Enzymatic synthesis of sugar fatty acid esters in ionic liquids

Abstract

Sugar fatty acid esters are nonionic biosurfactants widely used in food, cosmetic and pharmaceutical industries. Ionic liquids (ILs) have shown great potential as an alternative reaction medium to conventional organic solvents for their enzymatic synthesis, promoting a significant enhancement in sugar solubility, enzymatic reactivity and regioselectivity. In this perspective, we summarize the advances in the use of ILs as a solvent for sugar dissolution and lipase-catalyzed sugar ester synthesis, review the recent achievements in designing enzyme-compatible ILs for the synthesis, and discuss the challenges that the research on this topic is facing. Among the advanced ILs, deep eutectic solvents (DESs) merit a serious consideration as a potential reaction medium for this application.

---

Zhen Yang

Zhen Yang obtained her BSc degree in Chemistry from Zhongshan University, China, in 1985 and PhD degree in Biochemistry from the University of Strathclyde, UK, in 1992. This was followed by postdoctoral research with Prof. Alan Russell in the University of Pittsburgh, USA. Since 1995 she had been working as a Research Scientist in industry until joining Shenzhen University, China, in 2004 as an Associate Professor. She was promoted to Full Professor in 2008. Her current research interests involve the use of enzymes in nonconventional reaction media (e.g., in ionic liquids): their catalytic performance, biophysical chemistry, and applications.

Ze-Lin Huang

Ze-Lin Huang obtained his Bachelor's degree in Biotechnology from Shenzhen University, China, in 2010, and is currently a postgraduate student pursuing his Master's degree in Biochemistry from Shenzhen University. He has been working on research projects involving lipase-catalyzed biodiesel production and sugar ester synthesis.

---

1. Introduction

Sugar fatty acid esters are nonionic biosurfactants produced from naturally renewable resources such as carbohydrates (e.g., glucose and sucrose) and fatty acids (e.g., lauric and palmitic acids), and have been widely utilized in food, cosmetic and pharmaceutical industries because they are tasteless, odorless, nontoxic, nonirritant, and biodegradable. Moreover, their functional characteristics, such as critical micelle concentration (CMC) and hydrophilic–lipophilic balance (HLB), can be formulated over a wide range by tuning the constitutive sugar and fatty acid moieties of the sugar esters.¹

While almost all the sugar ester commercial products on the market are manufactured by chemical syntheses, there has been a growing interest in developing enzymatic approaches to synthesize them. The chemical processes usually require the use of reagents and catalysts that may cause safety concerns for food, cosmetic and pharmaceutical products, involve multiple protecting and deprotecting steps working under harsh conditions, and produce poorly regioselective sugar ester products either acylated at different positions or with different degrees of acylation.² Comparatively, the enzymatic method is more advantageous in terms of mild reaction conditions, simple operational procedures, high productivity, excellent regioselectivity, and easy product separation.³

Enzymatic synthesis of sugar esters can be achieved by direct esterification of sugars with fatty acids or by their transesterification with fatty acid vinyl esters, both being catalyzed by lipases from different sources (Scheme 1). So far Novozym 435 from Novozyme (Candida antarctica lipase B immobilized on acrylic resin) has been the most popular enzyme used for this application. The use of the vinyl esters is to improve the production yield by driving the reaction through the tautomerization of the enol product. These synthetic reactions are preferred to be performed in a nonaqueous medium. However, the enzymatic methods often suffer from slow reaction rates and low production yields, which are mainly caused by the poor solubility of carbohydrates in most conventional organic solvents. Sugars are only soluble in a few polar solvents such as pyridine, dimethyl sulfoxide (DMSO), and dimethyl formamide (DMF). However, these solvents have been found to be serious deactivators of enzymes and also unacceptable to be used in food applications. Alternatively, the solubility of the sugars can be increased by chemical modification such as complexation with phenylboronic acid⁴ and preparation of isopropylidene derivatives⁵ and alkyl glycosides,⁶ but additional problems arise due to the fact that either extra protecting and deprotecting steps are required or the products thus generated do not show the same properties that the nonderivatized sugar esters should possess. Although sugar esters can be quantitatively produced based on a solid-phase system with addition of an organic solvent (typically acetone) as adjuvant,⁷ this is not practical in consideration of its up-scaling and continuous operation. Therefore, for enzymatic synthesis of sugar esters, the major issue to be tackled is the selection of appropriate solvents.

Scheme 1 Lipase-catalyzed synthesis of glucose esters in nonaqueous media.

Ionic liquids (ILs) are organic salts which are liquids under ambient temperatures. Currently, they have been the focus of increasing attention as a promising new type of nonaqueous solvent for biocatalytic processes due to their low volatility, high chemical and thermal stability, tunable solvent properties, and the excellent enzymatic performance observed in these solvents.^8–11 One major attraction of using ILs as an alternative to conventional organic solvents is their “designer solvent” properties, because their physical and chemical properties, including melting point, polarity, hydrophobicity, viscosity, and solvent miscibility, can be widely tuned through appropriate modification of the cation, the anion, and their attached substituents, thus allowing them to be specifically designed for different reaction conditions. In particular, ionic liquids have shown the great power of dissolving compounds of varying polarity, especially proteins and carbohydrates, which are sparingly soluble in common organic solvents, and this solubilizing ability can even be further improved by taking advantage of this “designer solvent” property. All these features have made ionic liquids a perfect choice as a reaction medium for carbohydrate biotransformations.

Although extensive literature already exists for lipase-mediated synthesis of sugar fatty acid esters in organic media (for an extensive review see ref. 3), the reports on performing this synthesis in ionic liquids have been limited. In this article, we first review the recent advances in using ionic liquids as a solvent for sugar dissolution and as a reaction medium for enzymatic synthesis of sugar esters. This leads to our consideration concerning the topic of designing ILs for this application. Finally, we discuss the challenges that the research has to face, in the hope of paving the way for further exploration in this subject.

