Kinetic resolution of 1,3,6-tri-O-benzyl-myo-inositol by Novozym 435: Optimization and enzyme reuse

Article abstract of DOI:10.1021/op300063f

myo-Inositol derivatives bearing selectively protected hydroxyl groups are relevant precursors of high-value myo-inositols. In the present study, we applied the response surface method to the optimization of kinetic resolution of (±)-1,3,6-tri-O-benzyl-my

Full text of DOI:10.1021/op300063f

Article
pubs.acs.org/OPRD
Kinetic Resolution of 1,3,6-Tri‑O‑benzyl-myo-Inositol by Novozym
435: Optimization and Enzyme Reuse
Evelin A. Manoel,*,† Karla C. Pais,§ Aline G. Cunha,‡ Alessandro B. C. Simas,§ Maria Alice Z. Coelho,†
and Denise M. G. Freire*,‡
†Departamento de Engenharia Bioquímica, Universidade Federal do Rio de Janeiro, Ilha do Fundao, Rio de Janeiro, RJ, Brazil, EQ
̃
‡Departamento de Bioquímica, Programa de Pos
Brazil
́
-Graduaca̧ o em Bioquímica, Universidade Federal do Rio de Janeiro, Rio de Janeiro,
̃
§Nucleo de Pesquisas de Produtos Naturais (NPPN), Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil
́
ABSTRACT: myo-Inositol derivatives bearing selectively protected hydroxyl groups are relevant precursors of high-value myo-
inositols. In the present study, we applied the response surface method to the optimization of kinetic resolution of ( )-1,3,6-tri-
O-benzyl-myo-inositol by Novozym 435 (immobilized lipase B from Candida antarctica) with vinyl acetate in hexane. Reaction
temperature, substrate, acyl donor and enzyme concentrations were set as variables. Through the constructed mathematical
model, optimum condition for this enzymatic condition was established. The feasibility of enzyme recycle was demonstrated.
Lipase B from Candida antarctica (CaL-B) is one of the most
used lipases in biotransformations in many different applica-
tions.12−16 3D-Structures of the CaL-B lipase have shown the
presence of a short oligopeptide helix that may act as a lid and it
has been found to adopt different conformations as a function
of the medium, suggesting a great mobility of the active site
environment.8,17 CaL-B exhibits a very high degree of substrate
selectivity and also has the potential for use in different
reactions, for instance hydrolysis and alcoholysis reac-
tions.7,12,13,18−21 Different studies have discussed how the
immobilization technique modulate the catalytic ability of CaL-
B, endowing high performance22,23 as it is the case of Novozym
435.
In the present work, we disclose our results of optimization
of the kinetic resolution of myo-inositol derivative DL-1
(Scheme 1) by Novozym 435 using vinyl acetate, as activated
acyl donor, in hexane. A statistical method based upon the
response surface methodology (RSM) was applied in order to
determine the defining variables. By this, the statistically
significant variables (for the conversions) were determined by
the fractional factorial design. Once such variables are selected,
the CCRD was applied to define the effect of each factor, its
quadratic term and the expected curvature. The studied
variables in the experimental design were temperature reaction,
substrate, acyl donor and enzyme concentration, whose
interactions led to the mathematical model for this enzymatic
transformation. Thus, through such technique, we sought to
increase the efficiency and economy of this biocatalytic
transformation. In addition, the optimum conditions were
explored in successful reuse experiments.
INTRODUCTION
■
Chiral myo-inositol derivatives (e.g., inositol polyphosphates)
are important probes in cell biology studies. Commonly, the
synthetic routes to these compounds involve optical resolution
of racemic mixtures via derivatizations, which are neither
practical nor efficient.1,2 This explains in part the high prices of
commercially available inositols. Chemoenzymatic routes may
remove such process hurdles and increase the availability of
these important compounds.
