The Effect of Hydrophilic Ionic Liquids 1-Ethyl-3-Methylimidazolium Lactate and Choline Lactate on Lipid Vesicle Fusion

Ionic liquids (ILs) are room-temperature molten salts that have applications in both physical sciences and more recently in the purification of proteins and lipids, gene transfection and sample preparation for electron microscopy (EM) studies. Transfection of genes into cells requires membrane fusion between the cell membrane and the transfection reagent, thus, ILs may be induce a membrane fusion event. To clarify the behavior of ILs with cell membranes the effect of ILs on model membranes, i.e., liposomes, were investigated. We used two standard ILs, 1-ethyl-3-methylimidazolium lactate ([EMI][Lac]) and choline lactate ([Ch][Lac]), and focused on whether these ILs can induce lipid vesicle fusion. Fluorescence resonance energy transfer and dynamic light scattering were employed to determine whether the ILs induced vesicle fusion. Vesicle solutions at low IL concentrations showed negligible fusion when compared with the controls in the absence of ILs. At concentrations of 30% (v/v), both types of ILs induced vesicle fusion up to 1.3 and 1.6 times the fluorescence intensity of the control in the presence of [Ch][Lac] and [EMI][Lac], respectively. This is the first demonstration that [EMI][Lac] and [Ch][Lac] induce vesicle fusion at high IL concentrations and this observation should have a significant influence on basic biophysical studies. Conversely, the ability to avoid vesicle fusion at low IL concentrations is clearly advantageous for EM studies of lipid samples and cells. This new information describing IL-lipid membrane interactions should impact EM observations examining cell morphology.


Introduction
Ionic liquids (ILs), molten salts with low melting temperatures (<100 °C), are a complex of anionic and cationic ions. ILs have unique characteristics, including high ionic conductivity, a wide range of viscosities and are nonvolatile. These features make ILs suitable for many applications such as surface finishing, electrochemical deposition, electroplating and formation of carbon nanotubes. Recently, ILs have been used in analytical biochemistry examining proteins and lipids by replacing conventionally used organic solvents such as methanol/ chloroform and acetonitrile. For example, ILs provide matrices for matrix-assisted laser desorption and ionization (MALDI) of biomolecules in mass spectroscopy [1] and they improve the homogeneity of ionized molecules and ion yield [2,3]. ILs have also been used to enhance the separation of proteins [4] in sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), chromatography techniques including high performance liquid chromatography (HPLC) [5], hydrophobic interaction chromatograph [6], gas chromatography [7,8]. Furthermore, if suitable preparation conditions can be found that ensure regular, high quality crystals then various soluble and transmembrane proteins can be examined, thus promoting protein structure-function studies. However, preparing high quality protein crystals with good diffraction properties for better 3D structure estimation is often a difficult task. In this point, ILs also can provide better conditions for protein crystallization [9]. In addition, ILs have been studied as gene transfection reagents [10]. This application suggests that ILs may also aid membrane fusion.
Polyethylene glycols (PEGs) show features that are similar to the interactions that ILs make with biomolecules. PEGs are polymers of ethylene glycol and have average molecular weights of 180−3,500,000. PEG has been a common reagent used for protein precipitation and purification of proteins by liquid chromatography-mass spectrometry, because of their ability to induce liquid biphases [4]. PEG improves the protein crystallization process by eliminating water between protein molecules and the inner structures of proteins. PEGs are also useful for purification of antibodies [11], DNA precipitation [12] and are also used as carriers for DNA transfection of cells [13], initiators of membrane fusion between cells [14,15] and lipid vesicle-vesicle fusion [16].

Lipid vesicle preparation
Large unilamellar vesicles were prepared according to established methods [18]. To 1:1 (mol/mol %) DOPC/SM in a 3:1 chloroform/methanol solution was added 30 mol% chol of DOPC+SM and 0.5 mol % of N-NBD-PE and N-Rh-PE. After the solvent was removed by a stream of N 2 gas, the lipids were dried overnight in a vacuum and then hydrated in a phosphate buffer (pH 7.4) at 50−55 °C for 1 h and converted to multilamellar vesicles by rigorous mixing. Large unilamellar vesicles were prepared from this solution by 21 extrusion cycles at 55 °C with an Avanti Polar Lipid extruder equipped with a 100-nm pore size polycarbonate membrane (Whatman, Florham Park, NJ). The final lipid concentration was 0.52 mM.