2. Solubilization of sugars in ionic liquids

The potential of using ionic liquids as the reaction medium for carbohydrate transformations was first proposed by Sheldon's group¹² and was later confirmed by Park and Kazlauskas:¹³ among the 8 ILs tested, the best IL is [MOEMim][BF₄], which dissolves approximately 100 times more glucose (∼5 mg mL⁻¹ at 55 °C) than organic solvents such as acetone or THF; while the worst IL, [BMIm][PF₆], dissolves less than 1 mg mL⁻¹ of glucose at the same temperature.

Almost at the same time, Kimizuka and Nakashima¹⁴ developed two imidazolium ILs, [MOMMIm][Br] and [MOEMIm][Br], which were capable of dissolving carbohydrates such as β-D-glucose, α-cyclodextrin, amylose, and agarose. These two ILs had an ether moiety incorporated in the side chain of the imidazolium cation and a bromide as the anion. The “sugar-philic” property of these ILs was supported by the fact that they can both dissolve glucose oxidase, a highly glycosylated protein, but not other proteins (e.g., cytochrome C, myoglobin, hemoglobin, catalase). The authors suspected that this “sugar-philic” property was attributable to the ether linkage of the IL cation as a H-bond acceptor for the sugar hydroxyl groups. In fact, the bromide anion may also work as a H-bond acceptor and hence might also contribute to the high solubility of carbohydrates in these ILs.¹⁵

This later idea was supported by a discovery from Rogers' group:¹⁶ cellulose is soluble in a series of 1-methyl-3-alkylimidazolium ILs with chloride as the anion, and the solubility is lower in the IL with a longer alkyl chain (see entry 3 of Table 1); whereas ILs containing ‘non-coordinating’ anions, including BF₄⁻ and PF₆⁻ which exhibit lower H-bond basicity, were nonsolvents for cellulose. This can be explained by the speculation that the high chloride concentration and activity in the IL are responsible for breaking the extensive H-bonding network present in cellulose and solubilizing it through H-bonding with the hydroxyl groups of the polysaccharide; hence the lower solubilizing power obtained by the imidazolium ILs attached with a longer alkyl chain is simply due to the reduced effective chloride concentration within these liquids. The H-bonding between chloride anions and hydroxyl groups of glucose and cellulose has been confirmed by molecular dynamics simulations^21,22 and ¹³C and ^35/37Cl NMR relaxation measurements,²³ respectively.

Table 1 Solubility of carbohydrates in some ILs^a

Entry	Reference	IL	Carbohydrate	Solubility	Temperature
a MOMMIm = 1-methoxymethyl-3-methylimidazolium; MOEMIm = 1-methoxyethyl-3-methylimidazolium; EOEMIm = 1-ethoxyethyl-3-methylimidazolium; EMIm = 1-ethyl-3-methylimidazolium; Bt14 = 1-butyl-3-methylbenzotriazolium; Bt1Bn = 1-benzyl-3-methylbenzotriazolium; dca = dicyanamide, N(CN)2; Tf2N = bis(trifluoromethane)sulfonamide, (CF3SO2)2N; TfO = trifluromethylsulfonate, CF3SO3; Amm110 = cation of the commercial IL AMMOENG^TM 110 from Solvent Innovation GmbH; KGM = konjac glucomannan
1	13	[MOEMIm][BF4]	β-D-Glucose	∼5 g L^−1	55 °C
2	14	[MOEMIm][Br]	β-D-Glucose	450 g L^−1	Heating
			α-Cyclodextrin	350 g L^−1	Heating
			Amylose	30 g L^−1	Heating
			Agarose	20 g L^−1	Heating
3	16	[C4MIm][Cl]	Cellulose	3 wt.%	70 °C
		[C4MIm][Cl]	Cellulose	10 wt.%	100 °C
		[C6MIm][Cl]	Cellulose	5 wt.%	100 °C
		[C8MIm][Cl]	Cellulose	Slightly soluble	100 °C
4	17	[EMIm][dca]	Glucose	>10 wt.%	75 °C
		[EMIm][MeSO3]	Glucose	10 wt.%	75 °C
		[Bt14][dca]	Glucose	>10 wt.%	75 °C
		[Bt14][MeSO3]	Glucose	6 wt.%	75 °C
		[Bt1Bn][dca]	Glucose	6 wt.%	75 °C
		[Bt1Bn][Tf2N]	Glucose	2 wt.%	75 °C
		[Bt1Bn][MeSO3]	Sucrose	2 wt.%	75 °C
5	15	[MOMMIm][Tf2N]	β-D-Glucose	0.5 g L^−1	25 °C
		[MOMMIm][BF4]	β-D-Glucose	4.4 g L^−1	25 °C
		[MOMMIm][TfO]	β-D-Glucose	4.3 g L^−1	25 °C
		[MOMMIm][dca]	β-D-Glucose	66 g L^−1	25 °C
		[MOEMIm][Tf2N]	β-D-Glucose	0.5 g L^−1	25 °C
		[MOEMIm][PF6]	β-D-Glucose	2.5 g L^−1	25 °C
		[MOEMIm][BF4]	β-D-Glucose	2.8 g L^−1	25 °C
		[MOEMIm][TfO]	β-D-Glucose	3.2 g L^−1	25 °C
		[MOEMIm][dca]	β-D-Glucose	91 g L^−1	25 °C
		[EOEMIm][Tf2N]	β-D-Glucose	0.5 g L^−1	25 °C
		[EOEMIm][PF6]	β-D-Glucose	0.7 g L^−1	25 °C
		[EOEMIm][BF4]	β-D-Glucose	2.8 g L^−1	25 °C
		[EOEMIm][dca]	β-D-Glucose	70 g L^−1	25 °C
		[BMIm][BF4]	β-D-Glucose	<0.5 g L^−1	25 °C
		[BMIm][PF6]	β-D-Glucose	<0.5 g L^−1	25 °C
		[BMIm][dca]	β-D-Glucose	145 g L^−1	25 °C
			Sucrose	195 g L^−1	25 °C
			Lactose	51 g L^−1	25 °C
			β-Cyclodextrin	750 g L^−1	75 °C
			Amylose	4 g L^−1	25 °C
6	18	[C2MIm][BF4]	KGM	0.266 g L^−1	50 °C
		[C4MIm][BF4]	KGM	0.336 g L^−1	50 °C
		[C8MIm][BF4]	KGM	0.191 g L^−1	50 °C
		[C4MIm][PF6]	KGM	0.162 g L^−1	50 °C
7	19	[EMIm][MeSO4]	Glucose	89.6 g L^−1	25 °C
		[EMIm][TfO]	Glucose	6.1 g L^−1	25 °C
		[EMIm][BF4]	Glucose	1.1 g L^−1	25 °C
		[BMIm][TfO]	Glucose	4.8 g L^−1	25 °C
		[BMIm][BF4]	Glucose	0.9 g L^−1	25 °C
		[BMIm][PF6]	Glucose	<0.5 g L^−1	25 °C
		[OMIm][BF4]	Glucose	0.7 g L^−1	25 °C
		[EMIm][TfO]	Fructose	32.8 g L^−1	25 °C
		[EMIm][BF4]	Fructose	7.7 g L^−1	25 °C
		[BMIm][TfO]	Fructose	27.0 g L^−1	25 °C
		[BMIm][PF6]	Fructose	3.3 g L^−1	25 °C
8	20	[BMIm][Tf2N]	D-Glucose	<0.5 wt.%	60 °C
		[EMIm][Ac]	D-Glucose	60 wt.%	60 °C
		[Amm110][Ac]	D-Glucose	30 wt.%	60 °C
		[Amm110][dca]	D-Glucose	4.5 wt.%	60 °C
		[Me(OEt)3-Et-Im][Ac]	D-Glucose	80 wt.%	60 °C
		[Me(OEt)7-Et-Im][Ac]	D-Glucose	26 wt.%	60 °C
		[Me(OPr)3-Et-Im][Ac]	D-Glucose	45 wt.%	60 °C
		[Me(OEt)3-Et3N][Ac]	D-Glucose	16 wt.%	60 °C

