enantiomer that remain unaltered upon extended contact with ethanol, 1- and 2-propanol as acyl acceptors. However, theoretical calculations demonstrated that methanol introduces steric and electronic hindrance within the step of the coordination of the R/S-ketoprofen with the catalytic triad.

Full text of DOI:10.1016/j.molcatb.2015.05.003

Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Towards a green enantiomeric esterification of R/S-ketoprofen: A
theoretical and experimental investigation
María Victoria Toledoa, Carla Joséa, Sebastián E. Collinsb, María Luján Ferreirac,
Laura E. Brianda,∗
a Centro de Investigación y Desarrollo en Ciencias Aplicadas-Dr. Jorge J. Ronco. Universidad Nacional de La Plata, CONICET, CCT La Plata. Calle 47N 257,
B1900AJK La Plata, Buenos Aires, Argentina
b Instituto de Desarrollo Tecnológico para la Industria Química INTEC, Universidad Nacional del Litoral, CONICET, Güemes 3450, 3000 Santa Fe, Santa Fe,
Argentina
c Planta Piloto de Ingeniería Química – PLAPIQUI, CONICET, Universidad Nacional del Sur, Camino La Carrindanga km 7, 8000 Bahía Blanca, Argentina
a r t i c l e i n f o
a b s t r a c t
Article history:
Methanol, ethanol, 1- and 2-propanol were used as reactants and solvents in the esterification of R/S-
ketoprofen catalyzed with Novozym® 435. The interaction of the alcohols with Novozym® 435 was
studied at a molecular level through various spectroscopic techniques and molecular modeling. The
results evidenced the dissolution of the polymeric support, loss of active protein, strong adsorption of
the alcohols, modification of the secondary structure of the protein and smoothing of the inner struc-
ture of the biocatalyst’s beads upon extended contact with the alcohols. Nevertheless, none of those
drawbacks influences the specific activity and enantiomeric excess toward S(+)-enantiomer that remain
unaltered upon extended contact with ethanol, 1- and 2-propanol as acyl acceptors. However, theoretical
calculations demonstrated that methanol introduces steric and electronic hindrance within the step of
the coordination of the R/S-ketoprofen with the catalytic triad.
Received 12 March 2015
Received in revised form 2 May 2015
Accepted 7 May 2015
Available online 15 May 2015
Keywords:
Lipase
Novozym® 435
Alcohols
R/S-ketoprofen
Kinetic resolution
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
could produce side effects or being toxic. In this context, the pro-
duction of stereochemically pure drugs has many advantages, such
Lipases (triacylglycerol hydrolase, EC 3.1.1.3) are a family of
enzymes broadly applied in organic chemistry. They have wide
specificity and also the ability to react with different substrates [1].
They are able to catalyze not only hydrolytic but also synthetic reac-
tions of esterification, transesterification, interesterification and
alcoholysis among others. These enzymes are chosen for applica-
tion in several industries such as, food, detergent, pharmaceutical,
biodiesel, cosmetic and paper due to their high activity, substrate-,
regio- and stereospecificity along with their milder reaction condi-
tions [2–6].
Pharmaceutical industries apply the enantioselectivity of lipases
for preparing different pharmaceuticals and fine chemicals contain-
ing a chiral center [1,7]. Several drugs are found in their racemic
form. However, in many cases one enantiomer is responsible for the
activity of interest while the other could be inactive, be an antag-
onist of the active enantiomer or have a different activity which
as the reduction of the administered dose, improved therapeutic
window, prevents side effects and a more precise estimation of
dose-response relationship [8].
2-Arylpropionic acids (profens) constitute an important group
of non-steroidal anti-inflammatory substances (NSAIDs). They have
their pharmacological activity mainly on the (S)-enantiomer. These
are widely used for relieving pain and for the treatment of inflam-
matory diseases, such as rheumatoid arthritis and osteoarthritis [9].
In most cases, the (R)-enantiomer often has a very poor activity,
unwanted physiological side effects and toxicity [10]. Among the
side effects, gastrointestinal disorders are frequently present, being
ulceration and mucosal hemorrhage at different locations from the
gastrointestinal tract the most frequent adverse reactions [11].
Lipases are one of the most suitable enzymes for kinetic res-
olution of profens due to their wide substrate specificity, their
ability to recognize chirality and do not require labile cofactors [12].
Moreover, from an environmental point of view, processes that use
enzymes are greener, less hazardous and friendlier because they do
not contaminate as in the case of using chemical catalysts [13,14].
In this context, the investigations regarding the kinetic resolu-
tion of R/S-ketoprofen with lipases are summarized chronologically
∗
Corresponding author. Tel.: +54 221 4 211353/210711.
E-mail address: briand@quimica.unlp.edu.ar (L.E. Briand).
http://dx.doi.org/10.1016/j.molcatb.2015.05.003
1381-1177/© 2015 Elsevier B.V. All rights reserved.

M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
53
within the Supplementary Information section in Table 1S (see refs.
[15-29]. This table shows the various acyl acceptors, reaction condi-
tions (molar ratio of substrates, temperature, others), co-solvents,
conversion and enantiomeric excess of the esterification of the pro-
fen with alcohols. In this regard, the immobilized lipase B of Candida
antarctica known as Novozym® 435 is the most widely investigated
(see entries 1, 2, 4–7, 10, 12, 15 and references within Table 1S). The
enantiomeric esterification of racemic ketoprofen with alcohols
such as methanol, ethanol, 1-propanol, 1-butanol and dodecanol
between others, was assayed with various organic co-solvents
(1,2 diisopropyl ether; CCl4, hexane, diisopropyl ether, methylene
dichloride, isobutyl methyl ketone, acetonitrile, etc.). The investi-
gation reported by Park et al. is probably the most complete one
on that topic (entry 4 in Table 1S, ref. [18]). The authors found that
short chain alcohols enhanced the reaction rate in the presence
of a mixture of ethylene dichloride and n-hexane 20% (v/v) as co-
solvent. In particular, the use of ethanol as acyl acceptor and solvent
provided 6% of profen conversion with 2% of enantiomeric excess
toward the S-enantiomer at 37 ◦C in 75 h of reaction (see entry 4,
center paragraph in Table 1S). A more recent investigation by De
Crescenzo et al. reported that there was no reaction of 1-dodecanol
and ketoprofen in a solventless media (entry 6, ref. [20]). However,
they prepared propyl ketoprofenate by direct esterification with 1-
propanol for 4 days with the solely aim of obtaining the ester for
calibration purposes. The authors determined a much higher rate
of reaction when using 1-propanol (900 ␮mol h−1 g−1 of enzyme)
instead of 1-dodecanol (200 ␮mol h−1 g−1 of enzyme) with xylenes
as co-solvents. Additionally, the rate of reaction resulted lower
(48 ␮mol h−1 g−1 of enzyme) when solely 1-propanol was used.
