Solvent Effects on Equilibrium Position and Initial Rate of Lipase-catalyzed Esterification Reactions in Organic Solvents: Experimental Results and Prediction Capabilities

The solvent effect on the equilibrium position and the initial rate of esterification of 1-hexanol with acetic acid catalyzed by a lipase has been experimentally investigated. A variety of non-polar and polar solvents have been considered and the results obtained indicate that the solvent effect on the equilibrium conversion is very important compared to that for transesterification reactions. A theoretically sound methodology using the group-contribution UNIFAC model for the prediction of solvent effects on the equilibrium position of enzymatic reactions is presented and it is applied to the reaction of 1-hexanol with acetic acid as well as to a similar reaction from the literature. The results obtained are better than those from empirical methods proposed in the literature such as correlations with the octanol-water partition coefficient of the solvent, as well as the solubility of water in the solvent. Moreover, the proposed methodology can be used for the determination of the equilibrium constant of the reaction. For the prediction of the solvent effect on the initial rate of enzymatic reactions it is found that it is more accurately determined using the product of the activities of the reactants, which can be predicted by the UNIFAC model, than the octanol-water partition coefficient of the solvent or the solubility of water in the solvent.     

Enzymatic Production of Alkyl Esters Through Lipase-Catalyzed Transesterification Reactions in Organic Solvents: Solvent Effects and Prediction Capabilities of Equilibrium Conversions

The solvent effect on the equilibrium position of the transesterification reaction of hexanol with ethyl acetate catalyzed by a lipase has been investigated in a variety of non-polar and polar solvents - and binary mixtures. The results obtained indicate that the solvent effect on the equilibrium conversion is very small as compared to that for the direct esterification reactions.

Equilibrium conversions were then predicted using the equilibrium constant for the reaction obtained from Gibbs free energy of formation information for reactants and products in combination with the UNIFAC activity coefficient model. A solvent independent equilibrium conversion was obtained, which was in good agreement with the observed average value for all solvents. This indicates that UNIFAC provides satisfactory estimates of the activity coefficients but its group contribution structure does not allow the prediction of the small differences in conversion among the solvents examined.

Finally plots of these conversions versus the solvent octanol/water partition coefficient or the solubility of water in the solvent, that provide the correct trend in direct esterification reactions, did not achieve the same for transesterification.     

The effect of organic solvents on the equilibrium position of enzymatic acylglycerol synthesis

The effect of organic solvents on the equilibrium position of lipase-catalyzed esterification of glycerol and decanoic acid has been investigated. The reaction is carried out in an aqueous-organic two-phase system. In polar solvents, high mole fractions of monoacylglycerol and low mole fractions of triacylglycerol and measured, while in nonpolar solvents, the measured differences in the mole fractions of monodi-, and triacylglycerols are less. There is a good correlation between the ester mole fractions at equilibrium and the log P of the solvent (partition coefficient in n-octanolwater), however, only if the group of tertiary alcohols is excluded. In the plot of the easter mole fractions as a function of the logarithm of hte solubility of water in the organic solvent, the tertiary alcohols can be included; however, in this case other deviations appear.For the prediction of the effect of organic solvents on the ester mole fractions at reaction equilibrium in nondilute reaction systems with a water activity below 1, the program TREP (Two-phase Reaction Equilibrium Prediction) is developed, which is based on the UNIFAC group contribution method. With this model the equilibrium data are essentially predicted from basic thermodynamic data. The required equilibrium constants are estimated from experiments without an organic solvent in the reaction medium. The mole fractions calculated by TREP show the same trends as the experimentally measured mole fractions; however, some variation is observed in the absolute values. These deviations may be due to inaccuracies in the UNIFAC group contribution method. TREP is found to be a correct method to predict within some limits the ester mole fractions at equilibrium for all mixtures of solvents, substrates, and products. The production of monoester can be enhanced in reaction system with a sufficient high concentration of a polar solvent. In experiments with a triglymeto-decanoic acid ratio of 5, almost no di-and triesters can be detected at equilibrium. (c) 1993 John Wiley & Sons, Inc.     