In fact, the use of enzymes, especially lipases, in organic
solvents, has proved to be an excellent methodology for the
preparation of chiral drugs.3 Lipases enable high stereo-
selectivities and yields in kinetic resolutions or desymmetriza-
tions. However, notwithstanding the high efficiency and
economy that they bestow, few works have reported on their
application to syntheses of inositols.4−6
We have recently shown that an immobilized lipase B from
Candida antarctica (Novozym 435) is an outstanding
biocatalyst for the kinetic resolution of a extensively protected
myo-inositol.7,8 More recently,9,10 we determined that racemic
1,3,6-tri-O-benzyl-myo-inositol (DL-1) undergoes kinetic reso-
lution by Novozym 435 to afford the acylated product, L-(−)-5-
acetyl-1,3,6-tri-O-benzyl-myo-inositol (L-(−)-2) with a 48.30%
conversion and enantiomeric excess of 97% in 112 h (Scheme
1). Substance DL-1 is a known precursor of bioactive myo-
inositol phosphates and analogs and a potential precursor of
other inositol bisphosphates and trisphosphates.11
Scheme 1. Kinetic resolution of DL-1 by Novozym 435
(CaLB) using vinyl acetate in hexane
To the best of our knowledge, the optimization of
chemoenzymatic syntheses of myo-inositols, as disclosed in
this study, has not been previously reported in the literature.
Received: March 11, 2012
Published: June 21, 2012
© 2012 American Chemical Society
1378
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384

Organic Process Research & Development
Article
Table 1. Fractional factorial design matrix with real values and coded for the variables and responses for conversion (X%)
trials
substrate (mg/mL), S1
acyl donor (mg/mL), S2
enzyme (U), E
temperature (oC), T
conversion (%), X
1
2 (−1)
14 (1)
2 (−1)
14 (1)
2 (−1)
14 (1)
2 (−1)
14 (1)
8 (0)
186.8 (−1)
186.8 (−1)
560.4 (1)
560.4 (1)
186.8 (−1)
186.8 (−1)
560.4 (1)
560.4 (1)
373.6 (0)
373.6 (0)
373.6 (0)
54.56 (−1)
54.56 (−1)
54.56 (−1)
54.56 (−1)
136.4 (1)
136.4 (1)
136.4 (1)
136.4 (1)
81.84 (0)
81.84 (0)
81.84 (0)
30 (−1)
50 (1)
50 (1)
30 (−1)
50 (1)
30 (−1)
30 (−1)
50 (1)
40 (0)
40 (0)
40 (0)
21.46
6.15
2
3
31.21
4.10
4
5
42.00
11.86
41.00
42.65
37.00
35.50
36.00
6
7
8
9
10
11
8 (0)
8 (0)
The fractional design showed the coefficient of determi-
nation (R2 = 96.90%) was highly significant, meaning that the
model was unable to explain only 3.1% of the total variations
and also indicated the significance of the model.
Central Composite Rotatable Design. By building on the
results of the fractional design, the following variables were
selected for the CCRD: substrate concentration, enzyme
activity, and temperature. The results of the statistical treatment
are presented in Table 3. For all cases, eep remained >98%.
The highest conversion that was obtained was 51.66% in trial
2. For all runs, the results were similar to the predicted values
with a maximum deviation of −4.70% (trial 10).
RESULTS AND DISCUSSION
■
Optimization of Kinetic Resolution of DL-1 in Hexane.
Fractional Factorial Design. The first statistical treatment of
combinations of the test variables along with the measured
response values (conversion, X%) are shown in Table 1. For all
cases, 98% ( 0.8) eep were obtained.
The screening by fractional factorial design revealed that the
concentrations of enzyme and substrate are the most relevant
variables concerning the conversion response. Figure 1 shows
how significant the variables are to the conversion (X%).
The screening by means of CCRD revealed that the
interaction between the linear terms of substrate and
temperature (S. T) and the interaction between the linear
terms of enzyme and temperature (E. T) showed no
significance to conversion response (Table 4 shows only
significant variables).
The linear and quadratic terms of substrate concentration
and the quadratic terms of enzyme activity and temperature,
showed a negative effect in the studied range. On the other
hand, the linear terms of enzyme activity and temperature
showed a positive effect in the range studied. As expected,
increase of conversion was observed. In fact, a strong
dependence of conversion on temperature and enzyme activity
was established.
Figure 1. Pareto diagram after the fractional design showing the
variables significant to the conversion (X%). [* = significant factors (p
< 0.1).]
