Computational solvation analysis of biomolecules in aqueous ionic liquid mixtures

Computational solvation analysis of biomolecules in aqueous ionic liquid mixtures Based on their tunable properties, ionic liquids attracted significant interest to replace conventional, organic solvents in biomolecular applications. Following a Gartner cycle, the expectations on this new class of solvents dropped after the initial hype due to the high viscosity, hydrolysis, and toxicity problems as well as their high cost. Since not all possible combinations of cations and anions can be tested experimentally, fundamental knowledge on the interaction of the ionic liquid ions with water and with biomolecules is mandatory to optimize the solvation behavior, the biodegradability, and the costs of the ionic liquid. Here, we report on current computational approaches to characterize the impact of the ionic liquid ions on the structure and dynamics of the biomolecule and its solvation layer to explore the full potential of ionic liquids. Keywords Ionic liquid · Biomolecule · MD simulation Introduction in Fig. 1. The initial success of ionic liquids in (bio-) catalysis (van Rantwijk and Sheldon 2007) and solvation Ionic liquids (IL) are a unique class of high-performance (Welton 1999;Gutowskietal. 2003), e.g., dissolution of chemical compounds with many applications in electro- cellulose (Swatloski et al. 2002), served as a technology chemistry (Armand et al. 2009), synthesis (Itoh 2017), trigger. The vast number of 10 possible combinations of catalysis (Parvulescu ˆ and Hardacre 2007; van Rantwijk and cations and anions (Holbrey and Seddon 1999) seemed to Sheldon 2007) as well as solvation (Hallett and Welton promise the design of an optimal solvent for a particular 2011) and extraction processes (Ventura et al. 2017). Their application. As a result, the term “designer solvent” was tunable properties via variation or modification of either the coined (Freemantle 1998) and soon joined by the rash cation or the anion as well as their shared properties such as promise “ILs are green solvents” (Earle and Seddon 2009) low vapor pressure, low flammability, and high thermal and leading to an exponential number of paper submissions to electrochemical stability make them interesting solvents for the journal Green Chemistry. Also, the availability of ILs various applications and started a hype in the beginning of increased due to beginning industrial production (Plechkova the twenty-first century. and Seddon 2008). Kunz and Hackl ¨ (2016) as well as Wasserscheid pointed However, soon, first doubts were cast on the environmen- out that the expectations followed a Gartner cycle as visible tal benefit of ILs (Thuy Pham et al. 2010) and problems on the hydrolysis of some ILs were reported (Steudte et al. 2012). Additionally, many ionic liquids did not have a This article is part of a Special Issue on ‘Ionic Liquids and unique selling point or their improved performance did not Biomolecules’ edited by Antonio Benedetto and Hans-Joachim justify their increased cost (Plechkova and Seddon 2008; Galla Kunz and Hackl ¨ 2016). As a result, the hype cooled down Christian Schroder ¨ in the past decade reaching the so-called Trough of Disillu- christian.schroeder@univie.ac.at sionment of a Gartner cycle. Nevertheless, scientific research did not abate, visible by Faculty of Chemistry, Department of Computational the roughly exponential increase in publications indicated Biological Chemistry, University of Vienna, Wahringerstr ¨ . 17, Vienna, Austria by the blue dashed line in Fig. 1. Here, we normalized 826 Biophys Rev (2018) 10:825–840 Hype stage Implementation stage depicted in Fig. 2. Here, coordination numbers of particular solvent molecules and hydrogen bonding are interesting features to be computed from MD trajectories. Moreover, total publications on IL radial distribution functions (solid black line in Fig. 2) simulations on IL simulations on protein/IL detecting the accumulation or depletion of solvent species at the biomolecular surface and their interpretation in terms of Kirkwood-Buff theory (Lesch et al. 2015; Diddens et al. IL are IL are biodegradable green toxic IL 2017; Smiatek 2017) are reported in “The solvation layer”. Due to the surrounding hydrophilic or hydrophobic sol- vent, the structure and stability of the biomolecular solute may adapt to the environment. Depending on the size and shape of a solute, different mechanisms contribute to solvent-induced changes. Hence, we distinguish between large flexible biomolecules and small rigid drugs, employ- ing different analysis routines as sketched in Fig. 2: Technology Trough of Slope of trigger Disillusionment Enlightment 1. Large proteins, polypeptides, or polymeric biomolecules Fig. 1 Gartner hype cycle on ionic liquids adapted from Kunz and possess a secondary structure which can deform during Hackl ¨ (2016), Wasserscheid (2017) the solvation process (indicated by the gray arrows in Fig. 2). In a hydrophilic environment, for example, the publication growth obtained from a search in the “Web hydrophilic subunits are supposed to be at the surface of the of knowledge” to the current number of publications up biomolecule and in contact with the solvent. Hydropho- to 2017. This way, we can compare the development with bic subunits are covered in the core of the biomolecule. publication efforts concerning simulations of ionic liquids In addition to refolding, the size of a biomolecule shown as an orange dashed line in Fig. 1 which agrees may change when moving from a hydrophilic to a almost quantitatively with the overall publication output on hydrophobic environment or vice versa. Computational ILs. size and shape observables of this biomolecular class Quite generally, hype cycles can be decomposed into two are discussed in “Large flexible proteins”. processes (Sasaki 2015): a hype stage and an implementa- 2. In contrast, small rigid drugs cannot increase their size or tion stage. The latter can be fitted by an S-shaped Gompertz change their shape significantly. Here, other mechanisms function like solute aggregation can be monitored by molecular −k(t −t ) −e dynamics. The inclusion of the solute in the solvent net- f(t) = a · e.(1) work can also be computed by free energy calculations. Taking the publication efforts depicted in Fig. 1 as a Furthermore, functional groups and their respective measure of implementation, the “Slope of Enlightenment” influence on the solvation are of significant importance. phase can be extrapolated. In addition to the amplitude a This will be discussed in “Small rigid drugs.” and the stretching factor k, the inflection point t marks the year in which the implementations start to level off. In the case of the ILs, the year 2020 is obtained for both The solvation layer the total publications and the publications concerning the simulation of ionic liquids. The “Plateau of Productivity” The tunable solvation properties (Anderson et al. 2002; will be reached in twenty to thirty years based on the current Gutowski et al. 2003) of ILs gave reasons for the hype on data. ILs at the start of this century. They were considered as In order to get there, many scientific questions have green replacements for organic solvents in general (Cull yet to be answered. In particular, the knowledge gained et al. 2000) and for biological processes in particular (Yang by simulations on biomolecules in ionic liquids and their and Pan 2005; Park and Kazlauskas 2003). For example, mixtures is lagging behind the general ionic liquid trend as Rogers and co-workers reported on the dissolution of visible by the green dotted line in Fig. 1. The current review cellulose in 2002 (Swatloski et al. 2002). In 2005, BASF tries to summarize the current efforts of biomolecular licensed the exclusive use of various intellectual property solvation in ionic liquids and their mixtures from a rights from this group. In 2007, the Rogers group continued molecular dynamics (MD) perspective. their research on dissolution of lignocellulosic materials Solvation starts at the first solvation layer which is the (Fort et al. 2007). Mikkola et al. (2007), Anugwom et al. transition region between the bulk solvent and the solute as (2012), Raut et al. (2015), Kilpelainen ¨ et al. (2007), today inflection point Biophys Rev (2018) 10:825–840 827 Fig. 2 Computational analysis of the solvation interaction between biomolecules, their solvation layers, and the bulk solvent King et al. (2011), and Parviainen et al. (2013)worked on protein properties such as stability or solubility (Zhang on cellulose processing. Excellent reviews summarize the and Cremer 2006;Schroder ¨ 2017). Ions are classified either current development (Pinkert et al. 2009; Brandt et al. 2013; as kosmotropes (structure-makers) or chaotropes (structure- Maki-Arv ¨ ela et al. 2010; Zhang et al. 2017)and show that breakers) by their ability to influence water structure (Tung we are still in the “Slope of Enlightenment” phase during and Pfaendtner 2016; Constantinescu et al. 2007; Zhang and the implementation stage. Cremer 2006). Generally, kosmotropic anions stabilize pro- Due the high cost of ILs compared to the conventional teins, while chaotropic anions lead to destabilization (Tung alternatives, not many large-scale applications were intro- and Pfaendtner 2016). Still, exceptions to the Hofmeister duced. In order to reduce the costs, ILs were mixed with series have been observed (Zhang and Cremer 2009), not other cheap solvents, in particular water. Since only few pro- only due to the fact that specific protein properties such teins are soluble in pure ionic liquids (Shao 2013;Kragl as charge and surface characteristics (Schroder ¨ 2017; Con- et al. 2002), homogeneous and heterogeneous aqueous ionic stantinescu et al. 2007; Tung and Pfaendtner 2016) are not liquid mixtures were used (Kragl et al. 2002). The first ionic considered. Moreover, the ability of ions to influence bulk liquid-based aqueous biphasic system consisted of 1-butyl- water structure in a kosmo- or chaotropic manner has been 3-imidazolium chloride mixture in aqueous K PO and was questioned (Omta et al. 2003). 3 4 reported by Rogers and co-workers in 2003 (Gutowski et al. Less publications exist on the computational research of 2003) and may be considered as a “Technology Trigger” aqueous ionic liquid mixtures (Bhargava and Klein 2009; (see Fig. 1). Recently, biphasic systems regained interest, in Varela et al. 2015; Chang et al. 2010) without additional particular by the group of Coutinho for extraction (Ventura solutes as these studies focus on more fundamental prop- et al. 2009, 2017; Pei et al. 2009; Dreyer and Kragl 2008; erties of IL mixtures rather than a particular application. Pereira et al. 2010; Lee et al. 2017;Claudio ´ et al. 2010)and The effect of water on the viscosity of an IL/water mixture for purification of biomolecules (Pereira et al. 2013). An was studied by Kelkar and Maginn (2007). Furthermore, the excellent review was given by Freire et al. (2012). alkyl chain length of the IL cations changed the collective Although of great importance for the rational choice structure of the aqueous IL mixture (Bhargava and Klein of an ionic liquid for any given application (Weingartner ¨ 2009). We started our computational analysis on the col- et al. 2012), predicting the effect of an ionic liquid on lective network between water and various ionic liquids in the stability of a specific protein still remains difficult 2007 (Schroder ¨ et al. 2007, 2008, 2009, 2014) with a focus (Tung and Pfaendtner 2016) due to the complex and lit- on dielectric properties. Smiatek and co-workers published tle understood molecular interactions between ionic liquid interesting papers on Kirkwood-Buff analysis of protein sol- and protein (Burney et al. 2015) demonstrating that we vation in IL/water mixtures (Diddens et al. 2017; Smiatek are still in the “Slope of Enlightenment” phase and have 2017). Gupta and co-workers reported MD simulations on not reached the “Plateau of Productivity” in Fig. 1.A the interaction of cellulose with ionic liquids (Gupta et al. popular classification of the effect of ionic liquids on pro- 2013). As a complete review of computational analysis of teins already mentioned is the extension of the well-known interactions of biomolecules with IL mixtures is out of the Hofmeister series (Zhang and Cremer 2006) to aqueous scope of this review, we will focus on the basic techniques to ionic liquid solutions (Constantinescu et al. 2007). The encourage other authors to contribute to the understanding Hofmeister series ranks ions by their specific ion effect of biomolecular solvation in ionic liquid mixtures by MD 828 Biophys Rev (2018) 