Solvent selection for biocatalysis in mainly organic systems: predictions of effects on equilibrium position

Predictions may be made for the influence of solvent choice on the equilibrium position of biocatalyzed reactions, based on data for the liquid-liquid distribution of the reactants. The most reliable predictions are probably for dilute systems, based on partition coefficients or correlations derived from them. The effective equilibrium constant for esterification reactions is predicted to alter by more than four orders of magnitude on changing between different water-immiscible solvents. The equilibrium constant correlates well with the solubility of water in the solvent, and is most favorable for synthesis in the least polar solvents (aliphatic hydrocarbons). Similar effects seem to apply for other reactions, including oxidation of alcohols and hydrolysis of chlorides. Predictions can be made for nondilute systems using the UNIFAC system of group contributions, but the reliability of these is more questionable.     

Combining solvent engineering and thermodynamic modeling to enhance selectivity during monoglyceride synthesis by lipase-catalyzed esterification

Monoglyceride synthesis by Rhyzomucor miehei lipase was investigated via direct esterification between glycerol (adsorbed onto silica gel) and oleic acid in organic solvents. The main difficulty is to avoid the unwanted production of di- and tri-glycerides. It was demonstrated that an increase in solvent polarity, using mixtures of n-hexane and 2-methyl-2-butanol (2M2B), improves drastically the selectivity toward monoglyceride formation. In pure n-hexane, the monoglyceride represents only 6 molar % of the total products at the thermodynamic equilibrium (34 and 60% for di- and tri-glyceride respectively). Use of an equivolume mixture of n-hexane/2M2B enables a product mixture to be obtained containing 94% of monoglyceride at equilibrium (2.4 and 0% for di- and tri-glyceride respectively). This positive effect is counterbalanced by a decrease both in initial velocities and in substrate conversion at thermodynamic equilibrium.A modeling, able to predict the three thermodynamic equilibria governing the 3 consecutive reactions, based on activity coefficient calculations using the UNIFAC model, is proposed. It takes into account both the partition of water between solvent and immobilized catalyst, and the partition of glycerol between solvent and silica gel. A good correlation with experimental data obtained in n-hexane/2M2B mixtures was observed.     

Solvent effects on biocatalysis in organic systems: equilibrium position and rates of lipase catalyzed esterification

Porcine pancreatic lipase immobilized on celite particles has been employed as a catalyst for the esterification of dodecanol and decanoic acid in a predominantly organic system. Solvent influence on the equilibrium position and on the catalyst activity has been studied using 20 solvents, including aliphatic and aromatic hydrocarbons, ethers, ketones, nitro- and halogenated hydrocarbons, and esters. The equilibrium constant for esterification correlates well with the solubility of water in the organic solvent, which in turn shows a good relationship with a function of Guttman's donor number and the electron pair acceptance index number of the solvent. This may be rationalized in terms of the requirements for solvation of water and of the reactants. The catalyst activity, measured as the initial rate of the esterification reaction, is best correlated as a function of both n-octanol-water partition coefficient (log P) and either the electron pair acceptance index or the polarizability.     

BESSICC, a COSMO-RS based tool for in silico solvent screening of biocatalyzed reactions

Many enzymatic reactions are near-equilibrium reactions. This often limits final yield and hence application of biocatalyzed processes in the industrial production. The most widely applied strategy to overcome this issue is solvent selection. It must be underlined that measuring the equilibrium position experimentally is a difficult and time-consuming procedure and any tool for predicting the solvent effect on the reaction equilibrium can be very valuable. The present work reports on the development of BESSICC, an algorithm to calculate the effect of medium composition on biocatalyzed reactions equilibrium. It is based on COSMO-RS calculation of activity coefficients of all the species in the reaction mixture and minimization of Gibbs free energy of the reaction. Starting from one single experimental measurement of the equilibrium position for a given biocatalyzed reaction it can predict the yield of the reaction in any other solvent or solvent mixture. Predictions are accurate, the errors of prediction are in average below 25% for the esterification of dodecanoic acid with menthol and below 65% for esterification of 1-dodecanoic acid with 1-dodecanol. The best predictions show an error well below 5%.     