The following second-order model eq 1, which considers the
statistically significant parameters (p < 0.05) in the kinetic
resolution of DL-1, was obtained:
X% = 48.37 − 7.60S − 3.29S2 + 6.14E − 4.70E2
Although the temperature was not a significant factor in the
fractional factorial design, this variable was included in CCRD.
Previously in our studies, it proved to be the determining factor
(Table 2). As acyl donor (vinyl acetate) concentration (S2) was
the least significant variable, such parameter was then fixed (0.3
mL in 1.7 mL of hexane) in the CCRD trials for conversion (X
%) optimization. By means of the fractional design, we could
select a minimum concentration of acylating agent in the
experiments.
+ 7.80T − 5.94T2 + 3.04S·E
(1)
where S, E, and T were the coded values for the substrate
concentration, enzyme activity, and temperature, respectively.
E2 and T2 were the coded values for the quadratic terms of the
enzyme activity and temperature, respectively. The interaction
between the linear terms substrate concentration and enzyme
activity are represented by SE.
In addition, the fit of the statistical model to the experimental
data was confirmed by the ANOVA table (Table 5).
Table 2. Variation of conversion with the temperature in 48-
The F test value found for the regression (27.52) was higher
than the F tabulated value (3.50). Once the F test value was
14.1 times higher than F tabulated value, with high significance
(p = 0.000044), the model may be considered predictive.
Furthermore, the coefficient of determination (R2 = 96.87%)
was highly significant, meaning that the model was unable to
explain only 3.13% of the total variations. The adjusted R2 value
(93.35%) also indicated the significance of the model. Such
a
h reactions
temperature (°C)
conversion (X%)
ee
30
40
50
31
>98
>98
>98
41
45.5
a
Assays with 136.4 U of Novozym 435 and 5 mg of DL-1.
1379
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384

Organic Process Research & Development
Article
Table 3. CCRD matrix with real values and coded for the variables and conversion (X%) responses
trials
substrate (mg/mL), S, S1
enzyme (U), E
temperature (°C), T
conversion (%), X
predicted values (%)
error (%)
1
5 (−1)
5 (−1)
5 (−1)
136.4 (−1)
136.4 (−1)
300.1 (1)
300.1 (1)
136.4 (−1)
136.4 (−1)
300.1 (1)
300.1 (1)
218.2 (0)
218.2 (0)
80.5 (−1.68)
355.8 (1.68)
218.2 (0)
218.2 (0)
218.2 (0)
218.2 (0)
218.2 (0)
218.2 (0)
30 (−1)
50 (1)
27.89
51.66
39.00
51.00
10.46
21.04
28.30
38.00
50.03
31.00
25.00
48.00
18.00
48.00
50.00
48.00
48.00
47.00
27.61
50.24
36.98
53.28
10.20
25.09
31.74
40.31
51.85
26.30
24.72
45.40
18.44
44.67
48.37
48.37
48.37
48.37
−0.28
−1.42
−2.02
2.28
2
3
30 (−1)
50 (1)
4
5 (−1)
5
11.5 (1)
11.5 (1)
11.5 (1)
11.5 (1)
2 (−1.68)
15 (1.68)
8.5 (0)
30 (−1)
50 (1)
−0.26
4.05
6
7
30 (−1)
50 (1)
3.44
8
2.31
9
40 (0)
1.82
10
11
12
13
14
15
16
17
18
40 (0)
−4.70
−0.28
−2.6
0.44
40 (0)
8.5 (0)
40 (0)
8.5 (0)
23 (−1.68)
57 (1.68)
40 (0)
8.5 (0)
−3.33
−1.63
0.37
8.5 (0)
8.5 (0)
40 (0)
8.5 (0)
40 (0)
0.37
8.5 (0)
40 (0)
1.37
tests sufficed to confirm a high degree of adequacy of the
model. Then, the model was used to generate contour plots
and response surfaces (Figures 2 and 3).
As can be seen in Figure 2, the interaction of temperature
and enzyme activity leads to the increase of conversion,
indicating the beneficial effect of both variables.
The surface indicated that the best conversion was obtained
at midpoint of enzyme activity axis and of temperature.
Conversely, the response surface and contour diagram
(Figure 3) for the interaction of substrate concentration and
enzyme activity indicated contrary effects between such
variables.