10:825–840 simulations. Hereafter, we will give a compact overview of of xylanase II from trichoderma longibrachiatum in solvation layer properties obtainable by MD simulations as aqueous [C mim][EtSO ]and [C mim][OAc] (Jaeger and 2 4 2 well as their interpretation. Pfaendtner 2013). Steinhauser and co-workers noted that Eq. 3 not only Spatial distribution of the solvent requires a meaningful definition of a solvation shell radius R but also assumes sphericity of the solutes (Haberler The spatial distribution of molecular species j around and Steinhauser 2011; Haberler et al. 2011; Zeindlhofer et al. 2018). Although this approach is feasible for many a reference site i is often characterized by the radial distribution function g (r), small solutes, it may give highly inaccurate results in case ij of anisotropic solutes (Zeindlhofer et al. 2017, 2018) like δ(r − r ) ij large proteins (Neumayr et al. 2010) due to non-symmetric j 1 g (r) = · (2) ij excluded volume effects. Furthermore, for heterogeneous 4πr dr· ρ solvents with different molecular sizes, more than one shell which is the ratio between the local density of j in a radius would be necessary to describe a solvation shell spherical shell 4πr dr and the global density ρ of species appropriately (Haberler et al. 2011). j . The distance r is usually defined between the center of ij masses of the molecules i and j or between the coordinates Voronoi-based coordination numbers of particular atoms belonging to these molecules. An alternative method of spatial decomposition is Voronoi Coordination numbers decomposition (Schroder ¨ et al. 2009). It is parameter- free and also allows for straightforward decomposition of To calculate the coordination number n of species around ij multiple solvation shells. In Voronoi tessellation, each point the biomolecule i, spherical integration of the radial in space representing an atomic coordinate is assigned an distribution function is commonly used: irregular polyhedron that contains all space closer to its associated reference point than to any other point in the system. If two points share a polyhedron face, they are n = 4πr g (r)dr (3) ij ij defined as neighbors. Solvent molecules that share a face with the biomolecule are considered first-shell members, Micaelo ˆ and Soares found that the coordination num- molecules that share a face with first-shell members are ber of water around cutinase in aqueous 1-butyl-3- considered second shell, and so on. This analysis can methylimidazolium nitrate [C mim][NO ] and 1-butyl- be performed for each time step during a trajectory to 4 3 3-methylimidazolium hexafluorophosphate [C mim][PF ] yield mean residence times of particular species at the 4 6 mixtures is reduced at the protein surface (Micaelo ˆ and surface of the biomolecule (Haberler and Steinhauser 2011; Soares 2008). Similar removal of water was observed for Haberler et al. 2011). This shell assignment can also be hydrated Candida rugosa lipase 1 in [C mim][PF ]and applied to the radial distribution function (Neumayr et al. 4 6 [C mim][NO ] (Burney and Pfaendtner 2013). 2010; Zeindlhofer et al. 2018) as depicted in Fig. 3. 4 3 A problem that may arise when comparing coordination The decomposition of the displayed g (r) into spherical ij numbers of different IL ions is the fact that the coordination solvation shells discussed in the last section is not possible number depends on ion density. For the analysis of cation and anion coordination numbers around CAL-B, Klahn ¨ et al. (2011a) normalized the coordination number by the average number of the respective species within 10 Aof the protein to facilitate comparison. They observed the highest ion densities around charged residues. Interestingly, although density of cations and anions was highest around oppositely charged residues, densities around residues with the same charge was also higher than for polar and non- charged residues (Schroder ¨ 2017;Klahn ¨ et al. 2011a), which is an effect of the strong interionic interactions in ionic liquids (Klahn ¨ et al. 2011b). Furthermore, a diffusion Fig. 3 Radial distribution function g (r) of water oxygens (j ) around ij of cations into the active site of the protein occurred, the center of mass of calbindin i and the corresponding decomposition possibly hindering substrate access. A similar occupation into Voronoi shells. Reproduced from Neumayr et al. (2010) with of the active site by cations was observed in simulations permission of AIP Publishing Biophys Rev (2018) 10:825–840 829 −19 as no clear first minimum of the radial distribution can be e = 1.6022 × 10 Asand n± equals the number detected. Furthermore, the different shells overlap to a large of indistinguishable ion species. The binding coefficient extent. Consequently, the spherical integration performed in determines the transfer free energy (Smiatek 2017). Eq. 3 will not yield a correct coordination number. However, F (R) =−k T · ν (R) (7) 23 B 23 integrating the shell contributions in Fig. 3 results in the exact which estimates the excess energy at a given temperature coordination number obtained from counting neighbors by shared faces of the solute and solvent polyhedra. T that is needed to transfer the ionic liquid from infinite Employing Voronoi decomposition, Haberler and Stein- distance to close proximity around the biomolecule. hauser confirmed a surplus of C mim in the first solvation In a MD study of a small β -hairpin peptide in shell of various proteins despite the positive net charge of aqueous [C mim][OAc], where both native and denaturated the protein (Haberler and Steinhauser 2011; Haberler et al. conformations of the peptide were simulated, Kirkwood- 2011). This is very likely due to expulsion of the cations Buff preferential binding coefficients revealed that the unfolded conformation attracts more anions while no from the strong anion-water network, as also observed in other studies (Schroder ¨ et al. 2007, 2009; Haberler and difference in cation affinity is observed (Lesch et al. 2015). This agrees with the fact that the anionic influence on the Steinhauser 2011). Interestingly, the mean residence time of the anions at the protein surface is longer compared denaturation process is stronger (Klahn ¨ et al. 2011a). The unfolding effect of [C mim][OAc] on the peptide was also to the cations (Haberler and Steinhauser 2011) which can be explained by stronger interactions with the oppositely confirmed by experiment (Patel et al. 2014). In a follow- charged amino acid (Schroder ¨ 2017; Jaeger and Pfaendtner up publication by the same group, a similar approach was 2013;Klahn ¨ et al. 2011a;Micaelo ˆ and Soares 2008). used to study the behavior of the same β -hairpin peptide in aqueous [C mim][BF ]and[C mim]Cl. The binding 2 4 2 Kirkwood-Buff theory strength order of the anions was identified as OAc − − − BF > Cl , and again it was found that OAc strongly binds to the denatured state, which is not observable for Another approach utilizing radial distribution functions is − − the so-called Kirkwood-Buff theory, which relates structural BF and Cl . The authors suggested that the larger acetate anion exhibits a conformation-dependent binding behavior properties of the solutions due to specific interactions with the protein (Diddens et al. R 2017). Furthermore, they found only negligible dependence of the binding behavior of C mim on the anion. G (R) = g (r) − 1 4πr dr, (4) ij ij The interactions of caffeine, gallic acid, protocatechuic acid, and quercetin with aqueous mixtures of the ionic liquid [C mim][OAc] were studied by a shell-resolved to thermodynamic quantities (Kirkwood and Buff 1951). approach using Voronoi tessellation. As shown in the The Kirkwood-Buff integral is often interpreted as excess respective publications (Zeindlhofer et al. 2017, 2018), the volume around a solute species when compared to an investigated biomolecules are flat, aromatic systems of high ideal solution. It is also possible to compute Kirkwood- anisotropy, which makes Voronoi tessellation necessary Buff integrals of distinct solvation shells based on Voronoi to distinguish between several solvation shells. Since tessellation (Zeindlhofer et al. 2018). Voronoi tessellation allows straightforward calculation of As a standard notation, i, j = 1 ... 3 denote water, solvent shell volumes, a shell-wise calculation of Kirkwood- the biomolecular solute, and the co-solvent (in our case Buff interaction parameters analogous to Kirkwood-Buff the ionic liquid), respectively. Distance-dependent binding integrals (Zeindlhofer et al. 2018) can be computed. coefficients ν (R) between the solute i = 2 and the ionic However, the comparison of Kirkwood-Buff interaction liquid j = 3 can be obtained as follows: parameters between solutes of different size is complicated, since they depend on the solute size. In contrast, the shell- solute 1 q wise potential of mean force (Zeindlhofer et al. 2017) ν (R) = ρ (G (R) − G (R)) − (5) 23 3 23 21 n± e measures solvent affinities of solutes differing in size via solute 1 ρ q 3 the ratio of shell to bulk concentration of a species. = n (R) − n (R) − (6) 23 21 n± ρ e Solvation dynamics using coordination numbers defined in Eq.3 as well as The role of ionic liquids in solvation dynamics (Mandal the Donnan equilibrium condition (Pierce et al. 2008; solute and Samanta 2005; Samanta 2010; Roy and Maroncelli Smith 2004; Smiatek 2017). Here q corresponds to the 2012; Terranova and Corcelli 2013; Liang et al. 2014;Heid charge of the biomolecule in units of the elementary charge 830 Biophys Rev (2018) 10:825–840 and Schroder ¨ 2018) as well as the solvation dynamics of In contrast to neutral liquids like water, ionic liquids can biomolecules (Sajadi et al. 2014; Heid and Schroder ¨ 2016; also respond to the change in the local electric field via 2017) has attracted significant interest over the past decade. translation (Terranova and Corcelli 2013; Schmollngruber The time-dependent Stokes shift et al. 2013), in particular the anions. The response of induced solvent dipoles is restricted to the first solvation ν(t) − ν(∞) S(t) = (8) shell (Schmollngruber et al. 2013). Using polarizable non- ν(0) − ν(∞) equilibrium molecular dynamics simulations, not only the monitors the transient behavior of the solvent relaxation normalized Stokes shift relaxation function but also the pure after an electronic excitation of a dissolved chromophore. elec elec shift U (t ) − U (∞) can be reproduced (Heid and THz absorption bands of biomolecular hydration layers Schroder ¨ 2018). are generally swamped by absorption from bulk solvent. However, linking the chromophore to the biomolecule, the solvation properties of the immediate solvent molecules can Large flexible proteins be experimentally studied (Sajadi et al. 2014)asshown for the chromophore oxyquinolinium betaine attached to One of the “Technology Triggers” in Fig. 1 for protein sta- trehalose in Fig. 4. bility was published in 2000 by Summers et al. who showed In MD simulations, this experiment can be modeled by that ethylammonium nitrate can successfully refold dena- equilibrium (Roy and Maroncelli 2012; Schmollngruber tured egg-white lysozyme (Summers and Flowers 2000). et al. 2013; Terranova and Corcelli 2013) and non- Several groups observed that ionic liquids can significantly equilibrium simulations (Heid and Schroder ¨ 2016, 2017, stabilize certain proteins and enzymes (Lozano et al. 2001; 2018). However, the latter approach seems to be more Fujita et al. 2007;Weingartner ¨ et al. 2012), especially lipases appropriate (Heid and Schroder ¨ 2018). The change of the (Kaar et al. 2003; Ulbert et al. 2005; Lai et al. 2011), even at elec Coulomb interaction energy U (t ) of the chromophore high temperatures (Ulbert et al. 2005). In some cases, enzy- atoms j with the solvent atoms i me activity, enantioselectivity, and regioselectivity in ionic 1 q · q liquids are increased compared to conventional organic sol- j i elec U (t ) = (9) vents (Schofer ¨ et al. 2001) as discussed in more detail in 4π r (t ) 0 ij j i several reviews (Kragl et al. 2002; Park and Kazlauskas relaxes during the non-equilibrium simulation after switch- 2003;Yangand Pan 2005; van Rantwijk and Sheldon 2007; ing the partial charges q of the solute from ground to j Moniruzzaman et al. 2010; Lai et al. 2011; Naushad et al. excited state values: 2012; Gao et al. 2015;Chenetal. 