Thermodynamically based solvent design for enzymatic saccharide acylation with hydroxycinnamic acids in non-conventional media

Enzyme-catalyzed synthesis has been widely studied with lipases (EC 3.1.1.3), but feruloyl esterases (FAEs; EC 3.1.1.73) may provide advantages such as higher substrate affinity and regioselectivity in the synthesis of hydroxycinnamate saccharide esters. These compounds are interesting because of their amphiphilicity and antioxidative potential. Synthetic reactions using mono- or disaccharides as one of the substrates may moreover direct new routes for biomass upgrading in the biorefinery. The paper reviews the available data for enzymatic hydroxycinnamate saccharide ester synthesis in organic solvent systems as well as other enzymatic hydroxycinnamate acylations in ionic liquid systems. The choice of solvent system is highly decisive for enzyme stability, selectivity, and reaction yields in these synthesis reactions. To increase the understanding of the reaction environment and to facilitate solvent screening as a crucial part of the reaction design, the review explores the use of activity coefficient models for describing these systems and - more importantly - the use of group contribution model UNIFAC and quantum chemistry based COSMO-RS for thermodynamic predictions and preliminary solvent screening. Surfactant-free microemulsions of a hydrocarbon, a polar alcohol, and water are interesting solvent systems because they accommodate different substrate and product solubilities and maintain enzyme stability. Ionic liquids may provide advantages as solvents in terms of increased substrate and product solubility, higher reactivity and selectivity, as well as tunable physicochemical properties, but their design should be carefully considered in relation to enzyme stability. The treatise shows that thermodynamic modeling tools for solvent design provide a new toolbox to design enzyme-catalyzed synthetic reactions from biomass sources.     

QSAR study on solubility of alkanes in water and their partition coefficients in different solvent systems using PI index

The aqueous solubility (S(w)) of liquids and solids, expressed as log S(w) as well as their partition coefficients in different solvent systems viz. P(oct) (partition coefficient in octanol-water), P(16) (partition coefficient in water-hexadecane), P(alk) (partition coefficient in water-alkane), and P(cyc) (partition coefficient in water-cyclohexane), and aqueous solubility (S(w)) have been estimated using the PI (Padmakar-Ivan) index and the results compared with those obtained using the widely used Wiener index (W). Regression analysis of the data using n-alkanes show that the PI index gives better results than the W index.     

Solvent effects on lipase-catalyzed esterification of glycerol and fatty acids

The lipase-catalyzed acylglycerol synthesis with fatty acids of different chain length is studied. Measured ester mole fractions at equilibrium are compared with calculated mole fractions. For these calculations the computer program TREP (Two-phase Reaction Equilibrium Prediction) is used. This program is based on the UNIFAC group contribution method and is developed for nondilute two-phase reaction systems.With one set of equilibrium constants, namely 1.3, 0.8, and 0.6 for monoester, diester, and triester synthesis, respectively, the equilibrium position of the reaction between glycerol and all saturated fatty acids with a chain length from 6 to 18 and oleic acid (cis-9-octadecenoic acid) can be calculated. Deviations, expressed as the ratio between calculated and measured ester mole fractions, usually were between 0.7 and 1.2. In the presence of solvents, the deviations of the monoester mole fractions were higher and rose up to 3. Without addition of a solvent, the ester mole fractions at equilibrium are dependent on the fatty acid chain length. With the short-chain hexanoic acid, the monoester mole fraction is the highest ester mole fraction, while for the long-chain oleic acid, the diester mole fraction is the highest one. The ester mole fractions become independent on the chain length of the fatty acid with a solvent added in a sufficient high concentration. Both reactions, with saturated and unsaturated C(18) fatty acids, lead to the same equilibrium position. The program TREP is found to make good predictions of the equilibrium amounts of ester and fatty acid. However, systematic deviations arise between measured and calculated amounts of water and glycerol in the organic phase. The calculated water and glycerol amounts are always lower than the measured ones. These deviations seem to be highest in nonpolar media and are probably due to deficiencies in the UNIFAC calculation method. Some preliminary experiments show the effect of the choice of solvent on the reaction rates. In polar solvents, the monoester production rate is enhances by a factor of 1.5 as compared to the reaction rate in a system without solvent. (c) 1993 John Wiley & Sons, Inc.     

