Enantioselective C-C Bond Formation Enhanced by Self-Assembly of Achiral Surfactants.

The use of achiral surfactant assemblies as a reaction platform for an alkylation reaction resulted in a high enantiomeric excess. Dilauryldimethylammonium bromide (DDAB) vesicles were modified with cholesterol to promote alkylation of N-(diphenylmethylene)glycine tert-butyl ester (DMGBE) with benzyl bromide, resulting in high conversion (∼90%) and high enantioselectivity (up to 80%). The R-enantiomer was formed on using the DDAB vesicles, whereas the use of phospholipid liposomes prepared from 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) produced an excess of the S-enantiomer. Considering the chemical structures of the reaction substrates and amphiphiles as well as the membrane structures and properties of DDAB vesicles and DOPC liposomes, it is suggested that the enantiomeric excesses result from the location of the quaternary amine of the amphiphiles and the DMGBE at the outer surface of the membrane. We show that the enantioselective reaction at the surface of the self-assembly could be regulated by adjusting the chemical structures and resulting membrane properties of the self-assembly.

Introduction

C—C bond formation reactions are fundamental in many areas, like petrochemicals, materials science, and pharmaceutical chemistry.[1−3] Enantioselective C—C bond formations have attracted much attention from many scientists because these reactions can be an important tool for synthesizing pharmaceutical chemicals. There have been many reports on the development of chemical processes to obtain products enantioselectively, mostly focusing on the design of catalysts.[4−6] However, conventional enantioselective reactions need to be carried out in organic solvents as reaction media and have several reaction steps for synthesizing catalysts, which require additional energy for the whole reaction process. Therefore, there is a demand for enantioselective reaction processes that function in aqueous media.

Some difficulties must be overcome to achieve enantioselective reactions in water. Recently, powerful phase-transfer catalysts have been developed for enantioselective alkylation, utilizing the interface between the aqueous and organic phases.[7−9] However, many catalysts are unable to work in aqueous media; only a few metallic catalysts show activity in highly polar environments[10] and many organic reactants do not dissolve in water. Many efforts have been made to synthesize designed catalysts because of difficulties such as low yields and a lack of catalyst versatility because the mechanisms of the enantioselective reactions are unclear. Instead, regulation of the nature of the interface between aqueous and organic phases may positively affect interfacial reactions and improve the understanding of the mechanism of these reactions.

Self-assembling systems such as micelles, vesicles, and nanotubes can provide the microenvironment of an organic phase within aqueous media, in which reactions between organic species can occur.[11−13] It has also been reported that liposomes can be utilized as a platform for chemical processes that can achieve separation and conversion of reactants in aqueous media.[16,17] Our previous reports have shown that liposome membranes could enhance the enantioselective alkylation of an amino acid derivative.[18] In the previous report, it has also been revealed that the chirality of the product is independent of the chirality of the lipids (l-form or d-form) used for construction of the platform, suggesting that regulation of the reaction environment is more important than the chirality of the molecules. Assuming that the ability to localize reactants within the bilayer of a liposome is an essential requirement for a successful enantioselective reaction, a vesicle composed of achiral surfactants can also be used as a platform for an enantioselective reaction. Previous reports have shown that some reactions are enhanced with surfactants at the water–organic solvent interface[23] or with surfactants self-assembled as micelles, which act to enrich the concentration of reactants at the self-assembly surface.[14,15] It has also been reported that vesicles can be constructed using surfactants and surfactant vesicles can also work as a platform for some reaction processes.[19] Although these reports showed good examples of reactions with high yields, the possibility that achiral surfactants could work as catalysts for enantioselective reactions has never been suggested. Therefore, it would be very valuable if an enantioselective reaction can be carried out without chiral catalysts. Herein, for the first time, we propose an enantioselective reaction system with vesicles of dilauryldimethylammonium bromide (DDAB), which contains a quaternary ammonium moiety to enhance the reaction and no asymmetric carbon atom. The reaction of N-(diphenylmethylene)glycine tert-butyl ester (DMGBE) with benzyl bromide utilizing DDAB vesicles is investigated for the ability of vesicles constructed from achiral molecules to produce an enantioselective product and to reveal the mechanism of obtaining enantioselectivity. Further, we compared the surface properties of a DDAB vesicle with those of other self-assembling systems (liposomes and micelles, Figure ) to investigate a possible reaction model that regulates the observed conversion and enantioselectivity.