More recently, some of us assayed the esterification of racemic
ketoprofen with 2-propanol without success (see entry 15 in Table
1S, ref. [29]).
the minimum amount required to dissolve the profen. The molar
ratio of alcohol: R/S-ketoprofen is 9 for methanol and ethanol,
and 13 for 1-propanol. The esterification of R/S-ketoprofen was
catalyzed with 160 mg of Novozym® 435. This commercial bio-
catalyst was employed as received and previously pretreated with
the alcohols (according to the methodology described in Section
2.3) before the esterification reaction. All the reactions were per-
formed in closed 100 mL vials, at a constant temperature of 45 ◦C in
a shaker bath at 200 rpm for 72 h. The analysis of both enantiomers
was conducted by chiral HPLC analysis with a Nucleodex beta-PM
column (Macherey-Nagel, Germany) and an UV detector at 230 nm
Previously, the samples were diluted to 30 ppm and filtered with
nylon hydrophilic membrane filters (Osmonic Inc, 0.45 ␮m pore
size, 13 mm diameter). A volume of 10 ␮L was injected for anal-
ysis. The mobile phase (methanol/0.1% TEAA pH 4.0 (50/50 v/v))
was operated at a flow rate of 0.700 mL/min. The retention time
of S(+)-ketoprofen, R(−)-ketoprofen and esters were 11.665 min,
12.187 min and 22.844 min, respectively. The resolution factor Rs
was of 0.5. All samples were run a minimum of four times with a rel-
ative error of 2.4% and standard deviation of 0.5 units. Enantiomeric
excess (ee) referred to the form S(+) of the remaining ketoprofen
was calculated according to the Eq. (1), where Area S(+) and Area
R(−) account for the chromatographic areas of the S(+) and R(−)
enantiomers, respectively.
ꢀ
ꢁ
Area S(+) − FC ∗ Area R(−)
Area S(+) − FC ∗ Area R(−)
FC is a correction factor employed to solve the difference in sen-
sibility of the detector to each enantiomer. It was calculated as the
ratio between the area of the S(+) and R(−) enantiomer (Eq. (2)) of
a 30 ppm solution of R/S-ketoprofen.
ee S(+)% =
× 100
(1)
Area S(+)
Fc
=
(2)
To the knowledge of the authors there is no additional informa-
tion of the solventless esterification of racemic ketoprofen besides
those commented above. In this context, the kinetic resolution of
ketoprofen through the esterification with solely short chain alco-
hols (without the addition of co-solvents) is reported for the first
time in the literature in this contribution. A complete picture of
the effect of the short chain alcohols on the catalytic performance
of the CALB lipase in the esterification of profens is provided at
molecular level (through theoretical calculations) and experimen-
tal evidences.
Area R(−)
The conversion of profen was determined by titration of the final
reaction mixture with a basic solution of KOH in ethanol of known
concentration [30]. The titrations were performed in triplicate with
a relative error of 4% and a standard deviation of 3 units.
The specific activity of the biocatalyst was calculated as the ratio
between the conversion of the profen to esters (in ␮mol) per weight
(mg) of protein and the reaction time (min).
Relative errors involved in conversion, enantiomeric excess and
in the specific activity are indicated with errors bars in figures pre-
sented in this investigation.
2. Experimental
2.3. Procedure for the treatment of Novozym® 435 with the
alcohols
2.1. Materials
The commercial biocatalyst Novozym® 435 (batch LC200217)
was a gift of Novozymes Brasil (Paraná, Brazil). R/S-ketoprofen
(Parafarm, 99.80%, batch 030718 000928/004); methanol (Tedia),
ethanol (Carlo Erba) and 1-propanol (Sigma–Aldrich, ≥99.5%) and
potassium hydroxide (1 mol/L) in ethanol (Riedel-de Haën) were
used.
A sample of Novozym® 435 (1.000 g) was contacted with
10.00 mL of a mixture of the alcohols with 4.76% (v/v) H2O at 45 ◦C
and 200 rpm for a period of time of 8 days in order to investigate
the effect of the alcohol over the biocatalyst [29,31].
Then, the beads were washed 4 times with 5.00 mL of the
corresponding alcohol and filtered with nylon hydrophilic mem-
brane filters (0.45 ␮m pore size, 13 mm diameter) obtained from
Osmonic Inc. to retain the non-soluble substances as polymethyl-
methacrylate. The biocatalyst was dried in a desiccator for 8 days
(to dehydrate) and further heated for 10 min at a certain tempera-
ture necessary to desorb the alcohol (the desorption temperatures
were determined as described in Section 2.4), and after a step of
cooling down, the solid was weighed. This procedure allowed the
determination of the total weight loss of the biocatalyst and the
relative amount of alcohol adsorbed on the material.