Table 4. Effect estimates for the conversion (X%) in the
kinetic resolution
a
factor
effect
standard error
t (8)
p value
mean
S
48.3734
−15.1940
−6.5745
12.2914
−9.4129
15.5971
−11.8877
6.0875
1.706094
1.849379
1.921633
1.849379
1.921633
1.849379
1.921633
2.416330
28.35329
−8.21573
−3.42132
6.64625
0.000000
0.000036
0.009070
0.000161
0.001196
0.000030
0.000263
0.035847
(S)2
E
E2
−4.89836
8.43370
T
T2
−6.18626
2.51932
SE
a
Significant factors p < 0.05.
Although the best value of substrate concentration in the
kinetic resolution of DL-1 was 4 mg/mL (conversion as
response variable), it was desirable to increase it for the sake of
economy. Fortunately, the model shows that concentrations of
8.5 mg/mL still retain a high conversion (Figure 4).
Our data showed that, within that concentration range
(Figure 4 a and b), and with the use of 220 U of lipase,
conversions remained high. A further increase of concentration
up to 11 mg/mL is possible without significant conversion loss.
Raising the substrate concentration to 14 mg/mL (Figure 4c)
would require higher enzyme load (280 U) to obtain the same
conversion level, which would render the process less
Table 5. ANOVA of the quadratic model for conversion in
the kinetic resolution (X%)
a
sources of
variations sources
of variation
degrees
of
sum of
squares
mean
b
freedom squares F-test
p value
regression
residual
total
2892.52
93.42
9
8
321.39
11.67
27.52
0.000044
2985.93
17
a
Coefficient of determination (R2) = 0.9687; adj. R2 = 0.9335.
F0.05; 9; 8(Ftabulated) = 3.50
b
Figure 2. Response surface (left) and contour diagram (right) for conversion response (X%) to temperature (T) and enzyme activity (E). The value
of substrate concentration was fixed at 4 mg/mL.
1380
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384

Organic Process Research & Development
Article
Figure 3. Response surface (left) and contour diagram (right) for conversion response (X%) to substrate concentration (S) and enzyme activity (E).
The temperature was fixed at 40 °C.
Figure 4. Contour diagram for variable conversion response (X%) using 220 U of lipase. Values of substrate concentration were fixed at (a) 4 mg/
mL, (b) 8.5 mg/mL, and (c) 14 mg/mL.
economical. Although the productivity study pointed out to 57
°C as the temperature, we continued to employ 45 °C (the
optimal temperature by CCRD) in order to avoid enzyme
destabilization.
Thus, our model for kinetic resolution of DL-1 by Novozym
435 with vinyl acetate in hexane set as the optimum condition
(in the range) 45 °C, 4 mg/mL of substrate, and 222.8 U of
enzyme. The accuracy of the model was validated with three
replicates under the aforementioned optimum conditions. The
experimental conversion was 49.7 0.2% after 24 h of reaction
against a 52% predicted conversion. The eep was >98% (E >
200) in all samples collected during the 24-h reaction (Figure
5).
h in the original condition using hexane) in the kinetic
resolution of inositol DL-1.10
In fact, the effect of the optimization on the productivity and
reaction kinetics was dramatic (Figure 6). Despite the sharp
decay over time, the productivity of the optimized reaction
overcomes that of the nonoptimized one throughout the
reaction course (>20 h). Little variation of productivity is
shown in the latter condition. This explains the much higher
conversion in the optimized reaction.
Thus, through the experimental design strategy by CCRD, a
15-fold increase in productivity (0.006 mgproduct/mgenzyme/h for
an 8-h reaction) over that from the original protocol9 (0.0004
mgproduct/mgenzyme/h for a 112 h reaction). Besides boosting the
catalytic performance (higher conversion and shorter reaction
time) in the kinetic resolution of DL-1 by Novozym 435, the
present optimization fosters higher economy, as lesser
quantities of solvent and catalyst are required in the process.
The time course of this enzymatic transformation, under the
optimum conditions, was determined (Figure 5). Clearly, the
experimental design reduced the reaction time (48.3% after 112
Moreover, such high conversion (49.7
0.2%) enables the
synthesis of both the acylated product (ee > 99%) and the
remaining (unreacted) substrate (ee = 95%) in optically pure
form.