2010;Zhao 2010; elec elec Weingartner ¨ et al. 2012; Patel et al. 2014; Benedetto and U (t ) − U (∞) S(t) = (10) Ballone 2016; Smiatek 2017;Schroder ¨ 2017;Zhao 2016). elec elec U (0) − U (∞) An observation that many groups agree with is that the presence of water in the ionic liquid is crucial for protein stability (Klahn ¨ et al. 2011a; Lozano et al. 2001). It was also observed that the effect of the anion on protein structure is more pronounced than the effect of the cation in most cases (Weingartner ¨ et al. 2012; Constantinescu et al. 2007; Kaar et al. 2003). Furthermore, the destabilizing effect of ionic liquids in the aqueous mixture is a function of the ion concentration (Senske et al. 2016) as shown in Fig. 5. Stabilizing effects (T > 0) were only observed for dihydrogenphosphate salts as well as sodium and potassium chloride above 2 M. Here, “enlightenment” from MD simulations on the mechanism is highly welcome. In the quest of rationalizing ionic liquid protein interac- tions, MD simulations provide a useful tool for investiga- tions at a molecular level and can complement experimental findings (Haberler and Steinhauser 2011). Although the first Fig. 4 The chromophore oxyquinolinium betaine can be directly MD simulations of proteins were reported in the late 1970s linked to the biomolecule of interest. Upon laser excitation, the local (McCammon et al. 1977), the first MD simulation of a electric field (red lines) is changed and the response of the solvent protein in an aqueous ionic liquid was published only in molecules in the direct vicinity of the trehalose can be studied (Sajadi et al. 2014). Copyright 2014, American Chemical Society 2008 (Micaelo ˆ and Soares 2008) revealing the time delay of Biophys Rev (2018) 10:825–840 831 computational analysis in Fig. 1 concerning interaction of of large timescales by introducing an enhanced sampling proteins with ionic liquids. In Micaelo ˆ and Soares (2008), method called “infrequent metadynamics,” an algorithm investigated the behavior of the serine protease cutinase in that allows estimation of unfolding times on a greatly accel- [C mim][PF ]and [C mim][NO ] at various water con- erated timescale (Tung and Pfaendtner 2016). For assessing 4 6 4 3 tents. For [C mim][PF ], 5–10% of water was necessary protein stability, several structural quantities are accessible 4 6 for optimal enzyme stabilization (Micaelo ˆ and Soares 2008) via MD simulations: indicating the importance of even small amounts of water. A subsequent publication by the same group further Radius of gyration investigated the influence of the protein surface characteris- tics on stability in ILs by conducting simulations of Candida As unfolding increases protein volume, the loss of secondary rugosa lipase and Bos taurus α-chrymotopsine with mod- structure leads to an increase of the radius of gyration r (t ) gyr ified surface charges (Burney et al. 2015). By replacing surface lysines with glutamate, artificial mutations with m (r (t ) − r (t )) i i CM positive-to-negative surface charge ratios were examined in r (t ) = (11) gyr simulation. On a 50-ns timescale, no stability difference for the different enzymes was found, but the authors explic- itly state that such a timescale is likely too short (Burney calculated with respect to the current center of mass r (t ) CM et al. 2015). Due to the slow dynamics brought about by the of the protein, where m is the mass of the ith atom and M high viscosity of ionic liquids, the need for large timescales the total mass of the protein. Klahn ¨ et al. (2011a) reported when conducting simulations in ionic liquids is one of the that r (t ) of Candida antarctica lipase B (CAL-B) in aque- gyr great obstacles, and unfolding times are often too large to ous IL mixtures (hydration level 20% w/w) increased by be accessible by simulation (Tung and Pfaendtner 2016). 0.8, 1.1, and 2.8 Afor [C mim][PF ], [C mim][BF ], and 4 6 4 4 The problem of insufficient sampling in pure ionic liquids [C mim][NO ], respectively. They suggested that the strength 4 3 is also mentioned by Burney and Pfaendtner in their study of interactions between anions and enzyme is mainly gov- on the behavior of Candida rugosa lipase in [C mim][NO ] erned by ion density and ion surface charge, favoring inter- 4 3 and [C mim][PF ]. Despite sampling times of more than actions with smaller ions which can form stronger hydrogen 4 6 100 ns, no visible effects of the ionic liquid on the pro- bonds with the protein. This finding goes along with the tein could be observed, despite experimental evidence that denaturation capabilities of classical ILs in Fig. 5e. the enzyme is heavily denatured in [C mim][NO ](Bur- However, Chaban and co-workers (Chevrot et al. 2015) 4 3 ney et al. 2015). Tung and Pfaendtner address this problem showed that the radius of gyration does not increase in ionic Fig. 5 Salt-induced shifts of the melting temperature T of RNase in aqueous ionic liquid mixtures with various ion concentrations (Senske et al. 2016)—Published by the PCCP Owner Societies. (a,b) chlorides (c)bromides(d) sodium salts (e)C mim based salts (f) stabilizing oxo-anions 832 Biophys Rev (2018) 10:825–840 liquids with amino acid-based anions during the simulation The stronger destabilizing effect of nitrate compared period of 10 ns which may not be long enough to monitor to hexafluorophosphate was confirmed by Klahn ¨ et al. the unfolding process due to the high viscosity of these (2011a) for CAL-B in various aqueous ionic liquid mixtures ionic liquids. Interestingly, Shao (2013) reported that r by means of RMSD-values and agreed well with the gyr of the B domain of protein A from Staphylococcus aureus concurrent observation of an increased radius of gyration actually decreased with increasing [C mim]Cl content of r (t ) mentioned in the last section. Although to less 4 gyr the aqueous solution. Shao performed three independent extent than the anions, longer alkyl chains lead to higher simulations of 200 ns length to get meaningful values. values of these quantities, indicating destabilization (Klahn ¨ A decrease of the protein radius was also observed by et al. 2011b). Replacing butyl with a methoxyethyl group in Bhattacharyya and co-workers (Ghosh et al. 2015)for C mim also resulted in substantial protein destabilization, lysozyme in [C mim]Br. However, they computed the possibly by the increased hydrogen-bonding ability of the radius of hydration from the Stokes-Einstein relation: methoxyethyl group. Diddens et al. (2017) analyzed the RMSD of a β -hairpin k T D = (12) peptide by metadynamics simulations in aqueous mix- 6πηr tures of [C mim][OAc], [C mim][BF ], and [C mim]Cl. 2 2 4 2 using experimental diffusion coefficients from fluorescence The β -hairpin peptide unfolded in aqueous mixture of correlation spectroscopy and the experimental viscosity η. [C mim][OAc] and [C mim][BF ] as visible in the insets 2 2 4 The radius of gyration and the radius of hydration are linked of Fig. 6. However, the global energetic minima of the two by a factor of 5/3(Schroder ¨ et al. 2015). unfolded configurations in Fig. 6b and c are completely different. The authors argued that the strong denaturating Root-mean-squared displacements acetate attacks the intramolecular hydrogen bonds of the peptide. As a result, the peptide structure unfolds com- The time-dependent root-mean-square deviation RMSD pletely. In aqueous [C mim][BF ], the global minimum is 2 4 (Kufareva and Abagyan 2012) still located near that of the pure water solution; however, the RMSD is quite large. Since the anions are excluded 2 from the surface of the peptide, this effect was attributed to ref RMSD = r (t ) − r (13) the impact of the cations. In contrast to the aqueous mixture of [C mim][OAc] is defined as a function of the deviation of the current atomic and [C mim][BF ], the β -hairpin peptide did not unfold in 2 4 ref coordinates r (t ) from the coordinates r of a reference aqueous [C mim]Cl. The minimum structure of the peptide i 2 structure with N atoms, usually at the beginning of the resembles the structure found in pure water (see Fig. 6a simulation or the native configuration. and d). This finding is in agreement with the experiment in In the first MD simulation of a protein in an ionic Fig. 5b, where [C mim]Cl was characterized to be only a liquid by Micaelo ˆ and Soares mentioned above, the weak denaturant. authors analyzed the RMSD of cutinase in [C mim][PF ] In MD simulations of Shao (2013), the RMSD values 4 6 and [C mim][NO ] at different hydration levels and of the B domain of protein A from Staphylococcus aureus 4 3 temperatures (Micaelo ˆ and Soares 2008). In accordance decreased with increasing content of [C mim]Cl which was with experimental findings reporting that many enzymes interpreted as a stabilizing effect that may also be an effect retain activity in ILs with PF anions and lose activity of the increased viscosity. in ILs with NO anions (Lozano et al. 2001; Kaar et al. 2003; Persson and Bornscheuer 2003; Lou et al. 2006), they Root-mean-squared fluctuations observed generally larger RMSD values in [C mim][NO ]. 4 3 In contrast, Burney et al. did not find a significant The root-mean-squared fluctuation change in the RMSD of Candida rugosa lipase 1 in [C mim][PF ]and[C mim][NO ] (Burney and Pfaendtner 1 4 6 4 3 RMSF (i) = (r (t ) −r  ) (14) i i t 2013). Pfaendtner and co-workers (Burney et al. 2015) demonstrated that the RMSD is not only a function of the ionic liquid but also depends on the nature of the protein. For measures the fluctuation of atom i around its mean position the enzyme Candida rugosa lipase 1, the stability decreases r  (Pitera 2014). Experimentally, atomic flexibility is i t in the following order of the solvent: water > [C mim]Cl described with the crystallographic B-factor (Bornot et al. > [C mim][EtSO ]. However, no RMSD trend is visible 2011). Amino acids with high RMSF values have an 2 4 for Bos taurus α-chymotrypsin in the very same solvents increased mobility which may also be an indicator for (Burney et al. 2015). interactions with the ionic liquid (Attri et al. 2015; Burney Biophys Rev (2018) 10:825–840 833 Fig. 6 Free energy landscape with global energetic minimum conformation of a β -hairpin. a Native state of the peptide in water. b Denaturated state in aqueous [C mim]. c Denaturated state in [C mim]. d Denaturated state in [C mim]Cl. R is the end-to-end distance of the peptide. (Diddens et al. 2017) - Published by the PCCP Owner Societies and Pfaendtner 2013; Jaeger and Pfaendtner 2013)which index j may also comprise all atoms (separated by at least breaks or at least disturbs intramolecular hydrogen bonding three bonds) in the very same molecule to obtain intramo- elec vdW of the biomolecule. lecular energies U (t ) and U (t ). Steinhauser and co- PP PP elec For the analysis of simulations of three different cellu- workers (Haberler and Steinhauser 2011) computed U PP vdW lases in aqueous mixtures of the ionic liquid [C mim][OAc], and U  for a zinc finger protein in aqueous solutions PP elec Jaeger et al. (2015)usedthe RMSD and an RMSF (aver- of [C mim][CF SO ], and found a rise in U (x ) with 2 3 3 IL PP vdW aged by residue) of the C atoms of the protein. They increasing ionic mole fraction x , whereas U (x ) α IL IL PP found that the three investigated cellulases (a cellulase from decreased with increasing mole fraction. Moreover, the Thrichderma viride, a cellulase 5A from Thermotoga mar- sum of electrostatic and van-der-Waals energies exhibited itima, and an endoglucanase from Pyrococcus horikoshii) a maximum at an IL mole fraction of x = 0.073. The IL exhibited remarkably different stabilities in the aqueous authors termed this mole fraction a “magic point” since ionic liquid [C mim][OAc], highlighting that specific fea- other secondary structure descriptors also showed extrema tures of the protein, such as surface characteristics, and not at this point. This behavior can be attributed to a change only the ionic liquid itself determine stability. from dipolar screening by water to charge screening by the ionic liquid ions when the IL mole fraction increases, which elec Electrostatic interactions is consistent with a rise of U (x ). IL PP Klahn ¨ et al. analyzed the electrostatic and van-der- Non-bonded energies in MD simulation are most generally Waals-interaction energy of hydrated Candida antarctica elec calculated by the electrostatic energy U (t ) lipase B (CAL-B) with a set of ionic liquids composed of imidazolium- and guadinium-based cations and nitrate, 1 q · q i j tetrafluoroborate or hexfluorophosphate anions (Klahn ¨ et al. elec U (t ) = (15) 4π r (t ) 2011a, b). Here, the second atomic index j of Eq. 15 0 ij i j contains all solvent atoms of the particular solvent species. They found that interaction energies are dominated by between the partial charges q and q of the atoms i and i j elec Coulombic interactions U  between the anions (A) PA j at a distance of r (t ) as well as van-der-Waals energies ij and the protein, while smaller cation (C) contributions vdW U (t ) vdW stem