2.2. Esterification of profens with short chain alcohols
The esterification of R/S-ketoprofen with methanol, ethanol
and 1-propanol as reactants and solvents catalyzed by Novozym®
435 were carried out at the optimum operative conditions pre-
viously determined for the esterification of R/S-ibuprofen with
ethanol [30]. R/S-ketoprofen (0.5000 g; 1.966 mmol) was dissolved
in 0.70 mL (17.28 mmol) of methanol; 1.00 mL (17.15 mmol) of
ethanol; 1.9 mL (24.98 mmol) of 1-propanol with 4.76% of water
added corresponding to the optimum amount of water for the
performance of the biocatalyst [30]. Those volumes correspond to
The liquids remaining after each washing step were added to
the liquid phase initially separated and allowed to dry. The resul-
tant solid was dissolved with 2.00 mL of distilled water, centrifuged

54
M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
to separate the non-soluble substances and recover the enzyme
for further quantification through the bicinchoninic acid assay
as described in Section 2.5. The biocatalyst in contact with the
alcohol:H2O mixtures and the solids retained in the filters were
analyzed by infrared spectrometry.
vibration of O-H species (mainly from adsorbed water) that typ-
ically appears at 1640 cm−1. The percentage contribution of each
structure was obtained through the deconvolution, integration and
further normalization of the corresponding signals involved in the
Amide I. The methodology employed was described previously
when the secondary structure of Novozym® 435 in contact with
2-propanol was studied [29].
2.4. Temperature programmed desorption spectra
Novozym® 435 previously treated with the corresponding alco-
hol for 8 days was studied by temperature programmed desorption
TPD analysis. Details of the equipment used in this investigation
have been published before [29,31,32]. The samples (49.9 mg) were
heated up to 400 ◦C at 10 ◦C/min under a flow of pure helium
(35 cm3 (NTP) min−1) for the temperature programmed desorption
experiment. The species resulting of desorption and/or reaction of
the surface species were detected in the mass spectrometer and
recorded in a computer.
2.7. Scanning electron microscopy (SEM) analysis and fractal
dimension estimator calculation
The internal texture of Novozym® 435 before and after contac-
ting the alcohols for 8 days and the biocatalyst (with and without
previous exposure to the alcohols) after 72 h of reaction were inves-
tigated through scanning electron microscopy analysis with an
environmental ESEM FEI Quanta 400 microscope. The samples were
prepared as ultra thin specimens by embedding the biocatalyst
in a LR White resin and further sliced with a microtome. These
specimens, covered with a conductive gold layer in order to avoid
electrical charges on the surface, were observed in the microscope
in the high vacuum mode.
2.5. Quantification of the protein through the bicinchoninic acid
assay
The images of the samples at 20,000× magnifications were taken
and analyzed with the FERImage program to calculate the fractal
dimension D and dmin parameters by using the variogram method-
ology [33–36].
The amount of protein that was recovered from the treatment
of Novozym® 435 described in Section 2.3, was quantified through
the bicinchoninic acid assay. The BCA method involves the prepara-
tion of three aqueous solutions that are called A, B and C for brevity.
Solution A contains 0.4000 g of BCA (Fluka 90%), 0.8000 g Na2CO3
(Mallinckrodt), 0.1630 g NaHCO3 (Anedra), 0.1581 g NaOH (Carlo
Erba 97%) and 0.0641 g sodium tartrate dibasic dihydrate (Fluka
99%) in 40 mL of distilled water. Solution B contains 4% w/v of CuSO4
and solution C is a mixture of solution A and B in a 100:2 v/v ratio
(in this particular case, solution C corresponds to 30.00 mL of solu-
tion A and 0.60 mL of solution B, and is prepared just before the
quantification).
The calibration curve was obtained by employing highly pure
Candida antarctica lipase B (Sigma–Aldrich Argentina 10.9 U/mg).
The calibration was performed in distilled water. Initially, a CALB
solution (0.0044 g) in distilled water (4.00 mL) was obtained.
The actual concentration of the solution was obtained from the
absorbance value at 280 nm (A280) and the extinction coeffi-
cient of the enzyme (ε280 CALB = 41,285 M−1) according to the
Beer–Lambert law: A280 = ε280*b*c. The concentration of this start-
ing solution was 0.1956 mg/mL. Various calibration solutions were
obtained by dilution with distilled water, resulting in nine solutions
of concentrations between 0 and 200 ␮g/mL. The bicinchoninic acid
assay involved the reaction of 170 ␮L of each calibration solution
with 1.70 mL of solution C (working reagent) for 10 min at 60 ◦C. The
mixtures were allowed to cool down and then the absorbance was
measured at 562 nm with an equipment Perkin Elmer Lambda 35
Spectrophotometer using cuvettes Uvette® 220–1600 nm of 1 cm
path length. The calibration curve obtained fits the following linear
equation:
2.8. Theoretical study
To explore the importance of steric interactions at the level
of the active site and neighborhoods of the catalytic triad from
CALB, a simple molecular mechanics study was carried out, using
the same model and structures than previously reported in other
manuscripts from some of us [29]. The main interest was on the
alcohol therefore ketoprofen enantiomer was only of slight impor-
tance. Methanol, ethanol and 1-propanol were modeled and the
initial situation, the tetrahedral intermediary 1 and the acyl enzyme
were simulated.
3. Results
3.1. Esterification of R/S-ketoprofen with short chain alcohols
without co-solvents added
The esterification of R/S-ketoprofen with methanol, ethanol and
1-propanol (with 4.76% of water added) as reactants and solvents
was catalyzed with Novozym® 435. The biocatalyst was employed
with and without previous treatment with the alcohol for 8 days
(264 h) with the aim of obtaining evidences on the effect of the alco-
hol on the biocatalyst’s performance. Fig. 1A–D shows the specific
activity and the enantiomeric excess toward the S(+)-enantiomer
in the esterification of R/S-ketoprofen with the alcohols. The results
obtained with 2-propanol are also presented for comparison pur-
poses [29]. Additionally, the results of the esterification in terms
of conversion are presented in the supplementary information
accompanying this article (see Figs. S1–S4).
The comparison of the specific activities obtained with the bio-
catalyst with and without previous contact with the alcohols clearly
evidences that the esterification of the profen with 2-propanol is
not feasible. This behavior is attributed to the steric hindrance of
the secondary alcohol and is independent of the effect of the alco-
hol on the integrity of the biocatalyst itself [29]. In contrast, the
esterification with methanol proceeds with the starting biocatalyst
(with a similar specific activity as ethanol and 1-propanol) but falls
down after extended contact with the alcohol. This observation is
an evidence of the deleterious effect of the acyl acceptor on the
biocatalyst as will be discussed in the following sections.