Reaction Kinetics of Immobilized α-Chymotrypsin in Organic Media 1. Influence at Solvent Polarity

Esterification of N-acetyl phenylalanine with ethanol catalyzed by immobilized α-chymotrypsin (EC 3.4.21.1) was studied in various reaction media. The effect of reaction medium polarity on enzymatic activity as well as equilibrium yield was measured. The reaction rate increased with increasing amounts of added water, reaching an optimum corresponding to water saturation of the reaction medium. Further additions of water resulted in decreased activity. Bell-shaped activity profiles were obtained for all reaction media tested. Reaction media consisting of pure solvents and of mixtures of solvents were used. The enzymatic activity and the equilibrium yield increased with decreased polarity of the medium. Maximum activity was found in a reaction medium consisting of 80% diisopropyl ether and 20% heptane. The enzymatic activity obtained at optimal water additions in the different solvents and solvents mixtures could be correlated to the solubility of water and the log P of the medium. The equilibrium yield of the reaction was much more closely correlated to the solubility of water than the log P. Much lower enzymatic activity was obtained when solvent mixtures producing water-miscible media were created, than in mixtures producing water-immiscible media, such as mixtures of acetonitrile and diisopropyl ether. The equilibrium yield could be increased by decreasing the water content in the reaction medium, which reduced the water activity.     

A new approach to preparative enzymatic synthesis. Reprinted from Biotechnology and Bioengineering, Vol. XIX, No. 9, Pages 1351-1361

A new approach to preparative organic synthesis in aqueous-organic systems is suggested. It is based on the idea that the enzymatic process is carried out in a biphasic system "water-water-immiscible organic solvent." Thereby the enzyme is localized in the aqueous phase-this eliminates the traditional problem of stabilizing the enzymes against inactivation by a nonaqueous solvent. Hence, in contrast to the commonly used combinations "water-water-miscible organic solvent," in the suggested system the content of water may be infinitely low. This allows one to dramatically shift the equilibrium of the reactions forming water as a reaction product (synthesis of esters and amides, polymerization of amino acids, sugars and nucleotides, dehydration reactions, etc.) toward the products. The fact that the system consists of two phases provides another very important sources for an equilibrium shift, i.e., free energies of the transfer of a reagent from one phase to the other. Equations are derived describing the dependence of the equilibrium constant in a biphasic system on the ratio of the volumes of the aqueous and nonaqueous phases and the partition coefficients of the reagents between the phases. The approach has been experimentally verified with the synthesis of N-acetyl-L-tryptophan ethyl ester from the respective alcohol and acid. Porous glass was impregnated with aqueous buffer solution of chymotrypsin and suspended in chloroform containing N-acetyl-L-tryptophan and ethanol. In water (no organic phase) the yield of the ester is about 0.01%, whereas in this biphasic system it is practically 100%. The idea is applicable to a great number of preparative enzymatic reactions.     

A new approach to preparative enzymatic synthesis

A new approach to preparative organic synthesis in aqueous–organic systems is suggested. It is based on the idea that the enzymatic process is carried out in a biphasic system “water–water-immiscible organic solvent.” Thereby the enzyme is localized in the aqueous phase—this eliminates the traditional problem of stabilizing the enzyme against inactivation by a nonaqueous solvent. Hence, in contrast to the commonly used combinations “water–water-miscible organic solvent,” in the suggested system the content of water may be infinitely low. This allows one to dramatically shift the equilibrium of the reactions forming water as a reaction product (synthesis of esters and amides, polymerization of amino acids, sugars and nucleotides, dehydration reactions, etc.) toward the products. The fact that the system consists of two phases provides another very important source for an equilibrium shift, i.e., free energies of the transfer of a reagent from one phase to the other. Equations are derived describing the dependence of the equilibrium constant in a biphasic system on the ratio of the volumes of the aqueous and nonaqueous phases and the partition coefficients of the reagents between the phases. The approach has been experimentally verified with the synthesis of N-acetyl-L -tryptophan ethyl ester from the respective alcohol and acid. Porous glass was impregnated with aqueous buffer solution of chymotrypsin and suspended in chloroform containing N-acetyl-L -tryptophan and ethanol. In water (no organic phase) the yield of the ester is about 0.01%, whereas in this biphasic system it is practically 100%. The idea is applicable to a great number of preparative enzymatic reactions.     