This condition could be reproduced in a 1-g scale, which
afforded high conversion (50%
1) and selectivities (eep =
99% and ees = 99%, E > 100).
REUSABILITY
■
Naturally, the reuse of an immobilized enzyme without
appreciable loss of enzyme activity is important for the
economic viability of enzymatic processes. We inquired
whether the designed conditions for the enzymatic reaction
under study would enable such reusability. The results are in
Figure 7.
Figure 5. Time course of the kinetic resolution of DL-1 in hexane
under optimum conditions: 45 °C, 4 mg/mL of substrate and 222.8 U
of enzyme activity.
The immobilized lipase was operated over 14 d (with 24-h
cycles), nearly without loss of conversion values up to the
1381
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384

Organic Process Research & Development
Article
Figure 6. Productivity and conversion evolution in the kinetic resolution of DL-1 in hexane under the optimum (45 °C, 4 mg/mL of substrate and
222.8 U of enzyme activity) and nonoptimized conditions.9.
EXPERIMENTAL SECTION
■
Substrate. The substrate ( )-1,3,6-tri-O-benzyl-myo-inosi-
tol, DL-1, was synthesized by literature procedures.24,25
Enzyme. Novozym 435 (Lipase B of Candida antarctica
immobilized on a macroporous acrilic resin, 2728 U/g) was
́
purchased from Novozymes Brasil (Araucaria-PR). The enzyme
activity was determined as the initial rate in esterification
reactions between oleic acid and ethanol at a molar ratio of 1:1,
temperature of 40 °C and enzyme concentration of 5 wt % in
relation to the substrates.26 One lipase activity unit (U) was
defined as the amount of enzyme necessary to consume 1 μmol
of oleic acid per minute at the established experimental
conditions previously presented. All enzymatic activity
determinations were replicated at least three times.
Experimental Conditions. The enzymatic reactions were
realized under magnetic stirring in closed thermostatized flasks
(water bath) containing DL-1, Novozym 435 and vinyl acetate
in hexane (Final volume 2.0 mL) (vide discussion for details).
After 24 h, the reactions were stopped by catalyst removal by
filtration. The assays were run in triplicate. The samples
containing product L-(−)-2 and D-(+)-1 had the volatiles
evaporated and the resulting material was subjected to HPLC
analysis of conversion and separation for ee determination (vide
infra).
Figure 7. Operational stability in the kinetic resolution of DL-1 by
Novozym 435 under the designed conditions: 4 mg/mL of substrate,
222.8 U of enzyme, and 45 °C of temperature.
seventh recycle (49% 0.7). From that point up to the 10th
recycle, acceptable conversions were still attained (40% 0.5).
Throughout the process 98% eep (E > 100) was obtained.
These results fully justify the use of immobilized lipase and
qualify the optimized procedure for gram-scale chemo-
enzymatic syntheses of myo-inositols.
For the 1-g scale (DL-1) reaction under the optimized
protocol (45 °C), 10.2 g of Novozym 435 and a 250-mL vinyl
acetate/hexane mixture (containing 37.5 mL of vinyl acetate)
were employed (see text).
CONCLUSION
■
The successful optimization of the kinetic resolution of myo-
inositol DL-1 by Novozym 435 enables a practical and
economical synthesis of both enantiomorphs D-(+)-1 and (by
deacylation) L-(−)-1, besides acetate L-(−)-2, in high ee. This
process promises to make different high-value myo-inositols and
other synthetic derivatives more available for cell biology
investigations and medicinal applications. In fact, by means of
CCRD, the optimization of this enzymatic transformation in
hexane with vinyl acetate led to a 15-fold increase of
productivity over the original protocol. Moreover, the
constructed mathematical model informed us that higher
substrate concentration (over 4 mg/mL) can be employed
without any significant decrease of conversion. Finally, these
results were explored in assays that demonstrated the
reusability of Novozym 435 in the kinetic resolution of myo-
inositol DL-1.
HPLC Analysis of Conversion of DL-1 to L-(−)-2.