A562 = 0.0029 ∗ [protein (␮g/mL)] − 0.0354 r2 = 0.9917
(3)
2.6. Determination of the protein’s secondary structure with
Diffuse Reflectance Infrared Fourier Transform Spectroscopy
(DRIFTs)
The secondary structure of CALB in Novozym® 435 without pre-
vious treatment along with CALB in the biocatalyst exposed to the
various alcohols for 8 days and the evolution of the secondary struc-
tureuponreactionwasstudiedthroughDiffuseReflectanceInfrared
Fourier Transform Spectroscopy (DRIFTs). Previous to DRIFTs anal-
ysis, isotopic exchange of water molecules by D2O molecules was
carried out. This technique allows the investigation of the Amide I
signal (1700–1600 cm−1) without the interference of the bending

M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
55
C
0.05
A 0.05
0.04
Novozym® 435 as received
8 days with alcohol, 264 hs
1-propanol
2-propanol
methanol
ethanol
1-propanol
2-propanol
methanol
ethanol
0.04
0.03
0.02
0.01
0.00
0.03
0.02
0.01
0.00
0
10
20
30
40
50
60
70
80
0
10
20
30
40
50
60
70
80
Time (hs)
Time (hs)
Novozym® 435 as received
8 days with alcohol, 264 hs
B
D
60
60
50
40
30
20
10
0
50
40
30
20
10
0
1-propanol
2-propanol
methanol
ethanol
1-propanol
2-propanol
methanol
ethanol
0
10
20
30
40
50
60
70
80
0
10
20
30
40
50
60
70
80
Time (hs)
Time (hs)
Fig. 1. Enantiomeric excess and specific activity of the esterification of (R/S)-ketoprofen with methanol, ethanol, 1- and 2-propanol catalyzed with Novozym® 435 as received
(Fig. 1A and B) and previously treated with the corresponding alcohol for 8 days (Fig. 1C and D).
The esterification of R/S-ketoprofen with both ethanol and 1-
propanol in a solvent-free system is a feasible process and it is
not influenced by an extended contact of the biocatalyst with
those alcohols. The esterification of R/S-ketoprofen reaches 50%
conversion toward the methyl and propyl esters (specific activ-
ity ∼1.4 × 10−2 ␮mol min−1 mg−1) at 48 h of reaction regardless
of the treatment of the catalytic material (see Figs. S2 and S3 in
the supplementary information). A similar behavior is observed for
the enantiomeric excess toward the S(+)-enantiomer that climbs to
about 56% at 72 h of reaction (Fig. 1B and D).
The esterification of ibuprofen was also studied for comparison
purposes. These investigations demonstrated that in contrast with
the above observations, the extended contact of Novozym® 435
with the alcohols exerts a detrimental effect in the enantiomeric
esterification of ibuprofen (see Figs. S5 and S6 in the supplementary
information).
forms the internal core of the biocatalyst‘s beads [29,31]. Simi-
larly to those observations, the extended exposure to methanol
and 1-propanol shows a similar trend. In fact, the solids recov-
ered from the organic medium possess CALB enzyme and the
species attributed to the macroporous matrix which proves again
the disaggregation of the biocatalyst (see the infrared spectra in
the Fig. S7 in the supplementary information). The quantifica-
tion (see Table 1) demonstrates that the shorter is the alcohol the
higher is the percentage of protein loss. The effect of 1-propanol in
terms of protein desorption is much less pronounced than the one
observed with the other alcohols. Again, the total weight lost with
methanol and ethanol (11.6% and 16.6%, respectively) is an order
of magnitude higher than with 1- and 2-propanol (5.9% and 1.2%,
respectively).
The experiments of desorption at programmed temperature
demonstrated that methanol and 1-propanol adsorb over the sur-
face/bulk of the biocatalyst as already reported with other alcohols.
The TPD spectra are shown in Fig. 2A to D and the temperatures of
desorption and the amount of adsorbed alcohols are presented in
the Table 1.
3.2. Stability of Novozym® 435 upon extended exposure to short
chain alcohols
The desorption profiles show that methanol (m/e = 31), ethanol
(m/e = 31), 1-propanol (m/e = 31) and 2-propanol (m/e = 45) des-
orb molecularly at 154 ◦C, 184 ◦C, 200 ◦C and 187 ◦C, respectively.
These high temperatures of desorption indicate a strong alcohol-
biocatalyst interaction that can be ascribed as a physisorption
moreover considering that the alcohols desorbs as intact molecules
[37–39]. However, the presence of dimethyl ether (m/e = 45 in
Fig. 2A) and propene (m/e = 41 in Fig. 2C) is an evidence of a
Novozym® 435 was contacted with a mixture of the correspond-
ing alcohol with 4.76% (v/v) H2O for 8 days in order to elucidate
the effect on the physical integrity of the biocatalyst. Previous
investigations evidenced the dissolution of Novozym® 435 in con-
tact with ethanol and 2-propanol which in turn induces the loss
of the surface protein (lipase B of Candida antactica) and poly-
methylmethacrylate (PMMA) (known as Lewatit VP OC1600) that

56
M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
Table 1
Summary of the percentage of weight and proteins lost; temperature of desorption of the alcohols and alcohol adsorbed upon contacting Novozym® 435 with methanol,
ethanol, 2-propanol and 1-propanol.