---

Meanwhile, MacFarlane's group has prepared two new families of ionic liquids, ammonium and imidazolium ILs containing dicyanamide (dca, N(CN)₂⁻) as the anion, which have been found to be water-miscible and have low viscosity (e.g., η = 21 for [EMIm][dca]) and low melting points (most being liquid at room temperature).²⁴ The low viscosity and low melting point presented by these ILs may result from the charge delocalization of the anion, thus producing weak ion–ion interactions. Soon they found out that these new ILs are all good solvents for glucose (greater than 10 weight percent at room temperature) and interestingly, [EMIm][dca] can work not only as an effective solvent but also as an active base catalyst for O-acetylation of glucose with alcohol.²⁵ Later, they reported that all the dca-based ILs they tested, containing imidazolium, triazolium and benzotriazolium as the cations, were able to dissolve glucose and sucrose in a high concentration; glucose was soluble in most of the mesylate (CH₃SO₄⁻) ILs but almost completely insoluble in the ILs based on Tf₂N⁻ anion.¹⁷

The above observations led Sheldon and co-workers to undertake a systematic study of the effect of the anion and alkyl chain in the cation on the ability of dialkylimidazolium ILs to dissolve mono-, di- and polysaccharides (see also entry 5 of Table 1).¹⁵ The following conclusions can be drawn from their study:

(1) [BMIm][dca] can readily dissolve monosaccharides (e.g., β-D-glucose), disaccharides (e.g., sucrose, lactose), oligosaccharides (e.g., β-cyclodextrin), and polysaccharides (e.g., amylose) to a concentration much greater than the solubility of glucose in tert-butanol (t-BuOH, 0.3 g L⁻¹ at 25 °C) and the solubility is generally reduced when the length of the sugar chain gets longer.

(2) The solubility of glucose varies in a wide range with the IL, but it seems to be more dependent on the IL anion rather than on the cation. The influence of the IL anions on the glucose solubility seems to follow a common trend for all the IL cations tested: dca > TfO > BF₄ > PF₆ > Tf₂N. All the dca-based ILs tested dissolve glucose more than an order of magnitude better than ILs containing other anions, demonstrating again the superior solubilizing ability of these ILs, which may be attributed to dca's H-bond acceptor property.

(3) The influence of the IL cations also counts. For ILs containing BF₄⁻ or PF₆⁻ as the anion, the cations affect the glucose solubility following a similar pattern: MOMMIm > MOEMIm > EOEMIm > BMIm, showing again the positive effect of the ether-containing side chain of the IL cation. Surprisingly, however, [BMIm][dca] was a much better solvent for glucose than all its oxygenated counterparts.

(4) As expected, the sugar solubility in ILs increased with an increase in temperature.