Conversion analyses were done via HPLC on a Shimadzu-
C18 column (40 °C in a CTO-20A oven), eluted with an
acetonitrile/H2O (60:40) mixture (0.5 mL/min) by a
Shimadzu LC-20AT pump. A Shimadzu SPD-M20A variable-
wavelength UV/vis detector was employed, with the detection
set at 215 nm, and the Shimadzu LC solution software was used
for chromatogram integration. The samples to be analyzed were
filtered through a 0.22 μm PTFE filter. The retention times of
the substrate DL-1 and the acetate L-(−)-2 were 8 min and 13
min, respectively.
Determination of Enantiomeric Excesses (ee). Un-
reacted substrate, D-(+)-1, and monoacetylated product, L-
(−)-2, were separated by HPLC, under elution condition used
for the conversion determination (vide supra). Then, the
1382
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384

Organic Process Research & Development
Article
n
n−1
n
solvents of the obtained solutions were evaporated prior to
direct analysis (D-1) or further treatment (methanolysis) and
following analysis (L-2). In the case of L-(−)-2, the substance
was subjected to methanolysis reaction (MeOH/K2CO3) to
give triol L-(−)-1 prior to the HPLC analyses.5c
Chromatographic determination of the ee’s of both D-(+)-1
(ees) and L-(−)-1 (eep) were done on the same equipment
mentioned above, carrying a Chiralcel OD-H column (5 μm;
4.6 mm × 250 mm), eluted with a 7:3 hexane/2-propanol
mixture (0.6 mL/min). The retention times of these
enantiomorphs were 24.5 and 28.5 min, respectively. The
enantiomeric ratio (E) was calculated by using the equation of
Chen et al.27
Y = a +
a x +
aijxixj
∑
∑ ∑
0
i i
i=0
i=0 j=i+1
(1a)
where Y is the predicted response, i and j range from 1 to the
number of variables (n), a0 is the intercept term, ai values are
the linear coefficients, aij values are the quadratic coefficients,
and xi and xj are the levels of the independent variables.
All the experiments of the fractional design and the CCRD
were performed randomly, and the data were treated with the
aid of the software STATISTICA 7.0 (Statsoft Inc., Tulsa, OK,
USA).
Enzyme Reusability. In the assays of Novozym 435 reuse,
after each 24-h batch run (2 mL), the reaction mixture was
centrifuged. Then, the liquid phase (for chromatographic
analyses) was decanted and the solid catalyst was used in the
next run under the optimum condition. Such procedure was
repeated 13 times. In each run, a sample of 100 μL was taken
for determination of the enzyme activity. No decline in activity
was observed. When washing of the catalyst (with vinyl acetate/
hexane mixture) was carried out after each run, a significant loss
of activity occurred.
Experimental Design Strategy for Kinetic Resolution
of DL-1: Optimization Experiments. Five independent test
variables were chosen for the statistical experimental design:
substrate concentration (S1, mg/mL); acyl donor concentration
(S2, mg/mL), enzyme concentration (E, mg/mL) and temper-
ature (T, °C). A previous study by our group had shown
anhydrous hexane as the best solvent for this enzymatic
transformation.9,10
First, a fractional factorial design including 24−1 runs, with
three central points, was carried out to evaluate which variables
have a major effect on the kinetic resolution (see Table 6).
AUTHOR INFORMATION
■
Corresponding Author
Table 6. Experimental values and levels of the independent
variables of fractional factorial design for the kinetic
resolution of ( )-1 by Novozym 435
*(E.A.M.) Telephone: +55 21 3346 1306. Fax: +55 21 2562
7360. E-mail: evelin@ufrj.br or biorecados@yahoo.com.br;
(D.M.G.F.) Telphone: +55 21 2562 7360. E-mail: freire@
peq.coppe.ufrj.br.
range and levels
Notes
test variables
−1
0
1
The authors declare no competing financial interest.
substrate (mg/mL), S1
acyl donor (mg/mL), S2
enzyme (U), E
2
8
14
186.8
54.56
30
373.6
81.84
40
560.4
136.4
50
ACKNOWLEDGMENTS
■
temperature (°C), T
FAPERJ, CNPq, and CAPES for funding and/or fellowships;
́
LAMAR/NPPN and Central Analitica/NPPN for analytical
data.