Figure 1

Top: Illustrations of liposome, vesicle, and micelle. Bottom: chemical structures of DDAB, 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), and cetyltrimethylammonium bromide (CTAB). A liposome is constituted of natural amphiphilies (often phospholipids, such as DOPC and DPPC), and in this study, vesicles composed of synthetic, non-natural amphiphiles (such as DDAB) are called “vesicles”. CTAB is used for the formation of micelles.

Results and Discussion

We have previously reported that the alkylation of DMGBE can be achieved (conversion >90%) at a low enantiomeric excess (e.e. ≈ 5%) in a CTAB micelle solution, whereas a higher e.e. value (>90%) could be obtained in a DOPC liposome suspension.[18] DDAB was selected as an amphiphile that can form vesicles in an aqueous solution and possesses a tertiary ammonium group that is located on the hydrophilic side of the membrane.[20] Alkylation of DMGBE was carried out in a suspension of DDAB vesicles, as shown in Scheme . This reaction showed a high conversion rate (∼90%) and medium enantioselectivity (e.e. ≈ −50%) based on a high-performance liquid chromatography (HPLC) chromatogram obtained using a chiral column (Figure , Table ). Most importantly, the (R)-product was formed with the DDAB vesicles. The (R)-product was also obtained by using the phase-transfer catalyst,[8] whereas the (S)-product was formed in excess in the presence of DOPC liposomes.[18] The size-effect of the DDAB vesicles was also examined, but no significant differences in conversion and e.e. were observed, indicating that the formation of the vesicles was an important factor for the reaction. When the DDAB vesicles were modified with 20% cholesterol, the e.e. values increased compared to those of the unmodified DDAB vesicles, reaching approximately −80%. The reactants were added to the DDAB vesicle suspension in different orders to examine the effect of partitioning of the reactants in the DDAB vesicle membrane. Insertion of benzyl bromide into the DDAB vesicle membrane prior to the addition of DMGBE resulted in the formation of the (R)-product, with an e.e. value that reached −56%. These results support the theory that benzyl bromide can be oriented in the deeper region of the DDAB vesicle membrane and can react with the DMGBE molecule at the outer surface of the membrane.

Scheme 1

Asymmetric Reaction of DMGBE with Benzyl Bromide

1: DMGBE, 2: benzyl bromide, and 3: product.

Figure 2

HPLC chromatogram of the reaction with each vesicle. aFrom our previous report; see ref (18) for detailed information. bInitially, benzyl bromide was added to the vesicle suspension and mixed, and then, the mixture was incubated for 1 h. The reaction was initiated by adding NaOH (aq) and DMGBE. cThe vesicle composition was DDAB/cholesterol = 80:20. Chol., cholesterol.

Table 1

Alkylation of Amino Acid Derivative Using Various Kinds of Amphiphiles

entry	amphiphile	conversion (%)	e.e. ((S) – (R)%)
1	DDAB (vesicle, d = 100 nm)	90 ± 3	–49 ± 13
2	DDAB (vesicle, d = 50 nm)	91 ± 4	–47 ± 7
3	DDAB (vesicle, d = 200 nm)	94 ± 3	–44 ± 12
4	DDAB/chol. = 80:20 (vesicle)	87 ± 2	–77 ± 6
5a	DDAB (vesicle)	88 ± 4	–56 ± 8
6b	CTAB (micelle)	95 ± 2	5 ± 3
7b	DOPC (liposome)	62 ± 2	97 ± 1

Benzyl bromide was initially added and mixed with the vesicle suspension and then incubated for 1 h. The reaction was initiated by adding NaOH(aq) and DMGBE.

From our previous report; see ref (18) for detailed information.