Fluorescence resonance energy transfer (FRET) analysis
FRET signals were detected by a Hitachi F4500 fluorescence spectrophotometer (Hitachi High-Technologies Corporation, Tokyo, Japan), and measurements were carried out as described previously [18][19][20]. Each sample consisted of two types of DOPC-SM-chol unilamellar vesicles. One of the vesicles contained two fluorescent probes (0.5 mol% of N-NBD-PE and N-Rh-PE), whereas the other vesicles contained no FRET probes. The FRET donor NBD-PE was excited at 473 nm and the emission intensity of the Rho-PE as the acceptor were monitored at 580 nm. All FRET measurements were carried out at room temperature (25-26 °C), as controlled by a recirculating chiller (Cool Ace CAE-1000A, Tokyo Rikakikai Co., Ltd., Tokyo, Japan).

Dynamic light scattering (DLS) measurements
Lipid vesicle sizes were obtained by 633 nm light scattering with a Malvern Zetasizer Nano series zen3600 (Malvern Instruments Ltd., Worcestershire, UK) at room temperature. We measured unilamellar vesicle sizes in the absence of ILs, followed by measurements at various concentrations of ILs. We averaged nine independent measurements of vesicle sizes, where each measurement was a mean of 13−15 readings by the instrument. All data were analyzed by the Zetasizer software (Malvern Instruments Ltd.). To correct the DLS data for changes in viscosity, we used a DV-II+ Pro viscometer (Brookfield Engineering Laboratories, Boston, MA) and measured the viscosity of the ILs at each concentration (10, 20 and 30%) and at room temperature. The torque conditions were set at 60 and 100 rpm, and we confirmed there was no difference between 60 and 100 rpm due to the IL viscosities. Size distributions of liposomes were fitted with the Origin ver. 7.5J software (Origin Lab, Northampton, MA). Normally, we cannot detect Rho-PE light emission because of quenching caused by the close proximity of the NBD-PEs and Rho-PEs within the same vesicle membrane ( Figure 2A). However, if vesicle fusion occurs between the labeled and unlabeled vesicles, the distance between the NBD-PE and Rho-PE is increased by the expansion of the vesicle surface area and the NBD-PE energized Rho-PE emission is no longer quenched. Thus, we observe an increase in Rho-PE emission due to IL-induced vesicle fusion ( Figure 2B). We investigated lipid membrane (vesicle) fusion by FRET at IL concentrations (volume/volume) of 10% ( Figure 3A  In Figure 4, the circles indicate the vesicle diameter distribution before the addition of the ILs; the peak of the distribution is 122.4 nm (Figures 4A, B) Figure 4A, black square) and 255 nm ( Figure 4B, black square)), respectively. When the concentrations of both ILs were increased from 10 to 30%, the vesicle diameters increased (supporting data). This is depicted in Figure 5, which shows the relative increase in peak size of the vesicles as a function of IL concentrations. Viscosities of ILs used in this experiment are very low, thus changes in vesicle sizes were primarily due to the addition of ILs. The DLS data support the expansion of the vesicles sizes by ILs, as observed in the FRET experiments. [Lac] not only exhibit similar effects on protein crystallization, but they also have similar Kamlet-Taft parameters (α, β and π * ) [21] (see Table 1). The α, β and π * define the hydrogen bond acidity, hydrogen bond basicity and dipolarity/polarizability effects, respectively, in solvents. The Kamlet-Taft parameters for the ILs are given in Table 1 and are estimated by reference to literature. Detailed information describing these parameters are found in some literatures [22][23][24]. PEGs induce lipid membrane fusion via water interactions and when the β value of ILs is >0.7 the ILs show properties that are similar to each other.

Thus, [EMI][Lac] and [Ch]
[Lac] probably interact with water via hydrogen bonding in a manner similar to PEGs. Generally, the molecular structure of an anion group has an effect on hydrogen bond donor ability [25]. In this study, we have used ILs with the same anion groups, so this feature is probably the reason that both ILs induced membrane fusion. However, [EMI][Lac] was found to show much greater ability to induce vesicle fusion than [Ch] [Lac]. This observation may be due to the different cationic groups, which have different hydrogen bonding acceptor abilities (β values).
Since lipid vesicles exist in water, the interaction of vesicles with PEGs and ILs probably involves similar bonding interactions and may be the basis in which ILs induce vesicle fusion. PEGs absorb and remove water from the surface of  vesicles. Consequently, the polar head groups of the outer lipids bond to each other. The inner lipids of the bilayer then fuse, followed by total membrane fusion [26]. If this mechanism is applicable to ILs, they must be in close proximity of the lipid polar head group-water interface. Evans et al. reported that the ILs induce the formation of a cholesterol-water interface in 1-