Analysis of solvent-free ethyl oleate enzymatic synthesis at equilibrium conditions

This work reports experimental equilibrium data for the esterification of pure oleic acid and a fatty acid mixture with ethanol, using an immobilized Candida antarctica B lipase as catalyst. Reactions are performed in a solvent-free system, containing a mixture of substrates and different amounts of distilled water. According to the initial amount of water and the extent of the reaction, one or two liquid phases are present. Therefore, when the equilibrium is achieved, the liquid–liquid and chemical reaction equilibria have to be simultaneously satisfied.

Several reports dealing with enzymatic reactions performed in two-phase systems have found that the value of the reaction equilibrium constant calculated from overall experimental concentrations varies not only with temperature but also with substrate ratio and water content. Although this approach is a valuable way to explore equilibrium shifts in biphasic systems, it is limited to ideal systems with constant partition coefficients. The aim of this work is to consider the biphasic nature of the reactive mixture through a computational procedure that simultaneously takes into account liquid–liquid and reaction equilibria. This approach enables the determination of a classical temperature-dependent thermodynamic equilibrium constant, which accurately fits experimental equilibrium conversions over a wide range of operating conditions.     

In silico prediction of medium effects on esterification equilibrium using the COSMO-RS method

This paper presents a new approach for predicting solvent effects on esterification reactions of industrial importance in the field of biocatalysis. The COSMO-RS method has been used to calculate the activity coefficients of the chemical species involved in various reactions, carried out in different solvents. For comparison we also used the traditional UNIFAC method. Three lipase-catalyzed esterifications were considered: (1) 1-dodecanoic acid with menthol in n-hexane, n-heptane, cyclohexane, 2,2,4-trimethylpentane, toluene, acetonitrile, and 2-methyl-2-butanol; (2) 1-dodecanoic acid and 1-dodecanol in n-hexane, n-heptane, cyclohexane, 2,2,4-trimethylpentane, and toluene; and (3) glycerol and n-octanoic acid in acetonitrile, benzene, and toluene and in the neat reaction mixture (without any solvent). Predicted activities were used to calculate the thermodynamic equilibrium ratio. This should be independent of medium, and the variation in COSMO-RS values is at most 9-fold (corresponding to a DeltaG degrees of about 5.5 kJ/mol, which would still be a very useful prediction) and often only 2-fold (corresponding to less than 2 kJ/mol or 0.5 kcal/mol, therefore comparable with experimental error). UNIFAC is weaker, especially when important roles are played by conformational freedom, intramolecular interactions, strong polar effects, and charge distribution of molecules in the mixture. The relative percent deviations from the mean of equilibrium constants in different solvents range between 17 and 49 for COSMO-RS versus 32 to 65 for UNIFAC. The COSMO-RS method opens up new perspectives for the development of theoretical models for solvent selection with general applicability.     

Effect of water activity on enzyme hydration and enzyme reaction rate in organic solvents

The effects of water on enzyme (protein) hydration and catalytic efficiency of enzyme molecules in organic solvents have been analyzed in terms of the thermodynamic activity of water, which has been estimated by the NRTL or UNIFAC equations. When the amount of water bound to the enzyme was plotted as a function of water activity, the water adsorption isotherms obtained from the water-solvent liquid mixtures were similar to the reported water-vapor adsorption isotherms of proteins. The water adsorption of proteins from the organic media was not significantly dependent on the properties of the solvents or the nature of the proteins. It is also shown that there is a linear relationship between the logarithm of the enzyme reaction rate and water activity. However, the dependence of the enzyme reaction rate on water activity was found to be different depending on the properties of the solvent. The relationship between water activity and other solvent parameters such as solvent hydrophobicity and the solubility of water in the solvent is also discussed.     

