Optimization and kinetic modeling of lipase mediated enantioselective kinetic resolution of ( ± )-2-octanol

Chiral 2-octanol is one of the key intermediates for preparation of liquid crystal materials, as well as many optically active pharmaceuticals. Lipase catalyzed kinetic resolution has proved to be an efficient technique for synthesis of enantiomerically enriched compounds. In the present study, optimization and kinetic modeling of kinetic resolution of (±)-2-octanol was done by using vinyl acetate as an acyl donor in n-heptane as a solvent. Response surface methodology (RSM) and four-factor-five-level Centre Composite Rotatable Design (CCRD) were employed to evaluate the effect of various parameters such as speed of agitation, enzyme loading, temperature and acyl donor/alcohol molar ratio on conversion, enantiomeric excess (ee), enantioselectivity and initial rate of reaction. Acylation of 2octanol with vinyl acetate catalyzed by Novozyme 435 follows the ternary complex mechanism (ordered bi-bi mechanism) with inhibition by 2octanol.


INTRODUCTION
Enzymatic catalysis in non-aqueous media has been greatly pursued these days for the synthesis of a wide variety of pharmaceuticals, agrochemicals, perfumes, flavors and other fine-chemicals [1][2][3][4].In this regard, our group has contributed extensively to mechanistic studies, kinetic modeling and separation of enantiomers, covering several industrially relevant classes of reactions such as epoxidation/oxidation [5,6], hydrolysis [7], esterification [8][9][10], transesterification [11][12][13], amidation [14] and hydrazinolysis [15].The synergism with microwave irradiation in immobilized lipase catalysis [16][17][18][19][20] and scope of non-aqueous systems in pharmaceutical industries [21] have been embraced.Optimization of process parameters by using statistical methods has been reported in a number of cases in literature and the current investigation in an effort in that direction.
Kinetic resolution of chiral compounds using enzymes especially lipases has proven to be an effective technique vis-a-vis chemical methods.The main consideration for adding biotransformation in a synthetic route is the regio-and stereo-control that can be achieved elegantly using enzyme-catalyzed step(s) [4].Thus, chemo-enzymatic processes will see commercial utility in future.For instance, chiral aliphatic alcohols, which are important active pharmaceutical intermediates (API), have been obtained through lipase catalyzed kinetic resolution of corresponding racemic mixtures via esterification, transesterification or ester hydrolysis [21].There are a number of ways to resolve the racemic mixtures by using enzymatic catalysis: dynamic kinetic resolution (DKR) with racemization catalysts [22], combination of DKR with double kinetic resolution [23], deracemisation [1], and sequential kinetic resolution [24].Different lipases have been used for the kinetic resolution of aliphatic alcohols [23,[25][26][27][28][29][30].Various immobilization techniques for lipase immobilization have been reported; for instance, hexagonal mesoporous silica (HMS) [12], magnetic nanoparticles, Diaion HP20, ultrastable-Y molecular sieve [27], SBA 15 [29], and Sol-gel method [31].
Process optimization has a great relevance in complex reaction and has been done by two ways: one-factor-ata-time method and statistical analysis such as Response Surface Methodology (RSM).RSM is a collection of statistical and mathematical techniques useful for developing, improving, and optimizing processes in which a response of interest is influenced by several variables and the objective is to optimize this response [32].To avoid the disadvantages of the one-factor-at-a-time method since it does not illustrate interaction effect among various factors and gives only local optima of the reaction, in this work we have used the RSM for process optimization.The Centre Composite Rotatable Design of RSM has been previously been successfully applied in food technology [33,34], microbiology [35], biotechnological [36][37][38][39][40][41] and chemical processes [42].To the best of our knowledge, there is a dearth of literature on RSM for kinetic resolution of chiral compounds using enzymatic catalysis [40][41][42][43][44].
In the present study, Candida antartica lipase B, Thermomyces lanuginosus lipase and Rhizomucor meihei lipase, were employed for the kinetic resolution of (±)-2-octanol by using vinyl acetate as an acylating agent.Optimization of reaction parameters has been done by RSM and CCRD using four factors, each at five variables.