Alcohol
Mass (g)
T
b
Starting biocatalyst
Recovered biocatalyst
% weight loss
% protein loss a
(◦C)
Mass alcohol
D
adsorbed/gr biocatalyst
Methanol
Ethanolc
1.0002
1.0000
1.0005
1.0011
0.8846
0.8340
0.9412
0.9889
11.6
16.6
5.9
1.93
1.27
0.57
0.79
154
184
200
187
0.0212
0.3881
0.0404
0.0368
1-Propanol
2-Propanold
1.2
a
ratio between the amount of protein lost (mg) due to the alcohol treatment and the total amount of protein (mg) of Novozym® 435.
temperature of desorption of the alcohols obtained from the temperature programmed desorption analysis.
b
c
,
d
The data corresponding to Novozym® 435 in contact with ethanol and 2-propanol were taken from Ref. [31] and [29], respectively.
certain degree of chemisorption of methanol and 1-propanol. The
chemisorption of the alcohols generates intermediate methoxy and
propoxy species that further dehydrate over active sites of acidic
nature [40–42].
Two desorption events of water molecules (m/e = 18) are
observed. The first one below 100 ◦C corresponds to the loss of con-
stitutional water of the proteins. A second event happens when
the temperature rises (T > 239 ◦C) and is accompanied with the
A
B
154 ºC
CH3OH (m/e=31)
H2O (m/e=18)
CO (m/e=28)
184oC
H2O (m/e=18)
C2H6O (m/e= 45)
CH2O (m/e= 29)
ethanol (m/e=31)
CO2 (m/e=44)
50
100
150
200
250
300
50
100
150
200
250
300
350
TEMPERATUREO(C)
TEMPERATUREO(C)
D
C
1-C3H7OH (m/e=31)
200 ºC
H2O
2-C3H7OH (m/e= 45)
187oC
C3H6
m/e= 41
H2O
C3H6O (m/e=58)
m/e=18
C3H(m/e= 41)
6
C3H6O (m/e= 58)
50
100
150
200
250
300
0
50
100
150
200
250
300
Temperature (ºC)
TEMPERATUREO(C)
Fig. 2. In situ temperature programmed desorption spectra of Novozym® 435 after being in contact with methanol (A), ethanol (B), 1-propanol (C) and 2-propanol (D) for 8
days at 45 ◦C.

M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
57
Fig. 3. Images of the cross section of the beads of Novozym® 435 as received (A) and after being in contact with methanol (B), ethanol (C), 1-propanol (D) and 2-propanol (E)
for 8 days at 45 ◦C. The Images were taken with an environmental scanning electron microscope at 20,000×.
texture of starting Novozym® 435 and after being in contact with
the alcohols for 8 days at 45 ◦C.
desorption of CO and CO2 which corresponds to the decomposition
of amino-acids and the PMMA support [31].
The adsorption of the alcohols and the degradation of the poly-
meric matrix do not only occur at the surface but also in the
interior of the beads. The investigation of the cross section of the
beads of Novozym® 435 through scanning electron microscopy
and the parameters (fractal dimension and minimum cell size)
that describe the surface roughness and the texture are presented
in Fig. 3 and Table 2. The Fig. 3A–E shows images of the inner
The fractal dimension values ranging from 2 < D < 2.5 indicate
a persistent smooth surface while values 2.5 < D < 3 are an indica-
tion of anti-persistence being a completely rough surface when D
is close to 3 [33,34]. The starting Novozym® 435 possesses a fractal
dimension equals to 2.8347 indicating a certain rough texture and
a minimum cell size dmin of 0.0582 ␮m. Nevertheless, the extended
exposure to the alcohols and even under reaction conditions

58
M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
Table 2
Fractal dimension estimator D and minimum cell size dmin of starting Novozym® 435, after being in contact with short chain alcohols for 8 days at 45 ◦C.
Texture estimators
Starting Novozym® 435
Treated with alcohols for 8 days
Methanol
Ethanol
1-Propanol
2-Propanol
Fractal dimension D
dmin [␮m]
2.8347 0.0027
0.0582 0.0022
2.595 0.054
0.05952 0.0017
2.665 0.004
0.1015 0.022
2.460 0.042
0.0692 0.017
2.727 0.022
0.0651 0.0032
(without previous exposure to alcohols) diminishes the D parame-
ter and increases the dmin. This observation certainly evidences that
the alcohols diffuse into the beads and exert a smoothing action due
to the degradation of the polymeric matrix as discussed before (see
the descriptors of the texture of the biocatalyst after reaction in
Table S2 included in the supplementary information).
methanol. The investigations by Fang et al. also demonstrated that
methanol induces the modification of the conformation of CALB
[47]. Although the authors observed the decrease in the ␣-helix
structure (in agreement with our results) they did not detected the
formation of aggregates but certainly, they correlated the inacti-
vation of the enzyme with the protein’s unfolding caused by the
alcohol.
Similarly, a high content of the ␤-turns structure, a decrease of
the contribution of ␤-sheet and random structure and a percent-
age of ␣-helix almost constant was observed when the biocatalyst
(previously treated with the alcohols or not) was used in the ester-
ification of R/S-ketoprofen with the short chain alcohols for 72 h.
3.3. Evolution of the secondary structure of the lipase B of
Candida antarctica
The percentages of aggregates, ␣-helix, ␤-sheet, turns and ran-
dom in the secondary structure of starting Novozym® 435 and after
exposure to the alcohol-water mixtures for 8 days are presented in
the Table 3. Additionally, the Tables S3–S5 (within the supplemen-
tary information) show the secondary structure of the biocatalyst
during the esterification of R/S-ketoprofen with methanol, ethanol
and 1-propanol (the results obtained with 2-propanol were already
published in [29]). The analysis of the infrared spectra of enzymes
in the 1700–1600 cm−1 region (Amide I region) provided qualita-
tive and quantitative information about the secondary structure of
the protein from CALB in Novozym® 435.
The percentage contribution of each structure was obtained
through the deconvolution, integration and further normaliza-
tion of the corresponding signals of Amide I. The contribution of
the ␣-helix and random structures correspond to the area of the
signals at 1654 cm−1 and 1643 cm−1, respectively. The contribu-
tion of the ␤-sheet structure was obtained by adding the area
of the signals appearing at 1631 cm−1, 1637 cm−1 and 1686 cm−1
wavenumbers. Similarly, the contribution of the ␤-turn structure
correspo−n1ded to the addition of the areas of the signals located at
1664 cm , 1666 cm−1 and 1676 cm−1 [29,43–46]. Finally, a signal
at 1619 cm−1 corresponds to ␤ aggregates.