from van-der-Waals-interactions U . The charge PC on cations is more delocalized while the charge density 12 6 σ σ ij ij vdW on the smaller anions is higher, allowing them to form U (t ) = 4 − (16) ij r (t ) r (t ) ij ij stronger hydrogen bonds with the protein. In their study i j on xylanase II in [C mim][EtSO ]and [C mim][OAc], 2 4 2 based on the interatomic Lennard-Jones parameter σ and Jaeger and Pfaendtner also attributed the low affinity of ij . cations for negatively charged residues to the high charge ij Usually, the first atomic index i concerns all atoms in delocalization (Jaeger and Pfaendtner 2013). Klahn ¨ et al. vdW elec the biomolecule P under investigation. The second atomic reported the highest values of U +U  (strongest 834 Biophys Rev (2018) 10:825–840 hydrogen-bond interactions) for the NO anion, which is that these high solvation enthalpies were mainly due to consistent with experimental and simulation results that large cage formation energies. These large cage formation NO destabilizes CAL-B (Klahn ¨ et al. 2011a;Micaelo ˆ and energies are due to the strong interactions between Soares 2008). Weakest interactions were observed for the ionic liquid ions compared to the weaker interactions in PF anion, which stabilizes CAL-B. From experiment, it is water. An exception was the ionic liquid [C mim][PF ], 4 6 also known that [C mim][NO ] can dissolve solid CAL-B, in which the cage formation energy was even lower 4 3 while [C mim][PF ] cannot (Klahn ¨ et al. 2011a; Lau et al. than in water. Furthermore, upon transfer from water to 4 6 2004). IL, the internal energy of the protein increased, except in [C mim][PF ], corresponding to the experimentally 4 6 Solvation free energy observed stability increase in [C mim][PF ]. Another 4 6 approach to estimate the free energy landscape of a protein The general procedure of solvation free energy calculation is the use of metadynamics (Barducci et al. 2008), which is briefly described in the next section on small molecules. was already mentioned in the context of a small β - As already mentioned before, the simulations of proteins hairpin protein whose unfolding free energy landscape in in ionic liquids are hampered by the increased viscosity aqueous [C mim][OAc], [C mim][BF ], and [C mim][Cl] 2 2 4 2 in ionic liquids, requiring long simulation times for was calculated via metadynamics (Lesch et al. 2015; sufficient sampling (Tung and Pfaendtner 2016). Under Diddens et al. 2017). As mentioned earlier, the strong these circumstances, it is impossible to acquire sufficient denaturation of the peptide by [C mim][OAc] apparent from preferential binding parameters was also visible in the sampling for the calculation of solvation free energies (Klahn ¨ et al. 2011a). However, some other methods for free energy landscapes. obtaining thermodynamic information about proteins in ionic liquids have been employed. Klahn ¨ et al. (2011a) calculated the solvation free enthalpy H of CAL-B based Small rigid drugs solv on a linear response formalism (Klahn ¨ et al. 2011a), Ionic liquids are of special interest for the extraction of solute solvent compounds from biomass. They can dissolve a wide range H = U − U + U + PV , solv solv gas solute of biomatrices, even materials such as cellulose (Pinkert (17) et al. 2009;Wangetal. 2012; Zhu et al. 2006) or chitin (Qin et al. 2010), and can enhance solubility of hydrophobic solute where U is the potential of the system containing N solv compounds in aqueous solution by acting as hydrotropes solvent solute solvated solutes in N solvent molecules, and U gas (Claudio ´ et al. 2015) or surfactants. and U are the potential energies of the N solutes in solvent From the computational point of view, these molecules gas phase and the pure solvent, respectively. The pressure- are much more rigid compared to a large flexible protein. volume term is small and may be neglected, and when Consequently, observables concerning size and shape, solute N = 1 is assumed, Eq. 17 may be written as e.g., the radius of gyration as well as root-mean-squared solute solvent solute−solvent displacements and fluctuations, are of less use for the H = U + U + U solv solv solv small biomolecular drugs dissolved in the aqueous mixtures solute solvent − U + U (18) gas of ionic liquids. Hence, particular interactions like π -π - solute solute−solvent stacking or hydrogen bonding attract more attention. Also, = U + U + U . (19) cage free energy calculations, which are not feasible for complete solute Here, U is the potential energy of only the solute in proteins, may help to understand the biomolecular solvation solv solvent solvent, U is the potential energy of only the solvent of smaller solutes. However, the current knowledge on the solv solute−solvent in the system, and U is the solute-solvent interaction of small biomolecules with ILs is not as far solute solvent interaction energy. U = U −U is termed the progressed compared to protein solvation. cage solv cage formation energy and measures the energy required to solute solvent form a solute cavity in the solvent. U = U − Hydrotropic theory solv solute U is the change in internal energy of the solute upon gas insertion into solvent (Klahn ¨ et al. 2011a). It was found The hydrotropic effect describes the increased water that solvation enthalpies in ILs increased noticeably in solubility of hydrophobic organic compounds by the ionic liquids compared to water, which is in accordance addition of a co-solvent (Shimizu and Matubayasi 2014; with experimental observations that CAL-B is less soluble Eastoe et al. 2011;Claudio ´ et al. 2015; Zeindlhofer et al. in ionic liquids compared to water. It was also observed 2017, 2018) which usually consists of small, amphiphilic Biophys Rev (2018) 10:825–840 835 molecules (Booth et al. 2012). In principle, the following (Zeindlhofer et al. 2018) on the solvation of coffee ingredi- mechanisms are discussed: ents in [C mim][OAc] mixtures. The cations seem to repel water in general and also anions from the surface of the 1. The solute forms a stable complex with few hydrotropic solute at higher concentration. Even in the second solvation molecules via hydrogen bonding or π -π -stacking. The shell determined by Voronoi tessellation, the concentration complex has an increased polarity, hence increased of water is below bulk density which argues against mech- solubility in water. anism 1. Nevertheless, water still has access to the solute 2. The hydrotrope assembles in micelle-like structures in surface, even at higher concentrations, and micellar cationic which the solute can be accommodated. structures are not observed at the surface. Also, the last 3. The water-immiscible solute induces a particular struc- hydrotropic hypothesis 3 could not be proven. The inclusion ture of the hydrotrope in the aqueous solution. This new of acetate in the water network is usually not a problem and structure promotes the solubility of the solute. does not change the water structure in the proximity of the Imidazolium-based ILs seem to be promising candidates solute very much. However, a Voronoi shell-wise computa- for hydrotropy. Their charged, hydrophilic ring is accom- tion of the dielectric permittivities reveals a subtle transition panied by more or less hydrophobic alkyl side chains (Tan from the solute permittivity to the permittivity of the bulk et al. 2009;Claudio ´ et al. 2015). We recently reported solution (Zeindlhofer et al. 2018). Fig. 7 Distribution functions P(n ) of aggregate sizes in pure aqueous ionic liquid mixtures (a, b) and with vanillin (d, e). c depicts a vanillin-water mixture. Reproduced from Claudio ´ et al. (2015) with permission of The Royal Society of Chemistry 836 Biophys Rev (2018) 10:825–840 Furthermore, hydrogen bonding between the water and visible in Fig. 7c,whichisbrokenupintosmalleraggregates the solute is not changed by the cations (as suggested by in the aqueous IL solution by formation of larger vanillin- hypothesis 3) but in competition with the anions. Interest- cation clusters and smaller vanillin-anion clusters in Fig. 7d ingly, the solvation efficiency drops from tosylate to chlo- and e. − − ride for gallic acid (Claudio ´ et al. 2015): Cl <[MeSO ] − − < [CF SO ] < [N(CN) ] < [Tos ]. A similar ranking Solvation free energy 3 2 was found in vanillin. Although the cations are the most prominent species at the solute surface, particular anion Enthalpic and entropic effects of the dissolution of a solute interactions are very important for the solvation behav- in a solvent can be best characterized by the computation ior of the ionic liquid, i.e., hydrogen bonding. Cationic of solvation free energies. In contrast to large proteins, π -π -stacking is of minor importance. the amount of sampling needed for these calculations is feasible for small compounds. A common approach is Solute-solvent interactions and solute aggregation thermodynamic integration. During several simulations, a coupling parameter λ is stepwise increased from zero to The extraction of shikimic acid from star anise pods one corresponding to a switching off of solute-solvent solute−solvent using [C mim]-based ionic liquids with a set of different interactions U − − − − − − anions (PF ,NTf ,BF ,Cl ,OTf ,OAc )was 6 2 4 solute−solvent solvent U(λ) = (1 − λ) U + λ · U (20) reported in an experimental study (Zirbs et al. 2013). Accompanying MD simulations were performed to identify solvent U corresponds to the total energy of the solvent only. the relevant molecular interactions in the extraction process: The free energy difference between the initial and final No correlation of the experimental extraction yield with λ-state is given by the coordination number or the contact surface obtained from simulations could be established (Zirbs et al. 2013). ∂U (λ) G = dλ (21) solv Coordination numbers around shikimic acid are higher for ∂λ cations, but anions seem to interact stronger since they are To study molecular details of extraction of amino acids able to form hydrogen bonds, and their residence time in the first solvation shell around shikimic acid is longer than from aqueous solution using ILs (Absalan et al. 2010;Tome ´ et al. 2012), Seduraman et al. (2012) calculated the sol- for cations. The number of anion hydrogen bonds correlates with the extraction yield, highlighting the importance of the vation free energy of L-tryptophan in pure and aqueous [C mim][PF ], [C mim][PF ], and [C mim][BF ]. In con- anions in the extraction process. 4 6 8 6 8 4 In agreement with electrostatic analysis of the IL ion trast to experimental studies, they found only negligible with proteins (Klahn ¨ et al. 2011a;Klahn ¨ et al. 2011b), differences in the free energies in the pure ILs. However, elec U  of the anions (A) with the solute (S) is much including residual water in the IL simulations, which is SA elec larger compared to U  of the cations (C) (Zeindlhofer higher in [C mim][PF ], the experimental trend is repro- 4 6 SC duced illustrating the importance of water for the extraction et al. 2017;Tomee ´ tal. 2012). Furthermore, the van-der- vdW − Waals interaction U  between cations and the solute process. The solvation in BF -based ILs is more favorable, SC elec probably due to the fact that the water density is 7.5 times exceeds the corresponding Coulomb interaction U SC for C mim at the surface of various coffee ingredients higher compared to the hexafluorophosphate systems. Fur- thermore, electrostatic interactions of tryptophan with BF (Zeindlhofer et al. 2017). Cations are mainly found parallel to the aromatic solute rings in a stacked orientation, while are stronger than with PF . This anion dependency was anions preferred positions around the ring that allowed observed for other amino acids as well (Absalan et al. 2010). hydrogen bonding to ring substituents as visible from elec U . The same finding was observed for five different SA amino acids (glycine, alanine, valine, isoleucine, glutamic Conclusion acid) in aqueous solutions of [C mim][NTf ](Tomee ´ tal. 4 2 After the initial hype in the beginning of the twenty-first 2012). Another experimental study complemented by MD simu- century, expectations on ionic liquids decreased because of issues with the price, viscosity, the stability in water, lations concerned the extraction of vanillin in aqueous solu- tions of [C mim][N(CN) ]and[C mim][SCN] (Claudio ´ the toxicity, and biodegradability. Often, the unique selling 4 2 4 point of ionic liquids was missing in applications to et al. 2015). In MD simulations, the ionic liquid ions form several filamentous ionic strands and interact with water justify the higher cost compared to conventional solvents. in an anion-water hydrogen bond network. The hydropho- Moreover, exorbitant promises during the hype prompted bic vanillin molecules form a large droplet in pure water as several authors to question the current status of knowledge Biophys Rev (2018) 10:825–840 837 and the usefulness of IL in general (Kunz and Hackl ¨ Applying these suggestions to future simulations on 2016). Consequently, to prevent making mistakes again, biomolecular solvation in ionic liquids should help to reach which were made during the hype, good practice routines the “Plateau of Productivity” within reasonable time. should go without saying. In the current review, we briefly summarized meaningful computational methods to Acknowledgments Open access funding provided by Austrian Sci- investigate the solvation of large and small biomolecules in ence Fund (FWF). aqueous ionic liquid mixtures. The increasing number of publications concerning ionic Compliance with Ethical Standards liquids demonstrates their unbowed interest. However, in This article does not contain any studies with human participants or order to make real progress during the implementation of animals performed by any of the authors. ionic liquids and their computational analysis, several ideas should be taken into account: Conflict of interests Veronika Zeindlhofer declares that she has no conflict of interest. Christian Schroder ¨ declares that he has no conflict of interest. 