The conformation of DMGBE during its interaction with the DDAB vesicle membrane should differ from that on interaction with the DOPC liposome membrane, as previously reported.[18] The physicochemical properties of the membrane are important to the outcome of a reaction on the membrane of either a vesicle or a liposome. The surface properties of various self-assemblies were analyzed using the previously reported method[21,22] to investigate the differences between the DOPC liposomes, CTAB micelles, and DDAB vesicles. The fluorescent probes 6-lauroyl-2-dimethylamino naphthalene (Laurdan) and 1,6-diphenyl-1,3,5-hexatriene (DPH) have been used to evaluate the membrane polarity and fluidity, respectively.[21] The spectrum of Laurdan is strongly related to the hydration environment of the Laurdan embedded in the membrane: the spectrum of Laurdan is blue-shifted when its environment is dehydrated.[22] Therefore, the surface characteristics of the self-assembling system can be distinguished from those of others by analyzing the fluorescence spectra of Laurdan. Laurdan shows a peak at 440 nm when embedded in vesicles with a solid-ordered (so) phase, at 490 nm when in vesicles with liquid-disordered (ld) phase, and at 505 nm for micelles. Figure shows the fluorescence spectra of Laurdan embedded in a typical liposome (phospholipid vesicle), micelle, and a DDAB vesicle. The fluorescence spectrum of Laurdan in the presence of a DDAB vesicle indicates that the DDAB membrane has quite different characteristics from those of a micelle, showing it to be more hydrated than typical ld phase vesicles. The membrane polarity can be analyzed in detail by calculation of its GP340 values; vesicles in the so phase (such as DPPC liposomes) showed high GP340 values of ∼0.4, whereas the GP340 values were about −0.3 for ld phase vesicles, such as DOPC liposomes (Figure S1a). The GP340 value for a micelle was much lower (GP340 < −0.7), indicating that the micelle surface is a hydrated environment. The DDAB vesicles also showed low GP340 values (GP340 ∼ −0.6), indicating that DDAB vesicles also form highly hydrated environments, such as those found in micelles. These results were also proved by the evaluation of the membrane fluidity, which was analyzed by the fluorescent probe DPH (Figure S1b). The membrane fluidity (1/P) of the DDAB vesicles was much higher than that of conventional vesicles (such as DOPC liposomes).

Figure 3

(a) Fluorescence spectra of Laurdan embedded in each self-assembly system. (b) DDAB vesicle, (c) DDAB vesicle + benzyl bromide, and (d) DDAB/cholesterol = 80:20 vesicle. The lines indicate the experimental result (black), fitted line (red), spectral fraction of the ordered phase (blue), and spectral fraction of the disordered phase (green).

The addition of benzyl bromide to the DDAB vesicle caused the fluorescence spectrum of Laurdan and the value of GP340 to vary (Figure b,c), whereas the addition of benzyl bromide to the DOPC liposome did not change its membrane properties. The membrane of the DDAB vesicle changed to a more “ordered” structure through its interaction with benzyl bromide, indicating that the hydrophobic molecules, used as reaction substrates, can change the membrane properties of DDAB vesicles. In addition, modification of the DDAB vesicle membrane with cholesterol also resulted in a similar variation of the membrane, as seen for benzyl bromide. These findings indicate that the membrane properties of the DDAB vesicles can be regulated by the addition of hydrophobic molecules. However, DDAB vesicles modified with either benzyl bromide or cholesterol had more polar environments than those of the liposomes, indicating that the DDAB vesicles act as a platform different from those of liposomes. This finding can be taken as an explanation for the observed variations in the enantioselectivity of reactions on using these platforms.

The dependence of enantioselective alkylation of DMGBE on the membrane properties of the self-assemblies was studied by direct comparison of the obtained results (Figure ). The conversion values were plotted against the GP340 values, showing an approximately linear relationship (Figure a). It can be clearly seen that the self-assembling systems that showed lower (more negative) GP340 values could enhance the conversion in the reaction. The reason for this enhancement is probably related to the adsorption of the reactants onto the surface (membrane) of the self-assembling system. Our previous reports have shown that the self-assembly systems with lower GP340 values could attract a greater number of hydrophobic substrates due to the exposure of the hydrophobic interior of the vesicle membrane.[16] The adsorption behavior of the reactants was also investigated in this study. It was found that a greater quantity of reactants was adsorbed onto the self-assemblies with lower GP340 values (Figure a). The adsorption of DMGBE was strongly related to the conversion, indicating that the interaction between DMGBE and self-assembly was an important factor for the reaction. Although the e.e. values were low (∼5%) in the presence of systems with lower GP340 values, such as CTAB micelles, some liposomes that showed higher GP340 values achieved high enantioselectivity (e.e. ∼90%; Figure b). The conversion and enantioselectivity were, respectively, suppressed and increased to some extent when benzyl bromide was inserted into the DDAB vesicle prior to DMGBE addition. This effect was probably because of the associated change in the GP340 value. These findings revealed the importance of the properties of self-assembling systems in highly effective reactions on the surface of self-assemblies.