Experimental Setup
The experimental setup consisted of a 3 cm internal diameter (ID), fully baffled mechanically agitated reactor of 50 cm 3 capacity, which was equipped with four equispaced baffles and 1 cm diameter four bladed-pitchedturbine impeller.The entire reactor assembly was immersed in a thermostatic water bath which was maintained at a desired temperature with an accuracy of ±1˚C.

Kinetic Resolution of (R,S)-2-Octanol by Immobilized Lipase
The resolution was performed in the above reactor containing (±)-2-octanol, catalyst and solvent.When the set temperature was reached, vinyl acetate was added in the reactor, and agitation started.Samples were withdrawn periodically at regular time intervals, and the reso-lution process was monitored by GC.The total reaction mixture volume was 25 cm 3 which was made up with n-heptane as a solvent.The total reaction time was 6 h.

Determination of Enantiomeric Excess (ee) and Enantioselectivity (E)
Clear liquid samples were withdrawn periodically from the reaction mass and analyzed by using Ceres 800 plus GC instrument equipped with flame ionization detector (FID) and β-Dex 120 (30 m  0.25 mm  0.25 µm) chiral capillary column.The analytical conditions were: injector temperature 220˚C; FID temperature 220˚C; oven temperature held at 65˚C for 30 min, then increased at 10˚C•min −1 to a final temperature of 130˚C, which was thereafter maintained for 10 min.The enantioselectivity (E) was calculated from the enantiomeric excess of the substrate (ee s %) at a certain conversion (c, %) based on the following equations. where, and where, A (R) and A (S) denote (R)-2-octanol and (S)-2-octanol, respectively.

Design of Experiments and Statistical Analysis
RSM was used to optimize the process of resolution of (±)-2-octanol and to study the effect of different process variables on reaction along with the interactions among them.The experimental design applied to this study was CCRD (four factors, each at five levels).Compared with one-factor-at-a-time design, which has been adopted most often in the literature, the combination of RSM and four-factor-three-level CCRD employed in this study allowed us to reduce the number of experiments and time.The independent variables were: speed of agitation, catalyst loading, reaction temperature, and ester to alcohol molar ratio.Whereas the responses (dependent variables) chosen were 1) conversion of (±)-2-octanol; 2) ee of remaining alcohol; 3) E of the enzyme and 4) initial rate of reaction.Table 1 shows the independent variables and their levels.The responses were then analyzed using numerical tools provided by Design Expert, Version 6.0.10 (Stat Ease, Minneapolis, MN, USA).The second order polynomial coefficients were calculated and analysis of variance (ANOVA) was conducted by using analytical tools of Design Expert.Contour and response surface plots were obtained after analysis of each response.After each response had been analyzed, multiple response optimizations were done by numerical tools provided by the Design Expert.Separate experiments at the optimum process conditions were performed for validation of the response models.

RESULTS AND DISCUSSION
Lipase catalyzed kinetic resolution of (±)-2-octanol with vinyl acetate as an acyl donor in n-heptane as a solvent produces ester and acetaldehyde is given in Scheme 1.

Effect of Different Catalysts
The activity and selectivity of Novozyme 435, Lipozyme RM IM and Lipozyme TL IM were evaluated towards the acylation of (±)-2-octanol under otherwise similar conditions.Figure 1 shows the average conversion of three experiments at the end of 6 h for each enzyme.In the case of Novozyme 435, the conversion was 41.8% compared with 18.9% for Lipozyme RM IM and 9.6% for Lipozyme TL IM.The higher activity of Novozyme 435 is probably due to its stability in the presence of acetaldehyde which is liberated during the reaction.Novozyme 435 was selected for all further experiments as it gave highest conversion as compared to other catalysts.