The preliminary fractional factorial design allowed for the
selection of the statistically significant variables for conversion
of the reaction. After selection of the variables, a CCRD with 23
trials, including four replicates at the central point and six axial
points, were employed to obtain a second-order model for the
prediction of conversion (dependent variable) as function of
the studied variables (independent variables). The real variable
values and coded levels were varied according to the
experimental design of the kinetic resolution under study are
given in Table 7.
REFERENCES
■
(1) For reviews on biology and chemistry of inositols: (a) Conway, S.
J.; Miller, G. J. Nat. Prod. Rep. 2007, 24, 687−707. (b) Best, M. D;
Zhang, H.; Prestwich, G. D. Nat. Prod. Rep. 2010, 27, 1403−1430.
(c) Potter, B. V. L; Lampe, D. Angew. Chem., Int. Ed. Engl. 1995, 34,
1933−1972.
(2) For reviews more focused on synthesis: (a) Duchek, J.; Adams,
D. R.; Hudlicky, T. Chem. Rev. 2011, 111, 4223−4258. (b) Kilbas, B.;
Balci, M. Tetrahedron 2011, 67, 2355−2389. (c) Sureshan, K. M.;
Shashidhar, M. S.; Praveen, T.; Das., T. Chem. Rev. 2003, 103, 4477−
4504. (d) Billington, D. C. Chem. Soc. Rev. 1989, 18, 83−122.
(3) Gotor-Fenandez, V.; Brieva, R.; Gotor, V. J. Mol. Catal. B: Enzym.
2006, 40, 111−120.
(4) Representative examples of lipase-catalyzed reactions of myo-
inositol derivatives: (a) Ling, L.; Watanabe, Y.; Akiyama, T.; Ozaki, S.
Tetrahedron Lett. 1992, 33, 1911−1914. (b) Andersch, P.; Schneider,
M. P. Tetrahedron: Asymmetry 1993, 4, 2135−2138.
(5) For works dealing with with O-benzylated substrates, as in this
study: (a) Laumen, K.; Ghisalba, O. Biosci., Biotechnol., Biochem. 1999,
63, 1374−1377. (b) Ling, L.; Ozaki, S. Bull. Chem. Soc. Jpn. 1995, 68,
1200−1205. (c) Laumen, K.; Ghisalba, O. Biosci., Biotechnol. Biochem.
1994, 58, 2046−2049.
The data obtained from the CCRD were used to fit an
empirical quadratic polynomial model related to the response
by a multiple regression procedure.28−30 For three factors, the
model takes the form of eq 1a.
Table 7. Experimental values and levels of the independent
variables of CCRD for the kinetic resolution of ( )-1 by
Novozym 435
range and levels
test variables
−1.68
−1
0
1
1.68
15
(6) For cases of catalysis by esterase: (a) Baudin, G.; Glanzer, B. I.;
Swaminathan, K. S.; Vasella, A. Helv. Chim. Acta 1988, 71, 1367−1378.
(b) Liu, Y.; Chen, C. Tetrahedron Lett. 1989, 30, 1617−1620. (c) Gou,
D.; Chen, C. Tetrahedron Lett. 1992, 33, 721−724.
substrate (mg/mL), S1
enzyme (U), E
2
5
8.5
11.5
300.1
50
80.5
23
136.4
30
218.2
40
355.8
57
temperature (°C), T
1383
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384

Organic Process Research & Development
Article
(7) Cunha, A. G.; da Silva, A. A. T.; da Silva, A. J.; Tinoco, L. W.;
Almeida, R. V.; de Alencastro, R. B.; Simas, A. B. C.; Freire, D. M. G.
Tetrahedron: Asymmetry 2010, 21, 2899−2903.
(8) Simas, A. B. C.; da Silva, A. A. T.; Cunha, A. G.; Assumpca̧ o, R.
̃
S.; Hoelz, L. V. B.; Neves, B. C.; de Alencastro, R. B.; Freire, D. M. G.