Figure 4

Effects of polarity on the reaction and adsorption of the reactants onto self-assembly systems. (a) Top: conversion of the reaction, bottom: amount of DMGBE absorbed. (b) e.e. values of the product. The values of conversion and e.e. of DOPC and DPPC (labeled 4 and 6, respectively) were from ref (18).

We suggest that the reaction model for enantioselective alkylation at the surface of self-assemblies is also related to the membrane properties and chemical structures of the reactants and amphiphiles. As previously reported in phase-transfer catalysis, the Si face of DMGBE tends to interact with and be covered by the hydrophobic region of the catalyst, which resulted in the formation of (R)-form enantiomers.[8] On applying this tendency to the DDAB vesicle system, the Si face of DMGBE faces toward the hydrophobic region of the DDAB vesicle and is hindered by the hydrophobic and electrostatic interactions between DDAB and DMGBE. The DDAB–DMGBE complex formed shows high hydrophobicity on the basis of its chemical structure, and the complex can enter into a deeper area of the DDAB vesicle to form stronger hydrophobic interactions (Figure ). As a result, benzyl bromide only can react with the Re face of DMGBE, which can produce the (R)-form of the product. This simple interaction between DDAB and DMGBE could be supported by the time required to accomplish the reaction. It took only a short time (within a few hours) to complete the reaction as compared to that for the reaction with the DOPC liposome. DOPC and DMGBE might form complicated reaction intermediates via multiple forces, such as (i) electrostatic interactions, (ii) hydrophobic interactions, and (iii) hydrogen bonds of both molecules at the interface. In our previous report, multiple interactions were suggested to be induced in the case of chiral adsorption of l-amino acids onto the phospholipid membrane surface, in which an induction time is needed to obtain “mature” interactions at the liposome surface.[17] Similarly, the multiple interactions between the DMGBE and DOPC molecules are thought to be rather complicated and therefore it could require more than several hours to form an appropriate reaction intermediate to obtain the chiral ((S)-form) product (Figure S3). The difference in interactions is considered to have varied the covered plain face of DMGBE, which results in the production of opposite enantiomers.

Figure 5

Reaction model of the enantioselective alkylation of DMGBE (shown in its depronated form) on the surface of a DDAB vesicle membrane. In this figure, the Re face of DMGBE is on the top (outside of the vesicle membrane) and the Si face is at the bottom (inside of the vesicle membrane).

On modification of the DDAB vesicle with cholesterol, the value of GP340 increased, as mentioned above, which is related to the packing density of the vesicle. The steric hindrance for the attack of the Si face of DMGBE by benzyl bromide became larger because the packing density of DDAB increased in the presence of cholesterol. This effect of cholesterol leads to the enhanced enantioselectivity of the reaction. In general, the properties of the self-assembling system can regulate the conversion and enantioselectivity of interfacial alkylation reactions at the surface of the assembly.

Conclusions

Enantioselective alkylation was carried out by utilizing achiral surfactant vesicles as a platform. The experiments suggested that the essential factor for obtaining the enantioselectivity by the reaction at the vesicle surface was not the chirality of the vesicles and vesicle components but regulation of the reaction environment and interaction between vesicle components and the reactants. The membrane properties of the self-assembling system and, particularly, its polarity have a strong regulatory influence on the interaction of the platform with the reactants, which eventually influences the conversion and enantioselectivity. A reaction mechanism is proposed by considering the plausible interactions between the surfactant (or lipid) and reactants to explain the result of different enantiomers obtained by using the DDAB vesicle and liposomes. It is now expected that an interfacial enantioselective reaction could be regulated by adjusting the membrane properties of vesicles and the interactions, providing a new and clean process for asymmetric synthesis.

Experimental Methods

Materials

DPPC and DOPC were purchased from Avanti Polar Lipid (Alabaster, AL). Cholesterol was purchased from Sigma Aldrich (St. Louis, MO). CTAB, sodium hydroxide, and benzyl bromide were purchased from Wako Pure Chemicals (Osaka, Japan). DDAB and DMGBE were purchased from Tokyo Chemical Industries (Tokyo, Japan). These chemicals were used without further purification.