Process Optimization
The major objective of this work was the development and evaluation of a statistical approach to better understand the relationship between the independent and dependent variables of a lipase catalyzed acylation of (±)-2octanol.The experiments were performed as per design of experiments data.The order in which reactions were performed was randomized to minimize errors due to possible systematic trends in the variables.Six experiments were carried out at the center point, coded as "0", to minimize experimental error.Different models (linear, two factor interaction, quadratic and cubic) were tested for fitting experimental data.Based on p-value, the quadratic model was used for conversion, two-factor interaction for enantioselectivity and linear for initial rate (Table 2).The mean of the enantiomeric excess (ee) was taken into account as the values obtained were much closer and no model can be fitted to the experimental data.Correlation regression coefficients were obtained for all the response models indicating that second order polynomial model fitted well to the experimental data and was adequate to represent the relationship between the responses and significant variables with very small p value and a satisfactory coefficient of determination.The second order polynomial equations for conversion, enantioselectivity and initial rate of reaction are as follows, Enantioselectivity 84. 13 ANOVA was performed for the model fitted to the experimental data.The mean squares, F values and p values for the response surface models are given in Table 3.Low p-value indicates that the model term is significantly affecting the process.If it is a single order term, it indicates that process parameter is significantly affecting whereas if it is second order term, it shows that the interaction between the process parameters is significant.Temperature, catalyst loading and mole ratio are the significantly affecting parameters for conversion; whereas for the initial rate, catalyst loading and mole ratio are the significantly affecting parameters.Temperature and mole ratio also show interaction amongst them to affect enantioselectivity.For initial rate, there were no interacting parameters as the model fitted to these responses was a linear model.temperature; however, as the mole ratio was increased, there was a decrease in enantioselectivity at higher temperatures.Figure 3 shows the response surface plot for conversion, as a function of catalyst loading and temperature.Catalyst loading and temperature were investigated in the range of 1:1 -5:1 and 10˚C -70˚C, respectively.As the temperature and catalyst loading were increased, the conversion increased.Figure 3 shows that at higher temperature, as the catalyst loading increased, conversion was increased.
The lack of fit test is a measure of failure of a model to represent data in the experimental domain at which points were not included in the regression [45].The analysis of lack of fit was performed on all the dependent variables and it was insignificant for all the models.Correlation regression coefficients greater than 0.9 for conversion showed that models gave satisfactory predicttion for experimental data; whereas for enantioselectivity and initial rate correlation regression coefficients are less than 0.9 which showed no model could gave satisfactory prediction for experimental data (Table 3).
A plot of distribution of residuals values, defined as the difference between calculated and observed values over the predicted values, shows that the quality of fit is good because the distribution does not follow the trend with respect to the predicted values.An optimum resolution reaction for (±)-2-octanol represents conditions which would give high enantiomeric excess, high enantioselectivity, higher initial rate and 50 % conversion.Numerical tools provided by Design Expert were used to find out the optimum conditions.The optimum reaction conditions thus obtained for the desired isomer, (R)-2-octanol, were, mole ratio of vinyl acetate: (±)-2-octanol of 4:1, temperature of 25˚C, 0.05 g of catalyst loading and 400 rpm as speed of agitation with conversion: 43.1%, ee of remaining enantiomer: 71.8%, enantioselectivity: 203 and Initial rate: 0.0060 M•min −1 .
The second order polynomial equations were used to generate surface response plots and then finally to arrive at the optimum reaction conditions to maximize conversion and enantiomeric excess.Response surface and contour plots were generated for interacting parameters for each response.Figure 2 shows the surface response plot for enantioselectivity, as a function of the interacting parameters, i.e. temperature and mole ratio.Temperature and mole ratio were investigated in the range of 10˚C -70˚C and 1:1 -5:1, respectively, at a catalyst loading of 0.06 g and speed of agitation at 300 rpm.At a molar ratio of 2:1, enantioselectivity increased with an increase in

Model Validation
The validity of the predicted model was examined by carrying out additional independent experiments at the suggested optimum reaction conditions and three centre points.Table 4 shows the predicted and observed values for the responses at optimum conditions for resolution of (±)-2-octanol using vinyl acetate as an acyl donor.The experimental values were averages of three values and were close to the predicted values indicating that the second order polynomial models generated were acceptable.