On the other hand, instead of searching for ILs that can dissolve carbohydrates in high concentrations, Koo’s group has focused on developing a water-mediated supersaturation method to increase the solubility of sugars in ILs:¹⁹ ILs are mixed with an aqueous sugar solution followed by removal of the water from the solution through vacuum evaporation. As can be seen from Fig. 1, this new method can significantly increase the solubility of sugars in all the ILs tested. For instance, the concentrations of glucose, fructose and sucrose in supersaturated [EMIm][TfO] at 25 °C were 19, 7 and 5-fold increased respectively, relative to the normal solubilities of these sugars in the same IL. The supersaturated glucose solutions in ILs were fairly stable over a long period of time, presumably due to the high viscosity of the ILs that prevents sugars from crystallization and precipitation. Additionally, these solubility data¹⁹ also agree with the IL anion order pointed out above. In terms of the cation effect, the solubilities of the three sugars tested were all higher in the IL with a shorter alkyl chain attached to the imidazolium ring. This is in accord with the results reported previously,^16,18,20 suggesting that the IL cation with a lower hydrophobicity may facilitate the sugar dissolution.

Fig. 1 Comparison of the concentrations of glucose in different ILs prepared at 25 °C by using different methods: (1) normal saturation: excess amounts of sugars were mixed thoroughly with ILs at 25 °C before supernatant was obtained by centrifugation; (2) conventional supersaturation: saturated solutions of sugars in ILs were prepared at 60 °C and then slowly cooled down to 25 °C; (3) water-mediated supersaturation: ILs are mixed with an aqueous sugar solution followed by removal of the water through vacuum evaporation at 60 °C, and the solutions were then slowly cooled to 25 °C and centrifuged to obtain the supersaturated sugar solutions in ILs. The actual concentration of glucose in [EMIm][MeSO₄] prepared by the water-mediated supersaturation method should be greater than 500 g L⁻¹ according to the reference, but the actual result has not be specified. Reproduced by using data given in ref. 19.

3. Lipase-catalyzed synthesis of sugar esters in ionic liquids

Park and Kazlauskas¹³ were the first to introduce the lipase-catalyzed acylation reaction of unmodified sugars into ionic liquids. The regioselective acylation of both mono- and disaccharides (e.g., glucose and maltose, respectively) with vinyl acetate, catalyzed by Novozym 435, can proceed smoothly in ILs, not only with a high conversion but also with a high regioselectivity for monoacylation at the 6-position. Among the 8 ILs tested, [MOEMIm][BF₄] was the best, yielding a maximum conversion of 99% and an excellent monoacylation rate of 93% within a reaction period of 36 h. Comparatively, in THF a yield of 99% was also obtained, but of which only 53% was the desired 6-O-acetyl compound while the rest was the side product, 3,6-O-diacylglucose. Achieving such a high regioselectivity is very challenging for chemical synthesis, because the reactivities of different primary and secondary hydroxyl groups of the sugar do not vary much, although the primary ones do react slightly preferentially.² No correlation between the reaction rate of the glucose acylation and the solvent polarity was observed, and the authors speculated that it was the ability of the solvent to dissolve the substrate glucose that was responsible for the high reaction rate and high regioselectivity in the ILs.¹³ Glucose is up to one hundred times more soluble in ILs (such as [MOEMIm][BF₄]) than in acetone or THF, whereas in the latter organic solvents acetylation of insoluble glucose yielded the more soluble monoacetylated product, 6-O-acetyl glucose, which underwent further acetylation to give the diacylated product, 3,6-O-diacyl-glucose. It is worth mentioning that with [MOEMIm][BF₄] as the best solvent for this application, this experiment has again shown the expected positive effect of the oxygenated side-chains of the IL cations.

Later, Kim et al.²⁶ explored the selective acylation of alkyl glycosides, catalyzed by Candida rugosa lipase, in both ILs ([BMIm][PF₆] and [MOEMIm][PF₆]) and organic solvents (THF and chloroform), confirming again that both reaction rate and regioselectivity obtained in ILs were significantly enhanced relative to those obtained in organic solvents.

In 2005, Ganske and Bornscheuer²⁷ described for the first time that glucose fatty acid esters can be enzymatically synthesized from glucose and fatty acid vinyl esters through transesterification in ionic liquids. Candida antarctica lipase B (CALB) modified with polyethylene glycol was able to catalyze the transesterification of glucose with lauric and myristic acid vinyl esters in pure [BMIm][BF₄] and [BMIm][PF₆], with a conversion of 30% and 35%, respectively. The conversion was markedly improved (up to 90%) when the reaction was performed with an immobilized CALB (Chirazyme L2 C2 from Boehringer Mannheim) in a bisolvent (biphasic) system composed of an IL ([BMIm][BF₄] or [BMIm][PF₆], 60%) and t-BuOH (40%). NMR analysis indicated that the 6-O-acylated monoester was formed exclusively, confirming again the superior regioselectivity in the IL system. Although the free enzyme did not enable the direct esterification reaction to proceed in pure ILs, in the bisolvent system catalyzed by the immobilized CALB a high conversion of 64% was obtained by esterification between glucose and palmitic acid. This IL/t-BuOH bisolvent system was further optimized in terms of enzyme, chain-length and type of acyl donor, temperature, reaction time, and percentage of co-solvent, and the commercial CALB (Chirazyme L2 C2) was identified as the best enzyme for this synthetic reaction.²⁸ No substantial differences between [BMIm][BF₄] and [BMIm][PF₆] was observed in this study. Interestingly, with both lauric acid vinyl ester and free palmitic acid as acyl donors, there was always a bell-shaped relationship between the conversion in the [BMIm][PF₆]/t-BuOH bisolvent system and the percentage of t-BuOH in the solvent mixture, and the optimal t-BuOH content was always 40%. Undoubtedly, addition of the co-solvent can reduce the viscosity of the IL-containing medium and hence improve the mass transfer in the reaction system; this may be the major reason for the initial increase in the enzyme reactivity. However, too much t-BuOH added may lead to a reduction in the glucose solubility, thus resulting in a lower conversion yield.