1. The vast majority of computational works deal with aprotic imidazolium-based ionic liquids. Maybe, protic Open Access This article is distributed under the terms of the imidazoliums as well as other cations like choline Creative Commons Attribution 4.0 International License (http:// may be more biodegradable and cheaper. Therefore, creativecommons.org/licenses/by/4.0/), which permits unrestricted juxtaposition of computational results of these “new” use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a ionic liquids to the aprotic imidazolium-based ionic link to the Creative Commons license, and indicate if changes were liquid are interesting. made. 2. The same is true for the anions. Tetrafluorobo- rate and hexafluorophosphate can easily be coarse- grained and much experimental data exists, but these References anions are of less interest for the experimenter. Therefore, other anions such as dihydrogenphosphate Absalan G, Akhond M, Sheikhian L (2010) Partitioning of acidic, (dhp), methyl- and ethylsulfate, hydrophobic bis- basic and neutral amino acids into imidazolium-based ionic (trifluoromethylsulfonyl)-imide NTf , or hydrophilic liquids. Amino Acids 39(1):167–174 Anderson JL, Ding J, Welton T, Armstrong DW (2002) Characterizing dicyanamide N(CN) should also be included in the ionic liquids on the basis of multiple solvation interactions. J Am portfolio of the theoreticians. Even deep eutectic sol- Chem Soc 124(47):14247–14254 vents are interesting but currently still mostly neglected Anugwom I, Maki-Arv ¨ ela P., Virtanen P, Willfor ¨ S., Sjoholm ¨ in MD simulations on solvation of biomolecules. R., Mikkola JP (2012) Selective extraction of hemicelluloses 3. Many solvation effects depend on the concentration of from spruce using switchable ionic liquids. Carbohydr Polym 87(3):2005–2011 the co-solvent and should be given much more room Armand M, Endres F, MacFarlane DR, Ohno H, Scrosati B (2009) in current theoretical investigations as these simulations Ionic-liquid materials for the electrochemical challenges of the can be simply parallelized. future. Nat Mater 8(8):621–629 4. Particular effects of ionic liquids should be tested on Attri P, Sarinont T, Kim M, Amano T, Koga K, Cho AE, Ha Choi E, Shiratani M (2015) Influence of ionic liquid and ionic salt more than one protein as the observation made may be on protein against the reactive species generated using dielectric a peculiarity of the investigated protein. barrier discharge plasma. Sci Rep 5:17781 5. Polarizable force fields to include electronic effects Barducci A, Bussi G, Parrinello M (2008) Well-tempered metadynam- may also be an option. ics: a smoothly converging and tunable free-energy method. Phys Rev Lett 100(2):020603 6. The computational setup is of utmost importance for the Benedetto A, Ballone P (2016) Room temperature ionic liquids meet reliability of the MD results. Hence, simulation system biomolecules: a microscopic view of structure and dynamics. ACS sizes should be large enough to prevent artifacts from Sustain Chem Eng 4(2):392–412 self-interactions. For example, at least five complete Bhargava BL, Klein ML (2009) Aqueous solutions of imida- solvation shells around the biomolecule should be zolium ionic liquids: molecular dynamics studies. Soft Matter 5(18):3475–3480 considered. The simulation period should be at least Booth JJ, Abbott S, Shimizu S (2012) Mechanism of hydrophobic 50 ns in low viscous systems, i.e., mixtures with high drug solubilization by small molecule hydrotropes. J Phys Chem water content, and even much longer with increasing B 116(51):14915–14921 ionic liquid concentration. Of course, the simulation Bornot A, Etchebest C, De Brevern AG (2011) Predicting protein flexibility through the prediction of local structures. Proteins period has also to be prolonged when analyzing 79(3):839–852 folding/unfolding processes. Here, smaller proteins Brandt A, Grasvik ¨ J., Hallett JP, Welton T (2013) Deconstruction may be more useful than larger proteins to cover the of lignocellulosic biomass with ionic liquids. Green Chem complete process or at least major parts of it. 15(3):550–583 838 Biophys Rev (2018) 10:825–840 Burney P, Nordwald EM, Hickman K, Kaar JL, Pfaendtner J (2015) Gutowski KE, Broker GA, Willauer HD, Huddleston JG, Swatloski Molecular dynamics investigation of the ionic liquid/enzyme RP, Holbrey JD, Rogers RD (2003) Controlling the aqueous misci- interface: application to engineering enzyme surface charge. bility of ionic liquids: aqueous biphasic systems of water-miscible Proteins: Struct Funct Bioinf 83(4):670–680 ionic liquids and water-structuring salts for recycle, metathesis, Burney P, Pfaendtner J (2013) Structural and dynamic features of and separations. J Am Chem Soc 125(22):6632–6633 Candida rugosa lipase 1 in water, octane, toluene, and ionic liquids Haberler M, Steinhauser O (2011) On the influence of hydrated BMIM-PF6 and BMIM-NO3. J Phys Chem B 117(9):2662– ionic liquids on the dynamical structure of model proteins: a compu- 2670 tational study. Phys Chem Chem Phys 13(40):17994–18004 Haberler M, Schroder ¨ C, Steinhauser O (2011) Solvation studies of Chang TM, Dang LX, Devanathan R, Dupuis M (2010) Structure a zinc finger protein in hydrated ionic liquids. Phys Chem Chem and dynamics of N, N-Diethyl-N-methylammonium triflate ionic Phys 13(15):6955–6969 liquid, neat and with water, from molecular dynamics simulations. Hallett JP, Welton T (2011) Room-temperature ionic liquids: solvents J Phys Chem A 114(48):12764–12774 for synthesis and catalysis. 2. Chem Rev 111(5):3508–3576 Chen X, Liu J, Wang J (2010) Ionic liquids in the assay of proteins. Heid E, Schroder ¨ C (2016) Computational solvation dynamics Anal Methods 2(9):1222–1226 of oxyquinolinium betaine linked to trehalose. J Chem Phys Chevrot G, Fileti EE, Chaban VV (2015) Enhanced stability of the 145(16):164507 model mini-protein in amino acid ionic liquids and their aqueous Heid E, Schroder C (2017) Effect of a tertiary butyl group on solutions. J Comput Chem 36(27):2044–2051 polar solvation dynamics in aqueous solution: a computational Claudio ´ AFM, Freire MG, Freire CS, Silvestre AJ, Coutinho JA (2010) approach. J Phys Chem B 121(41):9639–9646 Extraction of vanillin using ionic-liquid-based aqueous two-phase Heid E, Schroder C (2018) Solvation dynamics in polar solvents and systems. Sep Purif Technol 75(1):39–47 imidazolium ionic liquids: failure of linear response approxima- Claudio ´ AFM, Neves MC, Shimizu K, Lopes JNC, Freire MG, tions. Phys Chem Chem Phys 20(7):5246–5255 Coutinho JAP (2015) The magic of aqueous solutions of ionic Holbrey JD, Seddon KR (1999) Ionic liquids. Clean Technol Environ liquids: ionic liquids as a powerful class of catanionic hydrotropes. Policy 1(4):223–236 Green Chem 17(7):3948–3963 Itoh T (2017) Ionic liquids as tool to improve enzymatic organic Constantinescu D, Weingartner ¨ H, Herrmann C (2007) Protein synthesis. Chem Rev 117(15):10567–10607 denaturation by ionic liquids and the hofmeister series: a case Jaeger VW, Pfaendtner J (2013) Structure, dynamics, and activity of study of aqueous solutions of ribonuclease–A. Angew Chem Int xylanase solvated in binary mixtures of ionic liquid and water. Ed 46(46):8887–8889 ACS Chem Biol 8(6):1179–1186 Cull S, Holbrey J, Vargas-Mora V, Seddon K, Lye G (2000) Room- Jaeger V, Burney P, Pfaendtner J (2015) Comparison of three temperature ionic liquids as replacements for organic solvents in ionic liquid-tolerant cellulases by molecular dynamics. Biophys J multiphase bioprocess operations. Biotechnol Bioeng 69(2):227– 108(4):880–892 Kaar JL, Jesionowski AM, Berberich JA, Moulton R, Russell AJ Diddens D, Lesch V, Heuer A, Smiatek J (2017) Aqueous ionic (2003) Impact of ionic liquid physical properties on lipase activity liquids and their influence on peptide conformations: denaturation and stability. J Am Chem Soc 125(14):4125–4131 and dehydration mechanisms. Phys Chem Chem Phys 19:20430– Kelkar MS, Maginn EJ (2007) Effect of temperature and water content on the shear viscosity of the ionic liquid 1-ethyl-3- Dreyer S, Kragl U (2008) Ionic liquids for aqueous two-phase methylimidazolium bis(trifluoromethanesulfonyl)imide as studied extraction and stabilization of enzymes. Biotechnol Bioeng by atomistic simulations. J Phys Chem B 111(18):4867–4876 99(6):1416–1424 Kilpelainen ¨ I, Xie H, King A, Granstrom M, Heikkinen S, Earle MJ, Seddon KR (2009) Ionic liquids. Green solvents for the Argyropoulos DS (2007) Dissolution of wood in ionic liquids. J future. Pure Appl Chem 72(7):1391–1398 Agric Food Chem 55(22):9142–9148 Eastoe J, Hatzopoulos MH, Dowding PJ (2011) Action of hydrotropes King AWT, Asikkala J, Mutikainen I, Jarvi ¨ P, Kilpelainen ¨ I and alkyl-hydrotropes. Soft Matter 7(13):5917–5925 (2011) Fort DA, Remsing RC, Swatloski RP, Moyna P, Moyna G, Rogers RD Distillable acid–base conjugate ionic liquids for cellulose (2007) Can ionic liquids dissolve wood? Processing and analysis dissolution and processing. Angew Chem Int Ed 50(28):6301– of lignocellulosic materials with 1-n-butyl-3-methylimidazolium 6305 chloride. Green Chem 9(1):63–69 Kirkwood JG, Buff FP (1951) The statistical mechanical theory of Freemantle M (1998) Designer solvents. Chem Eng News 76(13):32– solutions. I. J Chem Phys 19(6):774–777 37 Klahn ¨ M, Lim GS, Wu P (2011a) How ion properties determine the Freire MG, Claudio ´ AFM, Araujo ´ JMM, Coutinho JAP, Marrucho stability of a lipase enzyme in ionic liquids: a molecular dynamics IM, Lopes JNC, Rebelo LPN (2012) Aqueous biphasic systems: study. Phys Chem Chem Phys 13(41):18647–18660 a boost brought about by using ionic liquids. Chem Soc Rev Klahn ¨ M, Lim GS, Seduraman A, Wu P (2011b) On the different roles 41(14):4966–4995 of anions and cations in the solvation of enzymes in ionic liquids. Fujita K, MacFarlane DR, Forsyth M, Yoshizawa-Fujita M, Murata Phys Chem Chem Phys 13(4):1649–1662 K, Nakamura N, Ohno H (2007) Solubility and stability of Kragl U, Eckstein M, Kaftzik N (2002) Enzyme catalysis in ionic cytochrome c in hydrated ionic liquids: effect of oxo acid residues liquids. Curr Opin Biotechnol 13(6):565–571 and kosmotropicity. Biomacromolecules 8(7):2080–2086 Kufareva I, Abagyan R (2012) Methods of protein structure Gao WW, Zhang FX, Zhang GX, Zhou CH (2015) Key factors comparison. Methods Mol Biol 857:231–257 affecting the activity and stability of enzymes in ionic liquids and Kunz W, Hackl ¨ K (2016) The hype with ionic liquids as solvents. novel applications in biocatalysis. Biochem Eng J 99:67–84 Chem Phys Lett 661:6–12 Ghosh S, Parui S, Jana B, Bhattacharyya K (2015) Ionic liquid Lai JQ, Li Z, Lu ¨ YH, Yang Z (2011) Specific ion effects of ionic liquids induced dehydration and domain closure in lysozyme: FCS and on enzyme activity and stability. Green Chem 13(7):1860–1868 MD simulation. J Chem Phys 143(12):125103 Lau RM, Sorgedrager MJ, Carrea G, Rantwijk FV, Secundo F, Gupta KM, Hu Z, Jiang J (2013) Molecular insight into cellulose Sheldon RA (2004) Dissolution of Candida antarctica lipase B regeneration from a cellulose/ionic liquid mixture: effects of water in ionic liquids: effects on structure and activity. Green Chem concentration and temperature. RSC Adv 3(13):4425–4433 6(9):483–487 Biophys Rev (2018) 10:825–840 839 Lee SY, Khoiroh I, Ooi CW, Ling TC, Show PL (2017) Recent Persson M, Bornscheuer UT (2003) Increased stability