Operational Stability of Enzyme
The operational stability study was conducted under the optimum reaction conditions obtained from the RSM.After each run, the enzyme was allowed to settle and the supernatant liquid was removed.Then, n-heptane was added to the solid particles, and the mixture was shaken to wash away the remaining substrate and product.The washing was carried out thrice.Then the enzyme was filtered, air dried and used for the next run.To investigate the effect of substrate on the stability of the enzyme, the reusability study was carried out under otherwise similar conditions.It was found that there was a decrease in conversion from 42% to 36% after third reuse (Figure 4).There was no make-up catalyst added and there was loss of catalyst of 3% -4% during handling.Thus, the reusability of the enzyme also confirmed that acetaldehyde did not deactivate the enzyme.

Kinetic Modeling
The effect of concentration of both the reactants on the rate of reaction was investigated systematically over a wide range.For the determination of initial rates, two sets of experiments were conducted by using 0.05 g Novozyme 435 with appropriate quantities of (±)-2-octanol and vinyl acetate and the total volume was made up to 25 cm 3 with n-heptane.In one set of experiments, (±)-2octanol amount was varied from 0.0075 -0.06 mol at a fixed quantity of vinyl acetate (0.06 mol) and in another set, the amount of vinyl acetate was varied from 0.015 -0.06 mol at a fixed quantity of (±)-2-octanol (0.015 mol).The conversions were quantified by using synthetic mixtures.The initial rates were determined from the quantified data.
When the concentration of (±)-2-octanol (A) was increased, by keeping the concentration of vinyl acetate (B) constant, the initial rate of reaction (r 0 ) increased proportionally and reached a maximum at a critical concentration.A subsequent increase in 2-octanol concentration immediately led to a decrease in the initial rate.Increasing concentrations of vinyl acetate under otherwise similar conditions increased the rate and conversion.There was no evidence of inhibition by vinyl acetate (B) at all the concentration tested.The Lineweaver-Burk plot of 1/r 0 versus 1/[A] showed that at lower concentration of (±)-2-octanol, there was an increase in initial rates (Figure 5).Increase in the (±)-2-octanol concentration resulted in decrease in initial rates.It suggested that (±)-2octanol acts as a dead-end inhibitor of enzyme whereas vinyl acetate does not inhibit the reaction.
In the case of lipase-catalyzed reactions, it has been established that the lipase first forms an acyl-enzyme complex with the acyl donor, ruling out the random mechanism [46].As a consequence, it can only be the ordered bi-bi mechanism.Considering it as bi-bi reaction, two models were proposed, namely, the ternary complex mechanism with inhibition by (±)-2-octanol, and the ping-pong bi-bi mechanism with inhibition by (±)-2octanol.The synthesis of isoamyl acetate by transesterification of ethyl acetate with isoamyl alcohol in n-hexane using lipozyme for which they had found a ping-pong bi-bi mechanism with competitive inhibition by substrates and product ethanol [47].Since there was no reverse reaction in the current case, a possible inhibition by vinyl acetate at higher concentration was also considered whereby ping-pong bi-bi mechanism with inhibition by both (±)-2-octanol and vinyl acetate was also considered.The rate equation for ping-pong bi-bi mechanism with inhibition by (±)-2-octanol, for initial conditions [48], is as follows: The rate equation for ping-pong bi-bi mechanism with inhibition by (±)-2-octanol and vinyl acetate is as follows: The rate equation for the ternary complex mechanism, for initial conditions, is as follows: where, r is the rate of reaction, max , maximum rate of reaction, [A], initial concentration of (±)-2-octanol, [B], initial concentration of vinyl acetate, mA r K , Michaelis constant for (±)-2-octanol, mB K , Michaelis constant for vinyl acetate, iA K , inhibition constant for (±)-2-octanol, and iB K is the inhibition constant for vinyl acetate.Initial rates were calculated from the linear portion of the concentration-time profiles and the kinetic constants were obtained by non-linear regression analysis for the above models (Table 5).It is observed that the sum of the squared residuals was minimum in the case of ternary complex model with inhibition by (±)-2-octanol above certain concentration.In the other two cases, some of the estimated parameters were found to be negative and unrealistic.It is thus concluded that the reaction sequence follows the ternary complex mechanism with inhibition by (±)-2-octanol.The sequence is as follows: According to the ordered bi-bi mechanism, the acyl donor (B) first binds with the enzyme and forms an acylenzyme complex (EB).The second reactant (A) then combines with (EB) to form ternary complex EBA.This ternary complex then isomerizes to another ternary complex, which releases the first product vinyl alcohol.This vinyl alcohol is highly unstable and therefore it irreversibly tautomerizes to acetaldehyde and the binary complex of (±)-2-octanol and enzyme which subsequently releases (±)-2-octyl acetate.However, at high concentrations of (±)-2-octanol the dead-end binary complex between (±)-2-octanol and enzyme is formed instead of vinyl acetate and enzyme.The reaction mechanism may be depicted in Scheme 2.Where E, enzyme; A, (±)-2-octanol; B, vinyl acetate; EA, enzyme-(±)-2-octanol dead-end complex; Q, acetaldehyde; P, (±)-2-octyl acetate; EBA, ternary complex; and EPQ is the isomer of EBA.The theoretical (simulated) initial rates were calculated by using the parameters in Table 5 for the ternary model and are compared against the experimental values for different (±)-2-octanol concentrations (Figure 6).There is an excellent match between theory and experiment, proving the validity of the ternary model.tion conditions required to obtain well-defined amount of acetate, enantiomeric excess of remaining alcohol, enantioselectivity of enzyme and initial rate of reaction.These models are useful to determine the optimum operating conditions for the resolution reaction using the minimal number of experiments with the consequent economical benefit.The analysis of the kinetic data showed that the acylation of (±)-2-octanol with vinyl acetate catalyzed by Novozyme 435 follows the ternary complex mechanism (ordered bi-bi mechanism) with 2-octanol inhibition providing support for one of the two proposed mechanisms.The optimum reaction conditions thus obtained for the desired isomer, (R)-2-octanol, were, mole ratio of vinyl acetate: (±)-2-octanol of 4:1, temperature of 25˚C, 0.05 g of catalyst loading and 400 rpm as speed of agitation