Later on, Zong's group¹⁸ successfully carried out the lipase-catalyzed acylation of konjac glucomannan (KGM) with vinyl acetate in pure ILs and IL/t-BuOH bisolvent systems. They have investigated the effect of water activity, solvent, reaction temperature, and cosolvent content on the acylation of KGM catalyzed by Novozym 435. These lipase-mediated acylation reactions proceeded in the ILs ([C_nMIm][BF₄] (n = 2,4,8) and [C₄MIm][PF₆]) with a higher substitution degree and a higher thermostability, as compared to the results obtained in t-BuOH. The degrees of substitution and the production yields were further increased by addition of t-BuOH as the co-solvent. In all the solvent systems (ILs, t-BuOH, and IL/t-BuOH), the acylation occurred merely to the C-6 of the sugar units, as has been confirmed by ¹³C NMR analysis.

By employing the water-mediated supersaturation method they developed to produce a glucose solution with high concentration, Koo and co-workers¹⁹ have succeeded in synthesizing glucose fatty acid esters in ILs not only through transesterification but also through esterification reactions. [BMIm][TfO] has been shown to be the solvent that presented the highest conversion. As shown in Table 2, both the initial rates and conversions were notably increased when the glucose solution was prepared by using the water-mediated supersaturation method. It is worth mentioning that these three reactions were compared in the same solvent, [BMIm][TfO], containing the same glucose content (40 g L⁻¹): this amount of glucose was only completely dissolved in the water-mediated supersaturated system, while the other two systems still contained a fair amount of undissolved glucose during the whole course of the reaction. Therefore, this study has clearly indicated that the concentration of the homogeneously dissolved glucose in the IL is critical to the synthesis of the sugar esters.

Table 2 Enzymatic synthesis of glucose laurate via transesterification and esterification in [BMIm][TfO] (data were obtained from ref. 19)

Preparation of glucose solution	Glucose solubility (g L^−1)	Transesterification		Esterification
		Initial reaction rate (μmol min^−1 g^−1)	Conversion (24 h)	Initial reaction rate (μmol min^−1 g^−1)	Conversion (100 h)
Normal saturation	9.9	0.17	∼8%	0.10	27%
Conventional supersaturation	17.2	0.83	∼17%
Water-mediated supersaturation	40.0	15.0	96%	2.03	91%

---

Soon after this, the group endeavored to optimize the esterification process by performing the reaction in an IL mixture of [BMIm][TfO]/[OMIm][Tf₂N] for the purpose of achieving maximal productivity.²⁹ The selection of these two ILs was based on the fact that the lipase (Novozym 435) exhibited the highest activity but a rather low stability in the hydrophilic [BMIm][TfO], whereas the hydrophobic [OMIm][Tf₂N] offered a higher enzyme stability, although the enzyme activity in this IL was much lower than in [BMIm][TfO]. As a result of compromise, the mixture of these two ILs with a 1 : 1 volume ratio was selected as the best reaction medium for the lipase-catalyzed synthesis of sugar esters. By using this new solvent system, a series of glucose fatty acid esters have been synthesized through direct esterification with different fatty acids. As expected, use of a supersaturated glucose solution offered a higher conversion relative to the reaction with a saturated glucose solution. Interestingly, the conversion was higher for the glucose ester with a shorter fatty acid chain, and this may be related to the difference in the solubility of the fatty acid in the ILs.

Koo and colleagues have also introduced ultrasound irradiation into the lipase-catalyzed sugar ester synthesis, and has found that in both esterification and transesterification reactions, this new technique enabled the conversions to be greatly enhanced without significantly perturbing the enzyme stability.³⁰

Very recently, Abdulmalek et al.³¹ have developed another strategy to significantly improve the sugar ester synthesis through direct esterification: considering that DMSO is a good solvent for dissolving sugars, it was introduced into the IL system as a solubilizing agent and a co-solvent as well. After examining different reaction parameters (type of solvent and enzyme, amount of enzyme, reaction time, temperature, stirring rate and molar ratio of the two substrates), the authors worked out the optimal reaction conditions for the synthesis of galactose oleate. The esterification between galactose and oleic acid was undertaken at 60 °C with a stirring rate of 300 rpm in a [BMIm][BF₄]/DMSO (20 : 1 v/v) bisolvent system containing the two substrates in the molar ratio of 1 : 3, with addition of 2% (w/w) Lipozyme RM IM as the catalyst. A maximum conversion (over 90%) was achieved after reaction for only 4 h, a much faster process as compared to others that usually require a few days.^19,28

4. Designing ionic liquids for sugar ester synthesis

From what has been discussed above, it has become evident that ILs are good solvents for dissolving carbohydrates so as to increase their solubility, which is important for the lipase-catalyzed synthesis of sugar esters. However, whether an IL that affords a high sugar solubility is also a good solvent for the enzymatic reaction to take place is very questionable. A typical example is given by the dca-based ILs: although these ILs have been well known to be highly effective for dissolving carbohydrates in large quantities,^15,25 they may not necessarily be a good solvent for the lipase-catalyzed synthetic reactions, probably depending on factors such as the cation that the IL contains and the immobilization method that applies to the enzyme. In [BMIm][dca], which offers remarkable solubilities for different sugars,¹⁵ the free CALB was irreversibly deactivated after a transesterification reaction (ethyl butyrate with butanol), but its cross-linked enzyme aggregates (CLEAs) presented an activity twice as high as in t-BuOH.³² When catalyzing another transesterification reaction (N-acetyl-L-phenylalanine ethyl ester with 1-propanol), Novozym 435 (CALB immobilized on acrylic resins) showed a low activity in this IL as well, but was fairly active in the other two dca-based ILs, namely [choline][dca] and [Amm110][dca].²⁰ This latter IL has even afforded the same immobilized lipase the highest conversion when catalyzing the acylation of glucose with methyl methacrylate, relative to other ILs that dissolve much more glucose (16–80 wt.% in these ILs vs. 0.45% in [Amm110][dca]).²⁰