The secondary structure of the protein CALB of the starting
Novozym® 435 is composed by 26.4% of ␣-helix, 19.4% of ␤-
sheet, 25.3% of random structure, 28.7% of ␤-turns and 0.2% of
aggregates. The biocatalyst exposed to the alcohol-water mixtures
for a period of 8 days presents a modified secondary structure
where the percentage of ␤-turns increases and the contribution
of ␤-sheet and random structure diminishes although the con-
tent of ␣-helix remains almost unaltered. The aggregation of the
protein (6%), the increase in ␤-sheet and the modification of
the ␣-helix are only observed in the case of the exposure to
3.4. Esterification of ketoprofen at molecular level. Theoretical
calculations
Fig. 4 shows the interaction between ethanol and 1-propanol in
the neighborhood of the catalytic triad of CALB, at the coordina-
tion step. Additionally, Fig. 5 shows the results for coordination,
intermediary 1 and the acyl enzyme for methanol, ethanol and
1-propanol. It is important to take into account that the bonds
involved are similar but the total number of atoms is not. The com-
parison is only qualitative. As a general rule, the most negative the
steric energy, the better for the stabilization of the conformer.
The MM2 calculation detected the difference among methanol,
ethanol and 1-propanol in steric energies, being always the same
R/S-ketoprofen with the structure of the whole molecule at the
initial step, the intermediary 1 and finally the acyl enzyme.
With methanol the ketoprofen carbonyl is located at the pocket
where the alcohol has to be coordinated and interferes clearly with
the formation of the tetrahedral intermediary being positioned
between the ketoprofen and the Thr40. In the case of ethanol and 1-
propanol the position of the acyl enzyme is correct to react with the
alcohol later. The carbon 1 of propanol is not at the same plane than
the carbon from the acyl group. The problem with propanol seems
to be in the transition states more than in the stable intermediaries,
probably due to the impact of the additional methylene group. The
methanol introduces steric and electronic hindrance from the step
of the coordination of the R/S-ketoprofen.
The alcohol inhibition effect seems to be related to: (a) the
relative occupancy of the alcohol before the ketoprofen coordina-
tion and the steric hindrance involved, (b) how the alcohol affects
Table 3
Contribution of the aggregates, ␣-helix, ␤-sheet, random structure and ␤-turns to the secondary structure of the starting Novozym® 435, after treated with methanol,
ethanol, 1- and 2-propanol for 8 days at 45 ◦C without reaction.
Novozym® 435
PERCENTAGE CONTRIBUTION
Aggregatesa
␣-Helixb
␤-Sheetc
Randomd
␤-Turnse
Starting biocatalyst
Methanol
Ethanol
1-Propanol
2-Propanol
0.2
6.0
0.2
0.4
0.0
26.4
16.6
23.6
23.8
25.3
19.4
22.3
5.7
13.2
18.6
25.3
16.3
7.5
9.3
3.6
28.7
38.8
63.0
53.3
52.5
a
Corresponds to the signal at 1619 cm−1
Corresponds to the signal at 1654 cm−1
.
.
b
c
Corresponds to the sum of the contribution of the signals at 1631 cm−1, 1637 cm−1 and 1686 cm−1
.
d
e
Corresponds to the signal at 1643 cm−1
.
Corresponds to the sum of the signals at 1664 cm−1, 1666 cm−1 and 1676 cm−1
.

M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
59
Fig. 4. Representation of the interaction of ethanol and 1-propanol with the catalytic triad of CALB and R/S-ketoprofen at the coordination step.
Steric energy of ketoprofen and alcohols near CALB
-150
0
1
2
3
4
-160
-170
-180
-190
-200
-210
inicial
int tetra
acyl E
1= Methanol, 2= Ethanol, 3= 1-Propanol
Fig. 5. Steric energy of ketoprofen and methanol, ethanol and 1-propanol near CALB. The initial step, the intermediary 1 (tetrahedral) and the acyl enzyme are considered.
the coordination of the ketoprofen, and the carboxylate group of
ketoprofen with serine to produce the intermediary 1 and the
stabilization through the oxyanion hole, (c) how the acyl enzyme
formation is stabilized or not, and (d) proper orientation of the
alcohol to produce the reaction with the acyl enzyme, once it is
formed.
that the secondary alcohol interacts with the lipase B of Candida
antarctica inhibiting the catalytic performance of Novozym® 435
[29]. The evidences reported in this investigation demonstrate that
the esterification of R/S-ketoprofen with ethanol and 1-propanol
(without co-solvents added) at 45 ◦C is a feasible process and more
importantly, the activity of Novozym® 435 is not influenced by the
extended contact with the alcohols.
In contrast with that observation, the catalytic performance
decreases when methanol is used as the acyl acceptor. This behav-
ior is attributed to various detrimental effects of this alcohol on
the physical integrity of the polymer that holds the lipase, the
enzymatic structure itself and the interaction with the active site.
Previous investigations by some of us demonstrated that ethanol
and 2-propanol are able to diffuse inside the polymeric beads of the
4. Discussion
The present investigation is a continuation of previous ones
devoted to the kinetic resolution of profens (namely, ibuprofen
and ketoprofen) with solely alcohols catalyzed with Novozym® 435
[29,31]. Previous investigations, using isopropyl alcohol evidenced

60
M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
biocatalyst leading to the swelling and dissolution of the PMMA
[29–31,48]. The low molecular weight of methanol increases its
ability to diffuse onto polymeric structures as reported by Koenig
et al. [38,39]. Again this alcohol performs a smoothing action of the
internal texture as discussed before.
The exposure of the protein to the alcohols affects the contribu-
tion of the ␤-turn structure that increases from 28.7% in the starting
biocatalyst to more than 50% upon contacting with ethanol and
1- and 2-propanol for 8 days, with/without reaction with racemic
ketoprofen. In contrast, methanol induces to the decrease of the
contribution of the ␣-helix and the increase of the aggregates which
evidences the denaturation of the protein.