butyl-3-methylimidazolium chloride ([BMI]Cl) and 1-butyl-3methylimidazolium bis (trifluoromethylsulfonyl)imide ([BMI]
[Tf 2 N]) in a cholesterol bilayer solution [27]. This occurs even though the ILs have a different affinity owing to the anionic structures. Despite the fact that [BMI]Cl has a high β value and [BMI] [Tf 2 N] has a low β value [28], both ILs form a cholesterolwater interface. These observations suggest that [EMI] [Lac] and [Ch][Lac] will also localize around the lipid-water interface in a vesicle solution. Furthermore, another study reported that [EMI] ethyl sulfate in excess water interacts with water via the hydrogen atoms on the [EMI] imidazolium ring [29]. Taken  Figure 3 and Figure 4 revealed negligible vesicle fusion at low (10% v/v) concentrations for both ILs. Thus, the hydrogen bonding interaction between the C-H groups of [EMI] and water are weaker if there is excess water [29], and low concentrations of ILs will not absorb water surrounding the vesicles in sufficient quantities. Consequently, the probability of adhesion and fusion of the vesicles will be markedly reduced. Therefore, the ILs-water, ILs-lipid and lipidwater interactions depend on how much water is present. The subtle fluctuations in fluorescence intensities (data not shown) observed in the plateau region following the emission maxima may be due to perturbations of the lipid membrane caused by subtle changes in the hydrogen bonding forces in the ILswater-vesicle interaction.

Figure 5. Comparison of the vesicle diameters at various concentration of [Ch][Lac] and [EMI][Lac
Since PEGs have been used as a regent to induce cells/ membranes fusion, we measured how vesicle sizes changed by PEG (1.1mg/ml, [16] were mixed with lipid vesicle solutions, the PDI values were under 0.4 (primarily between 0.15−0.33) and we could carry out the DLS measurements successfully. This comparison also indicated PEGs cause significant aggregation with vesicles in their cluster. In contrast, both ILs used in this study do not induce aggregation like PEGs, even both ILs enlarged the size of the liposomes. These phenomena of PEG support the previous observation that PEGs cause aggregation of vesicles as an early process of phospholipid vesicle fusion [30] .
Recently, ILs have been used for preparing samples for electron microscopy (EM). ILs can prevent the electrization of samples by an electrification phenomenon known as "chargeup" on the surface of biological samples. This problem can be remedied by a simple procedure that introduces ILs to biomaterials to the sample, which allows significantly shorter preparation times. In addition, time-consuming metal coating procedures with Pt and/or C are no longer required when ILs are used. These benefits expand applications and increase the efficiency of sample preparations. Figure 6 shows TEM images using [EMI][Lac]-and [Ch][Lac]-liposome. Liposomes were extruded through a 100 nm pore membrane. Since the TEM environment is under a vacuum, it was not possible to directly compare the FRET, DLS and TEM imaging data under equivalent conditions. We found both ILs enlarged liposome sizes in TEM imaging, and the vesicle morphology was kept under vacuum conditions. From these additional results, the ILs that we used in this study should be available for liposome imaging by TEM.

Conclusions
This study is the first to report that high concentrations of ILs induce lipid membrane fusion without aggregation, as detected by both FRET and DLS. We found both ILs also enlarged liposome size by TEM. However, low concentrations of ILs hardly affected the vesicles and should represent suitable conditions for artifact-free preparation of EM samples [17]. EM is becoming a powerful tool for cell biology by imaging surface and intracellular structures at high resolution. Since ILs have high electrical conductivity [31], they behave like metal coatings for the imaging of cells. Furthermore, ILs emit secondary electrons more readily than metals, thus enabling secondary electron images without sample charging. Biological sample preparations are challenging, so it is anticipated that ILs will improve these technical issues associated with EM. Wider applications of ILs with a better understanding of their interactions with biomolecules could open new areas in the field of biological sciences. Therefore, future studies will investigate the relationship between ILs and biomaterials or biomolecules in order to expand the potential of EM.