Indeed, one can even notice a clear trend of higher lipase activity with lower sugar solubility for a series of homologous imidazolium ILs involved in Zhao's study (Fig. 2).²⁰ A likely explanation for this is that anions like acetate can form H-bonds with the hydroxyl groups of the sugar molecules, which has been considered as the mechanism of sugar dissolution in ILs. These anions can also form H-bonds with protein molecules by dissociating the H-bonds that originally existed in the protein molecules for maintaining their structural integrity, thus denaturing the enzyme.³¹ A longer side-chain attached to the IL cation induced a lower effective acetate concentration in these ILs, thus rendering the ILs less efficient at dissolving cellulose but were also more favorable to enzyme activity. A lower enzyme activity with a higher sugar solubility in the IL can also be explained by the substrate ground-state stabilization.²⁰

Fig. 2 Cellulose solubility (wt.%) and lipase activity (μmol min⁻¹ g⁻¹) in acetate-based imidazolium ILs with a side-chain attached on the imidazolium ring (the x axis gives the carbon numbers of these side chains). The cations of the ILs 1–5 shown in the figure refer to EMIm⁺ and Me(OEt)_n-Et-Im⁺ (n = 2,3,4,7), respectively. Reproduced by using data given in ref. 20.

Therefore, it is necessary to investigate whether there are some general structural patterns for the ILs that are favorable to sugar dissolution or enzyme activity or both. This can help us to set up some guidelines for our design of ILs suitable as a reaction medium for the enzymatic synthesis of sugar esters. A detailed study of 36 different ILs by Zhao et al.²⁰ has provided us such a platform to examine the correlation between the IL structure and both the sugar solubility and lipase activity in the ILs. Based on their study and also the preceding discoveries, we can reach the following key points:

(1) ILs containing anions such as acetate and dicyanamide are good solvents for sugar dissolution, but may also be enzyme-denaturing.

(2) Grafting an alkyloxy or alkyloxyalkyl chain onto the IL cation is beneficial both to sugar dissolution due to its H-bond forming ability and to enzyme activity, because the effective concentration of the denaturing anion in the IL can be reduced while a relatively low viscosity of the IL can be ensured.

(3) IL cations with bulky groups or with a hydroxyl group at the end of the side-chain may not favor sugar dissolution. The possible reason for the latter case may be that the terminal hydroxyl group of the cation might compete with the sugar for H-bonding with the IL anion, thus reducing the H-bonding interaction between the sugar molecule and the IL anion.

(4) ILs with aprotic cations or formate anions may not be suitable for enzymes because of the high acidity or basicity, respectively.

Following these guidelines, Zhao et al.²⁰ have proposed that ILs containing H-bond forming anions and oxygenated cations with low bulkiness will benefit the sugar dissolution, while a low concentration of these H-bond forming anions may be essential for promoting the enzyme activity. This is actually very well illustrated by the two types of ILs they designed (Scheme 2): Both the imidazolium and ammonium ILs showed a good solubility for glucose (16–80 wt.%) and cellulose (3–12 wt.%), and the lipase (Novozym 435) also presented a high activity when catalyzing the transesterification of glucose and cellulose with methyl methacrylate in these ILs.

Scheme 2 Two types of enzyme-compatible ionic liquids designed by Zhao et al.:²⁰ (1) [Me(OEt)_n-Et-Im][Ac]; (2) [Me(OEt)₃-Et₃N][Ac].

5. Summary, challenges and outlook

The experimental results accumulated so far have convinced us that ionic liquids hold a great potential as an ideal reaction medium for enzymatic synthesis of sugar esters, promoting a significant enhancement in sugar solubility, enzymatic reactivity and regioselectivity. ILs capable of solubilizing a large quantity of various carbohydrates, including mono-, oligo-, and polysaccharides, have been selected or designed; the water-mediated supersaturation method has been developed to dramatically increase the sugar solubility in the IL systems; high production of sugar esters has been achieved in a pure IL (e.g., [BMIm][TfO]), in IL mixtures (e.g., [BMIm][TfO]/[OMIm][Tf₂N]), or in IL/organic solvent bisolvent systems (e.g., [BMIm][PF₆]/t-BuOH, [BMIm][BF₄]/DMSO); and rational design of ILs that can both dissolve carbohydrates and also promote enzyme activity has been explored. However, we have to admit that the research in this field is only at an early stage undertaken in laboratories, and there are still quite a few fundamental and practical problems to be solved before it can eventually be industrialized.

The most challenging issue here is still to search for an appropriate solvent. The general criteria for such selection should include low cost, low viscosity, recyclability, greenness, and biocompatibility; but more specifically for sugar ester synthesis, the suitable ILs should not only be able to well dissolve the substrates, i.e., carbohydrates and fatty acids or their vinyl esters, but also be enzyme-friendly, in favor of enzyme activity, stability, and regioselectivity.