A three-dimensional solubility parameter approach to nonaqueous enzymology

Widespread commercial application of enzymes as catalysts for specialty or commodity chemical synthesis will require their use in nonaqueous systems. While a number of non-aqueous enzyme applications have been demonstrated, the lack of useful rules for predicting enzyme-solvent interactions has hindered the development of this technology. Both Hildebrand and solvent hydrophobicity (octanol-water partition coefficient) parameters have been used previously to correlate and predict enzyme activity in nonaqueous systems, with some success, but any single-parameter approach is inherently limited in its ability to reflect the spectrum of possible enzyme-solvent interactions. Therefore, this study evaluates the three-dimensional solubility parameter space, as proposed by Hansen, to correlate and predict enzyme activity in microaqueous, miscible, and biphasic nonaqueous systems. Preliminary results suggest that Hansen parameters may be useful for correlating nonaqueous enzyme activity, and that the dispersive and polar parameters may be disproportionately important in single-phase microaqueous systems. The Hansen hydrogen-bonding parameter appears to be the only parameter yet evaluated capable of correlating the water requirement for enzyme activity in microaqueous systems, suggesting that water affects protein structure through enthalpic rather than entropic processes in nonaqueous systems. Insufficient data are available for miscible and biphasic systems, but it is proposed that enzyme activity may correlate with the average solubility parameters of miscible systems and of the aqueous phase in biphasic systems.     

Rules for the regulation of enzyme activity in reserved micelles as illustrated by the conversion of apolar steroids by 20 beta-hydroxysteroid dehydrogenase

20 beta-Hydroxysteroid dehydrogenase was enclosed in reversed micellar media consisting of cetyltrimethyl-ammonium bromide, hexanol, organic solvent and Hepes buffer. The influence of the composition of these media on the enzymatic reduction of the apolar steroids progesterone and prednisone was investigated by varying the water content, concentration of hexanol and type of organic solvent. By changing the water content and the type of organic solvent, the hexanol to cetyltrimethylammonium bromide ratio in the interphase can be varied. This ratio was determined by phase boundary titrations. It was found that the higher this ratio, the higher the rate of steroid conversion. From variations of the hexanol content it was concluded that the rate of steroid conversion is determined by the hydrophobicity of the steroid relative to the hydrophobicity of the continuous phase and the hydrophobicity of the interphase. The hydrophobicity of the phases was expressed in log P-values. Log P is defined as the logarithm of the partition coefficient in an octanol-water two-phase system. This enabled us to derive the following relations between the hydrophobicity values for the substrate (log Ps), for the interphase (log Pi) and for the continuous phase (log Peph): [log Pi-log Ps] must be minimal to ensure a high steroid concentration in the interphase and [log Pcph-log Ps] must be large to keep the steroid concentration in the continuous phase low. With these considerations, for any given apolar compound, a medium can be composed that gives optimal enzymatic conversion.     

Computer-aided solvent screening for biocatalysis

A computer-aided solvent screening methodology is described and tested for biocatalytic systems composed of enzyme, essential water and substrates/products dissolved in a solvent medium, without cells. The methodology is computationally simple, using group contribution methods for calculating constrained properties related to chemical reaction equilibrium, substrate and product solubility, water solubility, boiling points, toxicity and others. Two examples are provided, covering the screening of solvents for lipase-catalyzed transesterification of octanol and inulin with vinyl laurate. Esterification of acrylic acid with octanol is also addressed. Solvents are screened and candidates identified, confirming existing experimental results. Although the examples involve lipases, the method is quite general, so there seems to be no preclusion against application to other biocatalysts.     

The mode of action of N-(n-Dodecyl)diethanolamine with particular reference to the effect of protonation on uptake by Escherichia coli.

In a homologous series of N-(n-alkyl)diethanolamines antimicrobial activity was related to surface activity and increasing octanol-water partition coefficient. Maximum activity was exhibited by the dodecyl-, tetradecyl- and hexadecyl-derivatives. Dodecyldiethanolamine (DDE) displayed a broad spectrum of activity. Towards Escherichia coli NCIB8277, its bacteriostatic and bactericidal activity increased as the degree of protonation lessened, and may have been influenced by the formation of micelles. Uptake of DDE by washed suspensions of E. coli was more rapid and more extensive at pH 7.0 than pH 4.0. Within this pH range, bacterial uptake, the octanol-water partition coefficient (lipid solubility) and the proportion of unprotonated DDE all increased. Uptake isotherms at pH values in the range 4.0 to 8.0 are interpreted as signifying different uptake mechanisms for the protonated and unprotonated forms.