The “real situation” many times is not considered from the the-
oretical and practical point of view. When the lipase is exposed to a
mixture of high molar ratio short, linear alcohol: carboxylic acid, it
is clear that the alcohol reaches faster the catalytic triad and it may
introduce additional steric problems functioning as solvent and as
substrate. From the mechanistic point of view, to have reaction,
first the tetrahedral intermediary 1 has to be formed and later the
acyl enzyme is generated. Later, the reaction with the alcohol takes
place. The first step, the step that is required to the alcohol to react,
is the formation of the acyl enzyme by the reaction of the carboxylic
acid of the ketoprofen with the serine alcohol group.
1-propanol and ethanol showed different conformations and
distributions than methanol around ketoprofen and induced or
contributed to different conformations for carboxylic acid group
in ketoprofen, around the catalytic triad of CALB. However, the ini-
tial conformer, intermediary 1 and acyl enzymes were very similar
when 1-propanol and ethanol were used. The difference probably
can be found in the transition states or unstable transition states
and higher activation energies in the case of 1-propanol compared
to ethanol. The inhibition of activity with methanol can be under-
stood, whereas the ethanol and 1-propanol differences were based
in steric interactions at unstable transition states. This proposal
should have to be tested with higher levels of molecular mechanics
and/or semiempirical or ab-initio calculations.
the smaller molecules. This observation is somehow expected
considering their capability of diffusion inside the biocatalyst’s
beads. The alcohols irreversible adsorb on the biocatalyst as spec-
tator species that desorbs above Td > 100 ◦C. The exposure to
the alcohols affects the contribution of the ␤-turn conformation
(from 28.7% in the starting biocatalyst to ∼55%) to the secondary
structure of the protein as already been observed previously.
The catalytic performance of Novozym® 435 in the enantiomeric
esterification of racemic ketoprofen with ethanol and 1-propanol
as acyl acceptors is not influenced by the variety of changes
induced by the alcohols. However, the kinetic resolution of
ibuprofen diminishes under similar conditions. This observation
can not be explained at the moment.
•
•
•
1-propanol and ethanol showed different conformations and
distributions than methanol around racemic ketoprofen and
induced or contributed to different conformations for carboxylic
acid group from ketoprofen.
There is preliminar evidence of the importance of an extended
H-bonding system in the coordination of short, linear alcohols
in CALB catalytic triad and its surroundings. The coadsorption
of alcohols and ketoprofen (or any other profen) has to be con-
sidered because it introduces additional steric hindrance to the
profen coordination near the serine of the catalytic triad of CALB.
Short, linear alcohols without steric hindrance are able to interact
electronically and sterically with different portions of the CALB
protein, not only changing the proportion of secondary struc-
tural facts, but also introducing additional steric problems to the
profen coordination at the catalytic triad of the lipase.
•
Acknowledgments
The authors acknowledge the financial support of CONICET
(project PIP 112 20130100171) and Universidad Nacional de La
Plata (project 11-X626); Prof Luis A. Gambaro for the TPD exper-
iments and Lic Mariela Theiller for the calculation of textural
parameters.
The temperature programmed desorption spectra of the alco-
hols from Novozym® 435 indicated an intense physical adsorption
on the protein and/or the polymeric matrix similarly to previous
findings with ethanol and isopropyl alcohol. The alcohols molec-
ularly desorbed at 154 ◦C, 184 ◦C, 187 ◦C and 200 ◦C for methanol,
ethanol, 2-propanol and 1-propanol, respectively. Interestingly, 1-
propanol adsorbed also as alkoxy species that are intermediates
species to propylene and this compound was also detected in
the TPD spectra. The temperatures of desorption (Td methanol < Td
ethanol < Td 1-propanol) are a measure of the intensity of the inter-
action of the alcohol with the biocatalyst and are in accordance with
the steric energy calculated through theoretical calculations.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.molcatb.2015.05.
003
References
[1] O. Barbosa, C. Ariza, C. Ortiz, R. Torres, New Biotechnol. 27 (2010) 844–850.
[2] M. Kapoor, M.N. Grupta, Process Biochem. 47 (2012) 555–569.
[3] J. Durand, J. Lecomte, B. Baréa, G. Piombo, E. Dubreucq, P. Villeneuve, Process
Biochem. 47 (2012) 2081–2089.
[4] S.K. Karmee, Appl. Microbiol. Biotechnol. 81 (2009) 1013–1022.
[5] M.L. Foresti, M. Pedernera, V. Bucalá, M.L. Ferreira, Enzyme Microb. Technol. 41
(2007) 62–70.
5. Conclusions
[6] M.D. Romero, L. Calvo, C. Alba, A. Daneshfar, H.S. Ghaziaskar, Enzyme Microb.
Technol. 37 (2005) 42–48.
[7] A.C. Wu, P.Y. Wang, Y.S. Ling, M.F. Kao, J.R. Chen, J.F. Ciou, S.W. Tsai, J. Mol. Catal.
B: Enzym. 62 (2010) 235–241.
The results discussed above allow concluding:
[8] H. Caner, E. Groner, L. Levy, Drug Discov. Today 9 (2004) 105–110.
[9] R. Morrone, N. Dı´Antona, D. Lambusta, G. Nicolosi, J. Mol. Catal. B: Enzym. 65
(2010) 49–51.
[10] M.G. Kim, E.G. Lee, B.H. Chung, Process Biochem. 35 (2000) 977–982.
[11] J.F. Ciou, P.Y. Wang, A.C. Wu, S.W. Tsai, Process Biochem. 46 (2011) 960–965.
[12] J.B. Sontakke, G.D. Yadav, Ind. Eng. Chem. Res. 50 (2011) 12975–12983.
[13] L.C. Bencze, C. Paizs, M.I. Tos¸ a, M. Trif, F.D. Irimie, Tetrahedron: Asymm. 21
(2010) 1999–2004.
[14] A. Kamal, M.A. Azhar, T. Krishnaji, M.S. Malik, S. Azeeza, Coord. Chem. Rev. 252
(2008) 569–592.