Early studies regarding the selection of ILs for biocatalytic reactions were more exploratory in nature, which obviously cannot meet the requirements of increasing interests in using ILs. Much to our delight, as the knowledge about the relationship between IL structures and properties accumulates, utilization of ILs for biotransformations has been switching from random selection to rational design, and a few guidelines for IL design have already been proposed.³³ Zhao's study²⁰ regarding the design of enzyme-compatible ILs with good sugar solubility has provided an excellent example for this. As has been pointed out in our book chapter,³³ in order to take full advantage of the “designer solvent” property of the ionic liquids, more work has to be done to allow us to have a thorough understanding of the following relationships: (1) the correlation between the structure of an IL and its physicochemical properties; (2) the correlation between the IL structure and its interactions with enzyme molecules; and (3) the correlation between the IL structure and enzyme functions. Nevertheless, one has to keep in mind that biocatalysis is a complicated system involving so many different interactions. In addition to the general solvent properties of the ILs, factors specific to the enzyme and the components involved in the reaction system may also be very important in affecting the enzymatic reaction, depending very much on their effects on substrate dissolution, substrate–solvent–enzyme interactions, and conformation, dynamics, active site structure, and catalytic mechanism of each specific enzyme. In the specific case of sugar ester synthesis, the impacts of ILs on both the lipase catalysis and sugar dissolution have to be considered. For instance, in all the four imidazolium ILs ([BMIm][Ac], [BMIm][lactate], [BMIm][NO₃] and [BMIm][dca]), CALB can dissolve, but showed a much lower activity than in [BMIm][BF₄], and the activity loss in the former three ILs was partially reversible while that in [BMIm][dca] was irreversible; but in [Et₃MeN][MeSO₄], CALB can not only dissolve but also preserve both its activity and native secondary structure.³² Additionally, evidence has shown that while H-bond forming functionality on either the cation or anion may favor the sugar dissolution due to its ability to form H-bonds with the sugar molecules, it may also be detrimental to the enzyme by competing for the H-bonds originally present in the enzyme molecules, thus causing deactivation and/or denaturation. Whether this impact is more anion- or cation-dependent is still in question.

In the consideration of IL design, one has to pay special attention to deep eutectic solvents (DESs), a new class of ionic solvents prepared by complexation of a quaternary ammonium salt (e.g., choline chloride) with a H-bond donor (e.g. amide, amine, alcohol and carboxylic acid), the development of which was pioneered by Abbott et al.³⁴ Although holding a molecular component, DESs share many attractive solvent features of conventional cation–anion paired ionic liquids, such as melting points below room temperature, low volatility, and high thermal stability. The interaction of the H-bond donor with the anion reduces the original cation–anion electrostatic force, thus inducing a depression in the melting point of the mixture. An exemplary DES is the eutectic mixture of choline chloride (mp 302 °C) and urea (mp 133 °C) in a 1 : 2 molar ratio with a low freezing point of 12 °C.³⁴ DESs have been regarded as promising green replacements for volatile organic solvents and also conventional first- and second-generation ionic liquids, because they are easily prepared with inexpensive, sustainable, non-toxic, and biodegradable components.^35,36 Importantly, DESs are capable of dissolving a variety of substances including organic acids and carbohydrates, making them attractive as a solvent for many applications including carbohydrate chemistry. Moreover, DESs can be obtained by a broader selection than the conventional ILs, in terms of different organic salts and H-bond donors in different ratios, thus making this new type of designer solvents even more designable.

As the first proof of concept, Gorke et al.³⁷ reported that several hydrolases (Candida antarctica lipase A and B, and epoxide hydrolase) presented high activity in DESs composed of choline chloride or ethylammonium chloride paired with H-bond donors such as acetamide, ethylene glycol, glycerol, urea, and malonic acid. These DESs can also work as a cosolvent in aqueous solution to enhance the hydrolase-catalyzed reactions up to 20-fold. The use of DESs as cosolvent in aqueous solution for epoxide hydrolase has also been examined by Lindberg et al.³⁸ Additionally, Zhao and co-workers have demonstrated that the DESs based on choline acetate and choline chloride coupled with glycerol have favorable properties including low viscosity (79–93 cP at 50 °C) and excellent compatibility with lipase (Novozym 435)³⁹ and two proteases (subtilisin and α-chymotrypsin).⁴⁰ Although only at a very early stage of development, these encouraging results have suggested that DESs merit serious consideration as a green solvent for enzymatic processes, including the synthesis of sugar fatty acid esters.

On the other hand, there are also practical challenges to be concerned with the enzymatic synthesis of sugar fatty acid esters. The synthetic processes need to be optimized, in terms of enzyme and solvent selection, reaction conditions, and reactor design. It is also imperative to develop effective methodologies for substrate dissolution, product separation, IL recycling, and enzyme reuse. Although the water-mediated supersaturation method developed by Lee et al.¹⁹ appears to be a good way to enhance the sugar solubility in the reaction medium, it may be costly and time-consuming. The sugar ester products can be separated by extraction with organic solvents, but this brings back the greenness issue because one advantage of using ILs is to avoid the need for organic solvents. Ionic liquids and immobilized enzymes have to be recycled in order to reduce the entire processing cost, but so far there do not seem to be any methods available which are practically feasible for these jobs, especially in consideration of scaling up and continuous processing. In fact, in addition to the relatively high cost, the rather high viscosity of ILs is also a big obstacle to hamper the development of carbohydrate biotransformations in these reaction media: not only limiting the mass transfer in the reaction system, thus affecting the reactivity, but also burdening the already problematic situations such as those listed above. While attempt to design less viscous ILs may help to circumvent this problem, a direct solution for this is still through processing designs.

Overall, ionic liquids represent a promising new reaction medium for enzymatic synthesis of sugar fatty acid esters. Research on IL design, especially on the exploration of the newly emerging deep eutectic solvents, will not only facilitate this application but also offer us the insight into the fundamentals behind biocatalysis in ionic liquids in general and may also open up a bright future for these enzymatic processes. Overcoming the challenges that have been posed above can surely improve their utility for sugar ester synthesis and also other biocatalytic transformations as well.

Acknowledgements

We thank Prof. Min-Hua Zong of South China University of Technology, China, for helpful discussion.