of an esterase advances in protein extraction using ionic liquid-based aqueous from Bacillus stearothermophilus in ionic liquids as compared to two-phase systems. Sep Purif Rev 46(4):291–304 organic solvents. J Mol Catal B: Enzym 22(1-2):21–27 Lesch V, Heuer A, Tatsis VA, Holm C, Smiatek J (2015) Peptides in Pierce V, Kang M, Aburi M, Weerasinghe S, Smith PE (2008) Recent the presence of aqueous ionic liquids: tunable co-solutes as denaturants applications of Kirkwood-Buff theory to biological systems. Cell or protectants? Phys Chem Chem Phys 17(39):26049–26053 Biochem Biophys 50:1–22 Liang M, Zhang X, Kaintz A, Ernsting NP, Maroncelli M (2014) Pinkert A, Marsh KN, Pang S, Staiger MP (2009) Ionic liquids and Solvation dynamics in a prototypical ionic liquid + dipolar aprotic their interaction with cellulose. Chem Rev 109(12):6712–6728 liquid mixture: 1-butyl-3-methylimidazolium tetrafluoroborate + Pitera JW (2014) Expected distributions of root-mean-square positio- acetonitrile. J Phys Chem B 118(5):1340–1352 nal deviations in proteins. J Phys Chem B 118(24):6526–6530 Lou WY, Zong MH, Smith TJ, Wu H, Wang JF (2006) Impact of Plechkova NV, Seddon KR (2008) Applications of ionic liquids in the ionic liquids on papain: an investigation of structure–function chemical industry. Chem Soc Rev 37(1):123–150 relationships. Green Chem 8(6):509–512 Qin Y, Lu X, Sun N, Rogers RD (2010) Dissolution or extraction Lozano P, de Diego T, Guegan JP, Vaultier M, Iborra JL of crustacean shells using ionic liquids to obtain high molecular (2001) Stabilization of α-chymotrypsin by ionic liquids in weight purified chitin and direct production of chitin films and transesterification reactions. Biotechnol Bioeng 75(5):563–569 fibers. Green Chem 12(6):968 Maki-Arv ¨ ela P, Anugwom I, Virtanen P, Sjoholm ¨ R, Mikkola JP Raut DG, Sundman O, Su W, Virtanen P, Sugano Y, Kordas K, (2010) Dissolution of lignocellulosic materials and its constituents Mikkola JP (2015) A morpholinium ionic liquid for cellulose using ionic liquids—a review. Ind Crops Prod 32(3):175–201 dissolution. Carbohydr Polym 130:18–25 Mandal PK, Samanta A (2005) Fluorescence studies in a pyrrolidinium Roy D, Maroncelli M (2012) Simulations of solvation and solvation ionic liquid: polarity of the medium and solvation dynamics. J dynamics in an idealized ionic liquid model. J Phys Chem B Phys Chem B 109(31):15172–15177 116(20):5951–5970 McCammon JA, Gelin BR, Karplus M (1977) Dynamics of folded Sajadi M, Berndt F, Richter C, Gerecke M, Mahrwald R, Ernsting NP proteins. Nature 267(5612):585–590 (2014) Observing the hydration layer of trehalose with a linked Micaelo ˆ NM, Soares CM (2008) Protein structure and dynamics molecular terahertz probe. J Phys Chem Lett 5(11):1845–1849 in ionic liquids. Insights from molecular dynamics simulation Samanta A (2010) Solvation dynamics in ionic liquids: what we have studies. J Phys Chem B 112(9):2566–2572 learned from the dynamic fluorescence stokes shift studies. J Phys Mikkola JP, Kirilin A, Tuuf JC, Pranovich A, Holmbom B, Kustov Chem Lett 1(10):1557–1562 LM, Murzin DY, Salmi T (2007) Ultrasound enhancement of Sasaki H (2015) Simulating hype cycle curves with mathematical cellulose processing in ionic liquids: from dissolution towards functions : some examples of high-tech trends in Japan. Int J Inf functionalization. Green Chem 9(11):1229–1237 Technol 7(2):1–12 Moniruzzaman M, Nakashima K, Kamiya N, Goto M (2010) Recent Schmollngruber M, Schroder ¨ C, Steinhauser O (2013) Polarization advances of enzymatic reactions in ionic liquids. Biochem Eng J effects on the solvation dynamics of coumarin C153 in ionic 48(3):295–314 liquids: components and their cross-correlations. J Chem Phys Naushad M, ALOthman ZA, Khan AB, Ali M (2012) Effect of ionic 138(20):204504 liquid on activity, stability, and structure of enzymes: a review. Int Schofer ¨ SH, Kaftzik N, Wasserscheid P, Kragl U (2001) Enzyme J Biol Macromol 51(4):555–560 catalysis in ionic liquids: lipase catalysed kinetic resolution of 1- Neumayr G, Rudas T, Steinhauser O (2010) Global and local phenylethanol with improved enantioselectivity. Chem Commun Voronoi analysis of solvation shells of proteins. J Chem Phys 0(5):425–426 133(8):084108 Schroder ¨ C (2017) Proteins in ionic liquids: current status of Omta AW, Kropman MF, Woutersen S, Bakker HJ (2003) Negligible experiments and simulations. Top Curr Chem 375(2) effect of ions on the hydrogen-bond structure in liquid water. Schroder ¨ C, Rudas T, Neumayr G, Benkner S, Steinhauser O (2007) Science 301(5631):347–349 On the collective network of ionic liquid/water mixtures. I. Park S, Kazlauskas RJ (2003) Biocatalysis in ionic liquids— Orientational structure. J Chem Phys 127(23):234503 advantages beyond green technology. Curr Opin Biotechnol Schroder ¨ C, Hunger J, Stoppa A, Buchner R, Steinhauser O (2008) 14(4):432–437 On the collective network of ionic liquid/water mixtures. II. Parviainen A, King AWT, Mutikainen I, Hummel M, Selg C, Hauru Decomposition and interpretation of dielectric spectra. J Chem LKJ, Sixta H, Kilpelainen ¨ I (2013) Predicting cellulose solvating Phys 129(18):184501 capabilities of acid—base conjugate ionic liquids. ChemSusChem Schroder ¨ C, Neumayr G, Steinhauser O (2009) On the collective 6(11):2161–2169 network of ionic liquid/water mixtures. III. Structural analysis of Parvulescu ˆ VI, Hardacre C (2007) Catalysis in ionic liquids. Chem ionic liquids on the basis of Voronoi decomposition. J. Chem. Phys Rev 107(6):2615–2665 130(19):194503 Patel R, Kumari M, Khan AB (2014) Recent advances in the Schroder ¨ C, Sega M, Schmollngruber M, Gailberger E, Braun applications of ionic liquids in protein stability and activity: a D, Steinhauser O (2014) On the collective network of ionic review. Appl Biochem Biotechnol 172(8):3701–3720 liquid/water mixtures. IV. Kinetic and rotational depolarization. J. Pei Y, Wang J, Wu K, Xuan X, Lu X (2009) Ionic liquid-based aqueous Chem. Phys 140(20):204505 two-phase extraction of selected proteins. Sep Purif Technol Schroder ¨ C, Steinhauser O, Sasisanker P, Weingartner ¨ H (2015) Ori- 64(3):288–295 entational alignment of amyloidogenic proteins in pre-aggregated Pereira JFB, Lima AS, Freire MG, Coutinho JAP (2010) Ionic liquids solutions. Phys Rev Lett 114(12). https://doi.org/10.1007/s41061- as adjuvants for the tailored extraction of biomolecules in aqueous 017-0110-2 biphasic systems. Green Chem 12(9):1661–1669 Seduraman A, Wu P, Klahn ¨ M (2012) Extraction of tryptophan with Pereira JFB, Ventura SPM, e Silva FA, Shahriari S, Freire MG, ionic liquids studied with molecular dynamics simulations. J Phys Coutinho JAP (2013) Aqueous biphasic systems composed of Chem B 116(1):296–304 ionic liquids and polymers: a platform for the purification of Senske M, Constantinescu-Aruxandei D, Havenith M, Herrmann biomolecules. Sep Purif Technol 113:83–89 C, Weingartner ¨ H, Ebbinghaus S (2016) The temperature 840 Biophys Rev (2018) 10:825–840 dependence of the Hofmeister series: thermodynamic fingerprints liquids: a review of molecular dynamics simulations. J Mol Liq of cosolute—protein interactions. Phys Chem Chem Phys 210:178–188 18(43):29698–29708 Ventura SPM, e Silva FA, Quental MV, Mondal D, Freire Shao Q (2013) On the influence of hydrated imidazolium-based MG, Coutinho JAP (2017) Ionic-liquid-mediated extraction and ionic liquid on protein structure stability: a molecular dynamics separation processes for bioactive compounds: past, present, and simulation study. J Chem Phys 139(11):115102 future trends. Chem Rev 117(10):6984–7052 Shimizu S, Matubayasi N (2014) Hydrotropy: Monomer–Micelle Ventura SPM, Neves CMSS, Freire MG, Marrucho IM, Oliveira equilibrium and minimum hydrotrope concentration. J Phys Chem J, Coutinho JAP (2009) Evaluation of anion influence on the B 118(35):10515–10524 formation and extraction capacity of ionic-liquid-based aqueous Smiatek J (2017) Aqueous ionic liquids and their effects on protein biphasic systems. J Phys Chem B 113(27):9304–9310 structures: an overview on recent theoretical and experimental Wang H, Gurau G, Rogers RD (2012) Ionic liquid processing of results. J Phys: Condens Matter 29(23):233001 cellulose. Chem Soc Rev 41(4):1519–1537 Smith PE (2004) Cosolvent interactions with biomolecules: relating Wasserscheid P (2017) Talk at the 655th Heraeus symposium: Surfaces computer simulation data to experimental thermodynamic data. J and Interfaces of Ionic Liquids Phys Chem B 108:18716–18724 Weingartner ¨ H, Cabrele C, Herrmann C (2012) How ionic liquids Steudte S, Neumann J, Bottin-Weber U, Diedenhofen M, Arning can help to stabilize native proteins. Phys Chem Chem Phys J, Stepnowski P, Stolte S (2012) Hydrolysis study of fluo- 14(2):415–426 roorganic and cyano-based ionic liquid anions—consequences Welton T (1999) Room-temperature ionic liquids. Solvents for for operational safety and environmental stability. Green Chem synthesis and catalysis. Chem Rev 99(8):2071–2084 14(9):2474–2483 Yang Z, Pan W (2005) Ionic liquids: green solvents for nonaqueous Summers CA, Flowers RA (2000) Protein renaturation by the liquid biocatalysis. Enzyme Microb Technol 37(1):19–28 organic salt ethylammonium nitrate. Protein Sci 9(10):2001–2008 Zeindlhofer V, Khlan D, Bica K, Schroder ¨ C (2017) Computational Swatloski RP, Spear SK, Holbrey JD, Rogers RD (2002) Dissolution analysis of the solvation of coffee ingredients in aqueous ionic of cellose with ionic liquids. J Am Chem Soc 124(18):4974–4975 liquid mixtures. RSC Adv 7(6):3495–3504 Tan SSY, MacFarlane DR, Upfal J, Edye LA, Doherty WOS, Patti AF, Zeindlhofer V, Berger M, Steinhauser O, Schroder ¨ C (2018) A shell- Pringle JM, Scott JL (2009) Extraction of lignin from lignocel- resolved analysis of preferential solvation of coffee ingredients in lulose at atmospheric pressure using alkylbenzenesulfonate ionic aqueous mixtures of the ionic liquid 1-ethyl-3-methylimidazolium liquid. Green Chem 11(3):339–345 acetate. J Chem Phys 148(19):193819 Terranova ZL, Corcelli SA (2013) On the mechanism of solvation Zhang Y, Cremer P (2006) Interactions between macromolecules and dynamics in imidazolium-based ionic liquids. J Phys Chem B ions: the Hofmeister series. Curr Opin Chem Biol 10(6):658–663 117(49):15659–15666 Zhang Y, Cremer PS (2009) The inverse and direct Hofmeister Thuy Pham TP, Cho CW, Yun YS (2010) Environmental fate and series for lysozyme. Proc Natl Acad Sci USA 106(36):15249– toxicity of ionic liquids: a review. Water Res 44(2):352–372 15253 Tome ´ LIN, Jorge M, Gomes JRB, Coutinho JAP (2012) Molecular Zhang J, Wu J, Yu J, Zhang X, He J, Zhang J (2017) Application dynamics simulation studies of the interactions between ionic of ionic liquids for dissolving cellulose and fabricating cellulose- liquids and amino acids in aqueous solution. J Phys Chem B based materials: state of the art and future trends. Mater Chem 116(6):1831–1842 Front 1(7):1273–1290 Tung HJ, Pfaendtner J (2016) Kinetics and mechanism of ionic-liquid Zhao H (2010) Methods for stabilizing and activating enzymes in ionic induced protein unfolding: application to the model protein HP35. liquids—a review. J Chem Technol Biotechnol 85(7):891–907 Mol Syst Des Eng 1(4):382–390 Zhao H (2016) Protein stabilization and enzyme activation in ionic Ulbert O, Belafi-Bak ´ o ´ K, Tonova K, Gubicza L (2005) Thermal liquids: specific ion effects: protein stabilization and enzyme acti- stability enhancement of Candida rugosa lipase using ionic liquids. vation in ionic liquids. J Chem Technol Biotechnol 91(1):25–50 Biocatal Biotransform 23(3–4):177–183 Zhu S, Wu Y, Chen Q, Yu Z, Wang C, Jin S, Ding Y, Wu G (2006) van Rantwijk F, Sheldon RA (2007) Biocatalysis in ionic liquids. Dissolution of cellulose with ionic liquids and its application: a Chem Rev 107(6):2757–2785 mini-review. Green Chem 8(4):325–327 Varela LM, Mendez-Morales ´ T, Carrete J, Gomez-Gonz ´ alez ´ V, Zirbs R, Strassl K, Gaertner P, Schroder ¨ C, Bica K (2013) Exploring Docampo-Alvarez B, Gallego LJ, Cabeza O, Russina O (2015) ionic liquid-biomass interactions: towards the improved isolation Solvation of molecular cosolvents and inorganic salts in ionic of shikimic acid from star anise pods. RSC Adv 3(48):26010 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Biophysical Reviews Springer Journals

Computational solvation analysis of biomolecules in aqueous ionic liquid mixtures

Free
16 pages

Loading next page...
 
/lp/springer_journal/computational-solvation-analysis-of-biomolecules-in-aqueous-ionic-vj9S0CzHST
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Life Sciences; Biochemistry, general; Biological and Medical Physics, Biophysics; Cell Biology; Membrane Biology; Biological Techniques; Nanotechnology
ISSN
1867-2450
eISSN
1867-2469
D.O.I.