[15] M. Arroyo, J.V. Sinisterra, J. Org. Chem. 59 (1994) 4410–4417.
[16] G. Duan, C.B. Ching, E. Lim, C.H. Ang, Biotechnol. Lett. 19 (1997) 1051–1055.
[17] T. López-Belmonte, A.R. Alcántara, J.V. Sinisterra, J. Org. Chem. 62 (1997)
1831–1840.
•
The feasibility of the enzymatic kinetic resolution of racemic
ketoprofen through the direct esterification with ethanol and 1-
propanol (with 4.76% v/v H2O added) as substrates and solvents
without the addition of organic co-solvents. The conversion and
enantiomeric excess (toward the S(+)-enantiomer) were ∼48%
after 48 h at 45 ◦C in batch reaction.
•
A careful analysis of the effect of extended contact of methanol,
ethanol, 1- and 2-propanol with Novozym® 435 (up to 8 days)
revealed that the alcohols diffuse inside the biocatalyst’s beads
causing dissolution/de-aggregation of PMMA with the loss of
protein. Interestingly the weight loss (accounting for polymeric
matrix and protein) is higher for those alcohols that possess
[18] H.J. Park, W.J. Choi, E.C. Huh, E.Y. Lee, C.Y. Choi, J, Biosci. Bioeng. 87 (1999)
545–547.

M.V. Toledo et al. / Journal of Molecular Catalysis B: Enzymatic 118 (2015) 52–61
61
[19] W. Du, M.N. Zong, W.F. Chen, Y. Guo, Q. Li, Microbiology 27 (2000) 429–432.
[20] G. De Crescenzo, A. Ducret, M. Trani, R. Lortie, J. Mol. Catal. B: Enzym. 9 (2000)
49–56.
[21] N. D’Antona, P. Lombardi, G. Nicolosi, G. Salvo, Process Biochem. 38 (2002)
373–377.
[32] L.A. Gambaro, L.E. Briand, Appl. Catal. A: Gen. 264 (2) (2004) 151–159.
[33] F.D. Bianchi, R.D. Bonetto, Scanning 23 (2001) 193–197.
[34] J.L. Ladaga, R.D. Bonetto, in: P.W. Hawkes (Ed.), Characterization of Texture
in Scanning Electron Microscope Images. Advances in Imaging and Electron
Physics, vol. 120, Academic Press, 2002, pp. 136–189.
[22] A.R. Alcántara, P. Dominguez de María, M. Fernández, M.J. Hernaíz, J.M. Sánchez
Montero, J.V. Sinisterra, Food Technol. Biotechnol. 42 (2004) 343–354.
[23] C.-S. Chang, C.-S. Hsu, J. Chem. Technol. Biotechnol. 80 (2005) 537–544.
[24] S. Zhu, Y. Wu, Z. Yu, Biocatal. Biotransform. 23 (2005) 97–101.
[25] A.L. Ong, A.H. Kamaruddin, S. Bathia, W.S. Long, S.T. Lim, R. Kumari, Enzyme
Microb. Technol. 39 (2006) 924–929.
[26] P. De Oliveira Carvalho, F. Jares Contesini, M. Ikegaki, Braz. J. Microbiol. 37
(2006) 329–337.
[27] A.L. Ong, A.H. Kamaruddin, S. Bathia, H.Y. Aboul-Enein, J. Sep. Sci. 31 (2008)
2476–2485.
[28] X. Li, T. Liu, L. Xu, X. Gui, F. Su, Y. Yan, Biotechnol. Bioprocess. E 17 (2012)
1147–1155.
[29] M.V. Toledo, C. José, S.E. Collins, R.D. Bonetto, M.L. Ferreira, L.E. Briand, J. Mol.
Catal. B: Enzym. 83 (2012) 108–119.
[30] M.L. Foresti, M. Galle, M.L. Ferreira, L.E. Briand, J. Chem. Technol. Biotechnol. 84
(2009) 1461–1473.
[31] C. José, R.D. Bonetto, L.A. Gambaro, M.P. Guauque Torres, M.L. Foresti, M.L.
Ferreira, L.E. Briand, J. Mol. Catal. B: Enzym. 71 (2011) 95–107.
[35] A.P. Pentland, IEEE Trans. Pattern Anal. Mach. Intell. PAMI-6 (6) (1984) 661–674.
[36] R.D. Bonetto, J.L. Ladaga, Scanning 20 (1998) 457–463.
[37] L.E. Briand, in: J.L.G. Fierro (Ed.), Metal Oxides: Chemistry and Applications, CRC
Press Taylor & Francis, Florida, 2005, pp. 353–390.
[38] J. González-Benito, J.L. Koenig, Polymer 47 (2006) 3065–3072.
[39] B.A. Miller-Chou, J.L. Koenig, Prog. Polym. Sci. 28 (2003) 1223–1270.
[40] L.J. Burcham, L.E. Briand, I.E. Wachs, Langmuir 17 (2001) 6164.
[41] L.J. Burcham, L.E. Briand, I.E. Wachs, Langmuir 17 (2001) 6175.
[42] L.J. Burcham, M. Badlani, I.E. Wachs, J. Catal. 203 (2001) 104.
[43] M.L. Foresti, G.M. Valle, R.D. Bonetto, M.L. Ferreira, L.E. Briand, Appl. Surf. Sci.
256 (2010) 1624–1635.
[44] A. Barth, Biochim. Biophys. Acta 1756 (2007) 1073–1101.
[45] A. Barth, C. Zscherp, Q. Rev. Biophys. 35 (2002) 369–430.
[46] C.R. Llerena Suster, C. José, S.E. Collins, L.E. Briand, S.R. Morcelle, Process
Biochem. 47 (2012) 47–56.
[47] X. Fang, Y. Zhan, J. Yang, D. Yu, J. Mol. Catal. B: Enzym. 104 (2014) 1–7.
[48] L.O. Wiemann, P. Weisshaupt, R. Nieguth, O. Thum, M.B. Ansorge-Schumacher,
Org. Res. Dev. (13) (2009) 617–620.