References

1. T. Kobayash, Biotechnol. Lett., 2011, 33, 1911–1919.
2. T. Plat and R. J. Linhardt, J. Surfactants Deterg., 2001, 4, 415–421.
3. A. M. Gumel, M. S. M. Annuara, T. Heidelberg and Y. Chisti, Process Biochem., 2011, 46, 2079–2090.
4. I. Ikeda and A. M. Klibanov, Biotechnol. Bioeng., 1993, 42, 788–791.
5. G. Fregapane, D. B. Sarney, S. G. Greenberg, D. J. Knight and E. N. Vulfson, J. Am. Oil Chem. Soc., 1994, 71, 87–91.
6. F. Theil and H. Schick, Synthesis, 1991, 533–535.
7. L. Cao, U. T. Bornscheuer and R. D. Schmid, Fett/Lipid, 1996, 98, 332–335.
8. Z. Yang and W. Pan, Enzyme Microb. Technol., 2005, 37, 19–28.
9. F. van Rantwijk and R. A. Sheldon, Chem. Rev., 2007, 107, 2757–2785.
10. M. Moniruzzaman, K. Nakashima, N. Kamiya and M. Goto, Biochem. Eng. J., 2010, 48, 295–314.
11. H. Zhao, J. Chem. Technol. Biotechnol., 2010, 85, 891–907.
12. R. M. Lau, F. van Rantwijk, K. R. Seddon and R. A. Sheldon, Org. Lett., 2000, 2, 4189–4191.
13. S. Park and R. J. Kazlauskas, J. Org. Chem., 2001, 66, 8395–8401.
14. N. Kimizuka and T. Nakashima, Langmuir, 2001, 17, 6759–6761.
15. Q. Liu, M. H. A. Janssen, F. van Rantwijk and R. A. Sheldon, Green Chem., 2005, 7, 39–42.
16. S. A. Forsyth and D. R. MacFarlane, J. Mater. Chem., 2003, 13, 2451–2456.
17. R. P. Swatloski, S. K. Spear, J. D. Holbrey and R. D. Rogers, J. Am. Chem. Soc., 2002, 124, 4974–4975.
18. Z.-G. Chen, M.-H. Zong and G.-J. Li, J. Chem. Technol. Biotechnol., 2006, 81, 1225–1231.
19. S. H. Lee, D. T. Dang, S. H. Ha, W.-J. Chang and Y.-M. Koo, Biotechnol. Bioeng., 2008, 99, 1–8.
20. H. Zhao, G. A. Baker, Z. Song, O. Olubajo, T. Crittlea and D. Peters, Green Chem., 2008, 10, 696–705.
21. T. G. A. Youngs, John D. Holbrey, M. Deetlefs, M. Nieuwenhuyzen, M. F. Costa Gomes and C. Hardacre, ChemPhysChem, 2006, 7, 2279–2281.
22. T. G. A. Youngs, C. Hardacre and J. D. Holbrey, J. Phys. Chem. B, 2007, 111, 13765–13774.
23. R. C. Remsing, R. P. Swatloski, R. D. Rogers and G. Moyna, Chem. Commun., 2006, 1271–1273.
24. D. R. MacFarlane, J. Golding, S. Forsyth, M. Forsyth and G. B. Deacon, Chem. Commun., 2001, 1430–1431.
25. S. A. Forsyth, D. R. MacFarlane, R. J. Thomson and M. von Itzstein, Chem. Commun., 2002, 714–715.
26. M.-J. Kim, M. Y. Choi, J. K. Lee and Y. Ahn, J. Mol. Catal. B: Enzym., 2003, 26, 115–118.
27. F. Ganske and U. T. Bornscheuer, Org. Lett., 2005, 7, 3097–3098.
28. F. Ganske and U. T. Bornscheuer, J. Mol. Catal. B: Enzym., 2005, 36, 40–42.
29. S. H. Ha, N. M. Hiep, S. H. Lee and Y.-M. Koo, Bioprocess Biosyst. Eng., 2009, 33, 63–70.
30. S. H. Lee, H. M. Nguyen, Y.-M. Koo and S. H. Ha, Process Biochem., 2008, 43, 1009–1012.
31. E. Abdulmalek, H. S. M. Saupi, B. A. Tejo, M. Basri, A. B. Salleh, R. N. Z. R. A. Rahman and M. B. A. Rahman, J. Mol. Catal. B: Enzym., 2012, 76, 37–43.
32. F. van Rantwijk, F. Secundo and R. A. Sheldon, Green Chem., 2006, 8, 282–286.
33. Z. Yang, in Ionic Liquids in Biotransformations & Organocatalysis Solvents and Beyond, ed. P. Domínguez de María, John Wiley, New York, 2011, ch. 2 (in press).
34. A. P. Abbott, G. Capper, D. L. Davies, R. K. Rasheed and V. Tambyrajah, Chem. Commun., 2003, 70–71.
35. J. Gorke, F. Srienc and R. Kazlauskas, Biotechnol. Bioprocess Eng., 2010, 15, 40–53.
36. P. Domínguez de María and Z. Maugeri, Curr. Opin. Chem. Biol., 2011, 15, 220–225.
37. J. Gorke, F. Srienc and R. Kazlauskas, Chem. Commun., 2008, 1235–1237.
38. D. Lindberg, M. de la Fuente Revenga and M. Widersten, J. Biotechnol., 2010, 147, 169–171.
39. H. Zhao, G. A. Baker and S. Holmes, J. Mol. Catal. B: Enzym., 2011, 72, 163–167.
40. H. Zhao, G. A. Baker and S. Holmes, Org. Biomol. Chem., 2011, 9, 1908–1916.

---