10.1007/s12551-018-0416-5
Publisher site
See Article on Publisher Site

Abstract

Based on their tunable properties, ionic liquids attracted significant interest to replace conventional, organic solvents in biomolecular applications. Following a Gartner cycle, the expectations on this new class of solvents dropped after the initial hype due to the high viscosity, hydrolysis, and toxicity problems as well as their high cost. Since not all possible combinations of cations and anions can be tested experimentally, fundamental knowledge on the interaction of the ionic liquid ions with water and with biomolecules is mandatory to optimize the solvation behavior, the biodegradability, and the costs of the ionic liquid. Here, we report on current computational approaches to characterize the impact of the ionic liquid ions on the structure and dynamics of the biomolecule and its solvation layer to explore the full potential of ionic liquids. Keywords Ionic liquid · Biomolecule · MD simulation Introduction in Fig. 1. The initial success of ionic liquids in (bio-) catalysis (van Rantwijk and Sheldon 2007) and solvation Ionic liquids (IL) are a unique class of high-performance (Welton 1999;Gutowskietal. 2003), e.g., dissolution of chemical compounds with many applications in electro- cellulose (Swatloski et al. 2002), served as a technology chemistry (Armand et al. 2009), synthesis (Itoh 2017), trigger. The vast number of 10 possible combinations of catalysis (Parvulescu ˆ and Hardacre 2007; van Rantwijk and cations and anions (Holbrey and Seddon 1999) seemed to Sheldon 2007) as well as solvation (Hallett and Welton promise the design of an optimal solvent for a particular 2011) and extraction processes (Ventura et al. 2017). Their application. As a result, the term “designer solvent” was tunable properties via variation or modification of either the coined (Freemantle 1998) and soon joined by the rash cation or the anion as well as their shared properties such as promise “ILs are green solvents” (Earle and Seddon 2009) low vapor pressure, low flammability, and high thermal and leading to an exponential number of paper submissions to electrochemical stability make them interesting solvents for the journal Green Chemistry. Also, the availability of ILs various applications and started a hype in the beginning of increased due to beginning industrial production (Plechkova the twenty-first century. and Seddon 2008). Kunz and Hackl ¨ (2016) as well as Wasserscheid pointed However, soon, first doubts were cast on the environmen- out that the expectations followed a Gartner cycle as visible tal benefit of ILs (Thuy Pham et al. 2010) and problems on the hydrolysis of some ILs were reported (Steudte et al. 2012). Additionally, many ionic liquids did not have a This article is part of a Special Issue on ‘Ionic Liquids and unique selling point or their improved performance did not Biomolecules’ edited by Antonio Benedetto and Hans-Joachim justify their increased cost (Plechkova and Seddon 2008; Galla Kunz and Hackl ¨ 2016). As a result, the hype cooled down Christian Schroder ¨ in the past decade reaching the so-called Trough of Disillu- christian.schroeder@univie.ac.at sionment of a Gartner cycle. Nevertheless, scientific research did not abate, visible by Faculty of Chemistry, Department of Computational the roughly exponential increase in publications indicated Biological Chemistry, University of Vienna, Wahringerstr ¨ . 17, Vienna, Austria by the blue dashed line in Fig. 1. Here, we normalized 826 Biophys Rev (2018) 10:825–840 Hype stage Implementation stage depicted in Fig. 2. Here, coordination numbers of particular solvent molecules and hydrogen bonding are interesting features to be computed from MD trajectories. Moreover, total publications on IL radial distribution functions (solid black line in Fig. 2) simulations on IL simulations on protein/IL detecting the accumulation or depletion of solvent species at the biomolecular surface and their interpretation in terms of Kirkwood-Buff theory (Lesch et al. 2015; Diddens et al. IL are IL are biodegradable green toxic IL 2017; Smiatek 2017) are reported in “The solvation layer”. Due to the surrounding hydrophilic or hydrophobic sol- vent, the structure and stability of the biomolecular solute may adapt to the environment. Depending on the size and shape of a solute, different mechanisms contribute to solvent-induced changes. Hence, we distinguish between large flexible biomolecules and small rigid drugs, employ- ing different analysis routines as sketched in Fig. 2: Technology Trough of Slope of trigger Disillusionment Enlightment 1. Large proteins, polypeptides, or polymeric biomolecules Fig. 1 Gartner hype cycle on ionic liquids adapted from Kunz and possess a secondary structure which can deform during Hackl ¨ (2016), Wasserscheid (2017) the solvation process (indicated by the gray arrows in Fig. 2). In a hydrophilic environment, for example, the publication growth obtained from a search in the “Web hydrophilic subunits are supposed to be at the surface of the of knowledge” to the current number of publications up biomolecule and in contact with the solvent. Hydropho- to 2017. This way, we can compare the development with bic subunits are covered in the core of the biomolecule. publication efforts concerning simulations of ionic liquids In addition to refolding, the size of a biomolecule shown as an orange dashed line in Fig. 1 which agrees may change when moving from a hydrophilic to a almost quantitatively with the overall publication output on hydrophobic environment or vice versa. Computational ILs. size and shape observables of this biomolecular class Quite generally, hype cycles can be decomposed into two are discussed in “Large flexible proteins”. processes (Sasaki 2015): a hype stage and an implementa- 2. In contrast, small rigid drugs cannot increase their size or tion stage. The latter can be fitted by an S-shaped Gompertz change their shape significantly. Here, other mechanisms function like solute aggregation can be monitored by molecular −k(t −t ) −e dynamics. The inclusion of the solute in the solvent net- f(t) = a · e.(1) work can also be computed by free energy calculations. Taking the publication efforts depicted in Fig. 1 as a Furthermore, functional groups and their respective measure of implementation, the “Slope of Enlightenment” influence on the solvation are of significant importance. phase can be extrapolated. In addition to the amplitude a This will be discussed in “Small rigid drugs.” and the stretching factor k, the inflection point t marks the year in which the implementations start to level off. In the case of the ILs, the year 2020 is obtained for both The solvation layer the total publications and the publications concerning the simulation of ionic liquids. The “Plateau of Productivity” The tunable solvation properties (Anderson et al. 2002; will be reached in twenty to thirty years based on the current Gutowski et al. 2003) of ILs gave reasons for the hype on data. ILs at the start of this century. They were considered as In order to get there, many scientific questions have green replacements for organic solvents in general (Cull yet to be answered. In particular, the knowledge gained et al. 2000) and for biological processes in particular (Yang by simulations on biomolecules in ionic liquids and their and Pan 2005; Park and Kazlauskas 2003). For example, mixtures is lagging behind the general ionic liquid trend as Rogers and co-workers reported on the dissolution of visible by the green dotted line in Fig. 1. The current review cellulose in 2002 (Swatloski et al. 2002). In 2005, BASF tries to summarize the current efforts of biomolecular licensed the exclusive use of various intellectual property solvation in ionic liquids and their mixtures from a rights from this group. In 2007, the Rogers group continued molecular dynamics (MD) perspective. their research on dissolution of lignocellulosic materials Solvation starts at the first solvation layer which is the (Fort et al. 2007). Mikkola et al. (2007), Anugwom et al. transition region between the bulk solvent and the solute as (2012), Raut et al. (2015), Kilpelainen ¨ et al. (2007), today inflection point Biophys Rev (2018) 10:825–840 827 Fig. 2 Computational analysis of the solvation interaction between biomolecules, their solvation layers, and the bulk solvent King et al. (2011), and Parviainen et al. (2013)worked on protein properties such as stability or solubility (Zhang on cellulose processing. Excellent reviews summarize the and Cremer 2006;Schroder ¨ 2017). Ions are classified either current development (Pinkert et al. 2009; Brandt et al. 2013; as kosmotropes (structure-makers) or chaotropes (structure- Maki-Arv ¨ ela et al. 2010; Zhang et al. 2017)and show that breakers) by their ability to influence water structure (Tung we are still in the “Slope of Enlightenment” phase during and Pfaendtner 2016; Constantinescu et al. 2007; Zhang and the implementation stage. Cremer 2006). Generally, kosmotropic anions stabilize pro- Due the high cost of ILs compared to the conventional teins, while chaotropic anions lead to destabilization (Tung alternatives, not many large-scale applications were intro- and Pfaendtner 2016). Still, exceptions to the Hofmeister duced. In order to reduce the costs, ILs were mixed with series have been observed (Zhang and Cremer 2009), not other cheap solvents, in particular water. Since only few pro- only due to the fact that specific protein properties such teins are soluble in pure ionic liquids (Shao 2013;Kragl as charge and surface characteristics (Schroder ¨ 2017; Con- et al. 2002), homogeneous and heterogeneous aqueous ionic stantinescu et al. 2007; Tung and Pfaendtner 2016) are not liquid mixtures were used (Kragl et al. 2002). The first ionic considered. Moreover, the ability of ions to influence bulk liquid-based aqueous biphasic system consisted of 1-butyl- water structure in a kosmo- or chaotropic manner has been 3-imidazolium chloride mixture in aqueous K PO and was questioned (Omta et al. 2003). 3 4 reported by Rogers and co-workers in 2003 (Gutowski et al. Less publications exist on the computational research of 2003) and may be considered as a “Technology Trigger” aqueous ionic liquid mixtures (Bhargava and Klein 2009; (see Fig. 1). Recently, biphasic systems regained interest, in Varela et al. 2015; Chang et al. 2010) without additional particular by the group of Coutinho for extraction (Ventura solutes as these studies focus on more fundamental prop- et al. 2009, 2017; Pei et al. 2009; Dreyer and Kragl 2008; erties of IL mixtures rather than a particular application. Pereira et al. 2010; Lee et al. 2017;Claudio ´ et al. 2010)and The effect of water on the viscosity of an IL/water mixture for purification of biomolecules (Pereira et al. 2013). An was studied by Kelkar and Maginn (2007). Furthermore, the excellent review was given by Freire et al. (2012). alkyl chain length of the IL cations changed the collective Although of great importance for the rational choice structure of the aqueous IL mixture (Bhargava and Klein of an ionic liquid for any given application (Weingartner ¨ 2009). We started our computational analysis on the col- et al. 2012), predicting the effect of an ionic liquid on lective network between water and various ionic liquids in the stability of a specific protein still remains difficult 2007 (Schroder ¨ et al. 2007, 2008, 2009, 2014) with a focus (Tung and Pfaendtner 2016) due to the complex and lit- on dielectric properties. Smiatek and co-workers published tle understood molecular interactions between ionic liquid interesting papers on Kirkwood-Buff analysis of protein sol- and protein (Burney et al. 2015) demonstrating that we vation in IL/water mixtures (Diddens et al. 2017; Smiatek are still in the “Slope of Enlightenment” phase and have 2017). Gupta and co-workers reported MD simulations on not reached the “Plateau of Productivity” in Fig. 1.A the interaction of cellulose with ionic liquids (Gupta et al. popular classification of the effect of ionic liquids on pro- 2013). As a complete review of computational analysis of teins already mentioned is the extension of the well-known interactions of biomolecules with IL mixtures is out of the Hofmeister series (Zhang and Cremer 2006) to aqueous scope of this review, we will focus on the basic techniques to ionic liquid solutions (Constantinescu et al. 2007). The encourage other authors to contribute to the understanding Hofmeister series ranks ions by their specific ion effect of biomolecular solvation in ionic liquid mixtures by MD 828 Biophys Rev (2018) 10:825–840 simulations. Hereafter, we will give a compact overview of of xylanase II from trichoderma longibrachiatum in solvation layer properties obtainable by MD simulations as aqueous [C mim][EtSO ]and [C mim][OAc] (Jaeger and 2 4 2 well as their interpretation. Pfaendtner 2013). Steinhauser and co-workers noted that Eq. 3 not only Spatial distribution of the solvent requires a meaningful definition of a solvation shell radius R but also assumes sphericity of the solutes (Haberler The spatial distribution of molecular species j around and Steinhauser 2011; Haberler et al. 2011; Zeindlhofer et al. 2018). Although this approach is feasible for many a reference site i is often characterized by the radial distribution function g (r), small solutes, it may give highly inaccurate results in case ij of anisotropic solutes (Zeindlhofer et al. 2017, 2018) like δ(r − r ) ij large proteins (Neumayr et al. 2010) due to non-symmetric j 1 g (r) = · (2) ij excluded volume effects. Furthermore, for heterogeneous 4πr dr· ρ solvents with different molecular sizes, more than one shell which is the ratio between the local density of j in a radius would be necessary to describe a solvation shell spherical shell 4πr dr and the global density ρ of species appropriately (Haberler et al. 2011). j . The distance r is usually defined between the center of ij masses of the molecules i and j or between the coordinates Voronoi-based coordination numbers of particular atoms belonging to these molecules. An alternative method of spatial decomposition is Voronoi Coordination numbers decomposition (Schroder ¨ et al. 2009). It is parameter- free and also allows for straightforward decomposition of To calculate the coordination number n of species around ij multiple solvation shells. In Voronoi tessellation, each point the biomolecule i, spherical integration of the radial in space representing an atomic coordinate is assigned an distribution function is commonly used: irregular polyhedron that contains all space closer to its associated reference point than to any other point in the system. If two points share a polyhedron face, they are n = 4πr g (r)dr (3) ij ij defined as neighbors. Solvent molecules that share a face with the biomolecule are considered first-shell members, Micaelo ˆ and Soares found that the coordination num- molecules that share a face with first-shell members are ber of water around cutinase in aqueous 1-butyl-3- considered second shell, and so on. This analysis can methylimidazolium nitrate [C mim][NO ] and 1-butyl- be performed for each time step during a trajectory to 4 3 3-methylimidazolium hexafluorophosphate [C mim][PF ] yield mean residence times of particular species at the 4 6 mixtures is reduced at the protein surface (Micaelo ˆ and surface of the biomolecule (Haberler and Steinhauser 2011; Soares 2008). Similar removal of water was observed for Haberler et al. 2011). This shell assignment can also be hydrated Candida rugosa lipase 1 in [C mim][PF ]and applied to the radial distribution function (Neumayr et al. 4 6 [C mim][NO ] (Burney and Pfaendtner 2013). 2010; Zeindlhofer et al. 2018) as depicted in Fig. 3. 