CONCLUSION
In the present study, three commercial lipases, Candida antartica lipase B (Novozyme 435), Thermomyces lanuginosus and Rhizomucor meihei, were employed for the kinetic resolution of (±)-2-octanol by using vinyl acetate as an acylating agent.The process of synthesis of (R)-2-octyl acetate using immobilized lipase, Novozyme 435 was optimized applying RSM with CCRD.Second order polynomial equations have been obtained for the conversion of alcohol, enantioselectivity of enzyme and initial rate of reaction.It was possible to predict the reac-OPEN ACCESS with conversion, 43.1%; ee, 71.8%; enantioselectivity, 203; and Initial rate, 0.0060 M•min −1 .

Figure 2 .
Figure 2. Effect of temperature, mole ratio and their mutual interaction on enantioselectivity.

Figure 3 .
Figure 3.Effect of catalyst loading and temperature on conversion (%).

Scheme 2 .
Scheme 2. Ternary complex mechanism with inhibition by A.

Table 1 .
Independent variables and their levels.

Table 2 .
ANOVA table for response variables a .

Table 3 .
ANOVA for Response Surface Models a .

Table 4 .
Predicted and observed values for the response variables at optimum conditions.

Table 5 .
Kinetic parameters for kinetic resolution of 2-octanol a .