J. Mol. Catal. B: Enzym. 2011, 70, 32−40.
(9) Manoel, E. A.; Pais, K. C.; Cunha, A. G.; Coelho, M. A. Z.; Freire,
D. M. G.; Simas, A. B. C. Tetrahedron: Asymmetry 2012, 23, 47−52.
(10) Manoel, E. A. M.S. Thesis, School of Chemistry, Biochemical
Engineering Departament, Universidade Federal do Rio de Janeiro,
Brazil, 2012.
(11) (a) Mills, S. J.; Liu, C.; Potter, B. V. L. Carbohydr. Res. 2002,
337, 1795−1801. (b) Desai, T.; Gigg, J.; Gigg, R.; Martin-Zamora, E.
Carbohydr. Res. 1994, 262, 59−77. (c) Watanabe, Y.; Nakahira, H.;
Bunya, M.; Ozaki, S. Tetrahedron Lett. 1987, 28, 4179−4180.
(12) Torres, R.; Claudia, O.; Pessela, B. C. C.; Palomo, J. M.; Mateo,
C.; Guisan, J. M.; Fernandez-Lafuente, R. Enzyme Microb. Technol.
2006, 39, 167−171.
(13) Anderson, E. M.; Larson, K. M.; Kirk, O. Biocatal. Biotrans.
1998, 16, 181−204.
(14) Barbosa, O.; Ortiz, C.; Torres, R.; Fernandez-Lafuente, R. J. Mol.
Catal. B: Enzym. 2011, 71, 124−132.
(15) Gotor-Fernan
2006, 348, 797−812.
́
dez, V.; Busto, E.; Gotor, V. Adv. Synth. Catal.
(16) Shimada, Y.; Watanabe, Y.; Sugihara, A.; Tominaga, Y. J. Mol.
Catal. B: Enzym. 2002, 17, 133−142.
(17) Uppenberg, J.; Hansen, M. T.; Patkar, S.; Jones, T. A. Structure
1994, 2, 293−308.
(18) Kirk, O.; Christensen, W. Org. Process Res. Dev. 2002, 6, 446−
51.
(19) Noel, M.; Lozano, P.; Vaultier, M.; Iborra, J. L. Biotechnol. Lett.
2004, 26, 301−306.
(20) Ottoson, J.; Fransson, L.; King, J. W.; Hult, K. Biochim. Biophys.
Acta 2002, 1594, 325−334.
(21) Ottoson, J.; Rotticci-Mulder, J. C.; Rotticci, D.; Hult, K. Protein
Sci. 2002, 10, 1769−1774.
(22) Arroyo, M.; Sanchez-Montero, J. M.; Sinisterr, J. V. Enzyme
Microb. Technol. 1999, 24, 3−12.
(23) Barbosa, O.; Ortiz, C.; Torres, R.; Fernandez-Lafuente, R. J. Mol.
Catal. B: Enzym. 2011, 71, 124−132.
(24) Simas, A. B. C.; Pais, K. C.; da Silva, A. A. T. J. Org. Chem. 2003,
68, 5426−5428.
(25) Simas, A. B. C.; da Silva, A. A. T.; Dos Santos Filho, T. J.;
Barroso, P. T. W. Tetrahedron Lett. 2009, 50, 2744−2746.
(26) de Oliveira, D.; Feihrmann, A. C.; Dariva, C.; Cunha, A. G.;
Bevilaqua, J. V.; Destain, J.; Oliveira, J. V.; Freire, D. M. G. J. Mol.
Catal. B: Enzym. 2006, 39, 117−123.
(27) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem.
Soc. 1982, 104, 7294−7299.
(28) Khuri, A. I.; Cornell, J. A. Response Surfaces: Design and Analyses;
Dekker: New York, 1987.
(29) Montgomery, D. C. In Design and Analysis of Experiments, 3rd
ed.; Wiley: New York, 1991.
(30) Colla, E.; Pereira, A. B.; Hernalsteens, S.; Maugeri, F.;
Rodrigues, M. I. Food Bioprocess. Technol. 2010, 3, 265−275.
NOTE ADDED AFTER ASAP PUBLICATION
■
This paper was published on the Web on August 2, 2012.
Equation 1 and Tables 3 and 7 have been updated. The
corrected version was reposted on August 7, 2012.
1384
dx.doi.org/10.1021/op300063f | Org. Process Res. Dev. 2012, 16, 1378−1384