Preparation of Vesicles

A chloroform solution containing lipids was dried in a round-bottom flask by evaporation under vacuum. The obtained lipid thin film was dissolved in chloroform again, and the solvent was evaporated. The lipid thin film was kept under high vacuum for at least 3 h and was then hydrated with distilled water at room temperature. The liposome suspension was frozen at −80 °C and thawed at 50 °C to enhance the transformation of small vesicles into larger multilamellar vesicles (MLVs). This freeze–thaw cycle was performed five times. MLVs were used to prepare large unilamellar vesicles by extruding the MLV suspension 11 times through 2 layers of polycarbonate membranes, with different mean pore diameters (50, 100, and 200 nm), using an extruding device (Liposofast; Avestin Inc., Ottawa, Canada).

Evaluation of Membrane Polarities

Laurdan is sensitive to the polarity around the molecule itself, and its fluorescence properties enable the evaluation of the surface polarity of lipid membranes. The Laurdan emission spectra exhibit a redshift caused by dielectric relaxation. The emission spectra were measured with an excitation wavelength of 340 nm, and the general polarization (GP340) or membrane polarity was calculated as followswhere I440 and I490 represent the fluorescence intensity of Laurdan at 440 and 490 nm, respectively. The total concentrations of lipid (or surfactant) and Laurdan were 1 mM and 10 μM, respectively. For the CTAB micelle, the concentration of CTAB and Laurdan were 10 mM and 100 μM due to the critical micelle concentration of CTAB (∼1 mM).

The fluorescence spectrum of Laurdan was deconvoluted into two spectra using Peakfit software (Systat Software Inc., CA): one originated from the localization of Laurdan in an ordered membrane (ordered phase) and the other originated from the localization of Laurdan in a disordered membrane (disordered phase).

Alkylation of DMGBE with Benzyl Bromide

DMGBE (0.3 mg (1.0 μmol)) was dissolved in pure water, and then, the aqueous solution was mixed with the vesicle suspension and NaOH aqueous solution (10%) to obtain 1 mL of reaction solution. Benzyl bromide (1.5 μL (12 μmol)) was finally added to initiate the reaction (Scheme ). The reaction solution was stirred at 500 rpm at room temperature for 24 h. The total concentrations were 1.0 mM DMGBE, 12 mM benzyl bromide, 10 mM lipid, and 0.3 M NaOH. In the case of micelles, instead of liposomes, an aqueous solution of CTAB was added. The total concentrations were fixed to the same values as those for the reaction with liposomes.

HPLC Measurements of the Reaction Solution

After the reaction completed, the product (and reactant) was extracted to the organic solvent using the Bligh–Dyer method. In brief, 1 mL of the reaction solution was mixed with 2 mL of methanol and 1 mL of chloroform, resulting in a homogeneous, colorless, and transparent liquid. Then, 1 mL each of chloroform and pure water were added to the solution to lead to a phase separation. Centrifugation was performed (1400 rpm, 5 min) to complete the phase separation using a Tabletop Centrifuge, KUBOTA 5200 (Kubota, Tokyo, Japan).

After the extraction, the organic phase was moved into a round-bottom flask and chloroform was removed by evaporation. Diethyl ether (1 mL) was added to the flask, and 10 μL was taken and dissolved in 1 mL of the mobile phase for HPLC analysis. HPLC analysis was carried out using a Warers 1515 Isocratic HPLC Pump and Waters 2489 UV/Visible Detector (Waters, MA). To evaluate the conversion of the reaction and enantiomeric excess, a HPLC column Chiralpak IA (Daicel Corp., Osaka, Japan) was used. The mobile phase was hexane/2-propanol = 99:1, flow rate = 0.5 mL/min.

Adsorption of Reactants onto Vesicle

The adsorption amounts of the reactant (DMGBE) were evaluated by a UV spectrophotometer (UV-1800; Shimadzu, Kyoto, Japan). The total concentration of the reactant and vesicles were same as those in the alkylation method (DMGBE = 1.0 mM, vesicle = 12 mM, NaOH = 0.3 M). The solution was incubated for an hour at room temperature with 500 rpm stirring. The adsorption percentages were calculated from the difference in UV absorbance of the solutions without and with vesicleswhere Ainitial and Afiltrated represent the absorbance of the reactant in the NaOH aqueous solution and in the vesicle suspension, respectively.