4 3 A problem that may arise when comparing coordination The decomposition of the displayed g (r) into spherical ij numbers of different IL ions is the fact that the coordination solvation shells discussed in the last section is not possible number depends on ion density. For the analysis of cation and anion coordination numbers around CAL-B, Klahn ¨ et al. (2011a) normalized the coordination number by the average number of the respective species within 10 Aof the protein to facilitate comparison. They observed the highest ion densities around charged residues. Interestingly, although density of cations and anions was highest around oppositely charged residues, densities around residues with the same charge was also higher than for polar and non- charged residues (Schroder ¨ 2017;Klahn ¨ et al. 2011a), which is an effect of the strong interionic interactions in ionic liquids (Klahn ¨ et al. 2011b). Furthermore, a diffusion Fig. 3 Radial distribution function g (r) of water oxygens (j ) around ij of cations into the active site of the protein occurred, the center of mass of calbindin i and the corresponding decomposition possibly hindering substrate access. A similar occupation into Voronoi shells. Reproduced from Neumayr et al. (2010) with of the active site by cations was observed in simulations permission of AIP Publishing Biophys Rev (2018) 10:825–840 829 −19 as no clear first minimum of the radial distribution can be e = 1.6022 × 10 Asand n± equals the number detected. Furthermore, the different shells overlap to a large of indistinguishable ion species. The binding coefficient extent. Consequently, the spherical integration performed in determines the transfer free energy (Smiatek 2017). Eq. 3 will not yield a correct coordination number. However, F (R) =−k T · ν (R) (7) 23 B 23 integrating the shell contributions in Fig. 3 results in the exact which estimates the excess energy at a given temperature coordination number obtained from counting neighbors by shared faces of the solute and solvent polyhedra. T that is needed to transfer the ionic liquid from infinite Employing Voronoi decomposition, Haberler and Stein- distance to close proximity around the biomolecule. hauser confirmed a surplus of C mim in the first solvation In a MD study of a small β -hairpin peptide in shell of various proteins despite the positive net charge of aqueous [C mim][OAc], where both native and denaturated the protein (Haberler and Steinhauser 2011; Haberler et al. conformations of the peptide were simulated, Kirkwood- 2011). This is very likely due to expulsion of the cations Buff preferential binding coefficients revealed that the unfolded conformation attracts more anions while no from the strong anion-water network, as also observed in other studies (Schroder ¨ et al. 2007, 2009; Haberler and difference in cation affinity is observed (Lesch et al. 2015). This agrees with the fact that the anionic influence on the Steinhauser 2011). Interestingly, the mean residence time of the anions at the protein surface is longer compared denaturation process is stronger (Klahn ¨ et al. 2011a). The unfolding effect of [C mim][OAc] on the peptide was also to the cations (Haberler and Steinhauser 2011) which can be explained by stronger interactions with the oppositely confirmed by experiment (Patel et al. 2014). In a follow- charged amino acid (Schroder ¨ 2017; Jaeger and Pfaendtner up publication by the same group, a similar approach was 2013;Klahn ¨ et al. 2011a;Micaelo ˆ and Soares 2008). used to study the behavior of the same β -hairpin peptide in aqueous [C mim][BF ]and[C mim]Cl. The binding 2 4 2 Kirkwood-Buff theory strength order of the anions was identified as OAc − − − BF > Cl , and again it was found that OAc strongly binds to the denatured state, which is not observable for Another approach utilizing radial distribution functions is − − the so-called Kirkwood-Buff theory, which relates structural BF and Cl . The authors suggested that the larger acetate anion exhibits a conformation-dependent binding behavior properties of the solutions due to specific interactions with the protein (Diddens et al. R 2017). Furthermore, they found only negligible dependence of the binding behavior of C mim on the anion. G (R) = g (r) − 1 4πr dr, (4) ij ij The interactions of caffeine, gallic acid, protocatechuic acid, and quercetin with aqueous mixtures of the ionic liquid [C mim][OAc] were studied by a shell-resolved to thermodynamic quantities (Kirkwood and Buff 1951). approach using Voronoi tessellation. As shown in the The Kirkwood-Buff integral is often interpreted as excess respective publications (Zeindlhofer et al. 2017, 2018), the volume around a solute species when compared to an investigated biomolecules are flat, aromatic systems of high ideal solution. It is also possible to compute Kirkwood- anisotropy, which makes Voronoi tessellation necessary Buff integrals of distinct solvation shells based on Voronoi to distinguish between several solvation shells. Since tessellation (Zeindlhofer et al. 2018). Voronoi tessellation allows straightforward calculation of As a standard notation, i, j = 1 ... 3 denote water, solvent shell volumes, a shell-wise calculation of Kirkwood- the biomolecular solute, and the co-solvent (in our case Buff interaction parameters analogous to Kirkwood-Buff the ionic liquid), respectively. Distance-dependent binding integrals (Zeindlhofer et al. 2018) can be computed. coefficients ν (R) between the solute i = 2 and the ionic However, the comparison of Kirkwood-Buff interaction liquid j = 3 can be obtained as follows: parameters between solutes of different size is complicated, since they depend on the solute size. In contrast, the shell- solute 1 q wise potential of mean force (Zeindlhofer et al. 2017) ν (R) = ρ (G (R) − G (R)) − (5) 23 3 23 21 n± e measures solvent affinities of solutes differing in size via solute 1 ρ q 3 the ratio of shell to bulk concentration of a species. = n (R) − n (R) − (6) 23 21 n± ρ e Solvation dynamics using coordination numbers defined in Eq.3 as well as The role of ionic liquids in solvation dynamics (Mandal the Donnan equilibrium condition (Pierce et al. 2008; solute and Samanta 2005; Samanta 2010; Roy and Maroncelli Smith 2004; Smiatek 2017). Here q corresponds to the 2012; Terranova and Corcelli 2013; Liang et al. 2014;Heid charge of the biomolecule in units of the elementary charge 830 Biophys Rev (2018) 10:825–840 and Schroder ¨ 2018) as well as the solvation dynamics of In contrast to neutral liquids like water, ionic liquids can biomolecules (Sajadi et al. 2014; Heid and Schroder ¨ 2016; also respond to the change in the local electric field via 2017) has attracted significant interest over the past decade. translation (Terranova and Corcelli 2013; Schmollngruber The time-dependent Stokes shift et al. 2013), in particular the anions. The response of induced solvent dipoles is restricted to the first solvation ν(t) − ν(∞) S(t) = (8) shell (Schmollngruber et al. 2013). Using polarizable non- ν(0) − ν(∞) equilibrium molecular dynamics simulations, not only the monitors the transient behavior of the solvent relaxation normalized Stokes shift relaxation function but also the pure after an electronic excitation of a dissolved chromophore. elec elec shift U (t ) − U (∞) can be reproduced (Heid and THz absorption bands of biomolecular hydration layers Schroder ¨ 2018). are generally swamped by absorption from bulk solvent. However, linking the chromophore to the biomolecule, the solvation properties of the immediate solvent molecules can Large flexible proteins be experimentally studied (Sajadi et al. 2014)asshown for the chromophore oxyquinolinium betaine attached to One of the “Technology Triggers” in Fig. 1 for protein sta- trehalose in Fig. 4. bility was published in 2000 by Summers et al. who showed In MD simulations, this experiment can be modeled by that ethylammonium nitrate can successfully refold dena- equilibrium (Roy and Maroncelli 2012; Schmollngruber tured egg-white lysozyme (Summers and Flowers 2000). et al. 2013; Terranova and Corcelli 2013) and non- Several groups observed that ionic liquids can significantly equilibrium simulations (Heid and Schroder ¨ 2016, 2017, stabilize certain proteins and enzymes (Lozano et al. 2001; 2018). However, the latter approach seems to be more Fujita et al. 2007;Weingartner ¨ et al. 2012), especially lipases appropriate (Heid and Schroder ¨ 2018). The change of the (Kaar et al. 2003; Ulbert et al. 2005; Lai et al. 2011), even at elec Coulomb interaction energy U (t ) of the chromophore high temperatures (Ulbert et al. 2005). In some cases, enzy- atoms j with the solvent atoms i me activity, enantioselectivity, and regioselectivity in ionic 1 q · q liquids are increased compared to conventional organic sol- j i elec U (t ) = (9) vents (Schofer ¨ et al. 2001) as discussed in more detail in 4π r (t ) 0 ij j i several reviews (Kragl et al. 2002; Park and Kazlauskas relaxes during the non-equilibrium simulation after switch- 2003;Yangand Pan 2005; van Rantwijk and Sheldon 2007; ing the partial charges q of the solute from ground to j Moniruzzaman et al. 2010; Lai et al. 2011; Naushad et al. excited state values: 2012; Gao et al. 2015;Chenetal. 2010;Zhao 2010; elec elec Weingartner ¨ et al. 2012; Patel et al. 2014; Benedetto and U (t ) − U (∞) S(t) = (10) Ballone 2016; Smiatek 2017;Schroder ¨ 2017;Zhao 2016). elec elec U (0) − U (∞) An observation that many groups agree with is that the presence of water in the ionic liquid is crucial for protein stability (Klahn ¨ et al. 2011a; Lozano et al. 2001). It was also observed that the effect of the anion on protein structure is more pronounced than the effect of the cation in most cases (Weingartner ¨ et al. 2012; Constantinescu et al. 2007; Kaar et al. 2003). Furthermore, the destabilizing effect of ionic liquids in the aqueous mixture is a function of the ion concentration (Senske et al. 2016) as shown in Fig. 5. Stabilizing effects (T > 0) were only observed for dihydrogenphosphate salts as well as sodium and potassium chloride above 2 M. Here, “enlightenment” from MD simulations on the mechanism is highly welcome. In the quest of rationalizing ionic liquid protein interac- tions, MD simulations provide a useful tool for investiga- tions at a molecular level and can complement experimental findings (Haberler and Steinhauser 2011). Although the first Fig. 4 The chromophore oxyquinolinium betaine can be directly MD simulations of proteins were reported in the late 1970s linked to the biomolecule of interest. Upon laser excitation, the local (McCammon et al. 1977), the first MD simulation of a electric field (red lines) is changed and the response of the solvent protein in an aqueous ionic liquid was published only in molecules in the direct vicinity of the trehalose can be studied (Sajadi et al. 2014). Copyright 2014, American Chemical Society 2008 (Micaelo ˆ and Soares 2008) revealing the time delay of Biophys Rev (2018) 10:825–840 831 computational analysis in Fig. 1 concerning interaction of of large timescales by introducing an enhanced sampling proteins with ionic liquids. In Micaelo ˆ and Soares (2008), method called “infrequent metadynamics,” an algorithm investigated the behavior of the serine protease cutinase in that allows estimation of unfolding times on a greatly accel- [C mim][PF ]and [C mim][NO ] at various water con- erated timescale (Tung and Pfaendtner 2016). For assessing 4 6 4 3 tents. For [C mim][PF ], 5–10% of water was necessary protein stability, several structural quantities are accessible 4 6 for optimal enzyme stabilization (Micaelo ˆ and Soares 2008) via MD simulations: indicating the importance of even small amounts of water. A subsequent publication by the same group further Radius of gyration investigated the influence of the protein surface characteris- tics on stability in ILs by conducting simulations of Candida As unfolding increases protein volume, the loss of secondary rugosa lipase and Bos taurus α-chrymotopsine with mod- structure leads to an increase of the radius of gyration r (t ) gyr ified surface charges (Burney et al. 2015). By replacing surface lysines with glutamate, artificial mutations with m (r (t ) − r (t )) i i CM positive-to-negative surface charge ratios were examined in r (t ) = (11) gyr simulation. On a 50-ns timescale, no stability difference for the different enzymes was found, but the authors explic- itly state that such a timescale is likely too short (Burney calculated with respect to the current center of mass r (t ) CM et al. 2015). Due to the slow dynamics brought about by the of the protein, where m is the mass of the ith atom and M high viscosity of ionic liquids, the need for large timescales the total mass of the protein. Klahn ¨ et al. (2011a) reported when conducting simulations in ionic liquids is one of the that r (t ) of Candida antarctica lipase B (CAL-B) in aque- gyr great obstacles, and unfolding times are often too large to ous IL mixtures (hydration level 20% w/w) increased by be accessible by simulation (Tung and Pfaendtner 2016). 0.8, 1.1, and 2.8 Afor [C mim][PF ], [C mim][BF ], and 4 6 4 4 The problem of insufficient sampling in pure ionic liquids [C mim][NO ], respectively. They suggested that the strength 4 3 is also mentioned by Burney and Pfaendtner in their study of interactions between anions and enzyme is mainly gov- on the behavior of Candida rugosa lipase in [C mim][NO ] erned by ion density and ion surface charge, favoring inter- 4 3 and [C mim][PF ]. Despite sampling times of more than actions with smaller ions which can form stronger hydrogen 4 6 100 ns, no visible effects of the ionic liquid on the pro- bonds with the protein. This finding goes along with the tein could be observed, despite experimental evidence that denaturation capabilities of classical ILs in Fig. 5e. the enzyme is heavily denatured in [C mim][NO ](Bur- However, Chaban and co-workers (Chevrot et al. 2015) 4 3 ney et al. 2015). Tung and Pfaendtner address this problem showed that the radius of gyration does not increase in ionic Fig. 5 Salt-induced shifts of the melting temperature T of RNase in aqueous ionic liquid mixtures with various ion concentrations (Senske et al. 2016)—Published by the PCCP Owner Societies. (a,b) chlorides (c)bromides(d) sodium salts (e)C mim based salts (f) stabilizing oxo-anions 832 Biophys Rev (2018) 10:825–840 liquids with amino acid-based anions during the simulation The stronger destabilizing effect of nitrate compared period of 10 ns which may not be long enough to monitor to hexafluorophosphate was confirmed by Klahn ¨ et al. the unfolding process due to the high viscosity of these (2011a) for CAL-B in various aqueous ionic liquid mixtures ionic liquids. Interestingly, Shao (2013) reported that r by means of RMSD-values and agreed well with the gyr of the B domain of protein A from Staphylococcus aureus concurrent observation of an increased radius of gyration actually decreased with increasing [C mim]Cl content of r (t ) mentioned in the last section. Although to less 4 gyr the aqueous solution. Shao performed three independent extent than the anions, longer alkyl chains lead to higher simulations of 200 ns length to get meaningful values. values of these quantities, indicating destabilization (Klahn ¨ A decrease of the protein radius was also observed by et al. 2011b). Replacing butyl with a methoxyethyl group in Bhattacharyya and co-workers (Ghosh et al. 2015)for C mim also resulted in substantial protein destabilization, lysozyme in [C mim]Br. However, they computed the possibly by the increased hydrogen-bonding ability of the radius of hydration from the Stokes-Einstein relation: methoxyethyl group. Diddens et al. (2017) analyzed the RMSD of a β -hairpin k T D = (12) peptide by metadynamics simulations in aqueous mix- 6πηr tures of [C mim][OAc], [C mim][BF ], and [C mim]Cl. 2 2 4 2 using experimental diffusion coefficients from fluorescence The β -hairpin peptide unfolded in aqueous mixture of correlation spectroscopy and the experimental viscosity η. [C mim][OAc] and [C mim][BF ] as visible in the insets 2 2 4 The radius of gyration and the radius of hydration are linked of Fig. 6. However, the global energetic minima of the two by a factor of 5/3(Schroder ¨ et al. 2015). unfolded configurations in Fig. 6b and c are completely different. The authors argued that the strong denaturating Root-mean-squared displacements acetate attacks the intramolecular hydrogen bonds of the peptide. As a result, the peptide structure unfolds com- The time-dependent root-mean-square deviation RMSD pletely. In aqueous [C mim][BF ], the global minimum is 2 4 (Kufareva and Abagyan 2012) still located near that of the pure water solution; however, the RMSD is quite large. Since the anions are excluded 2 from the surface of the peptide, this effect was attributed to ref RMSD = r (t ) − r (13) the impact of the cations. In contrast to the aqueous mixture of [C mim][OAc] is defined as a function of the deviation of the current atomic and [C mim][BF ], the β -hairpin peptide did not unfold in 2 4 ref coordinates r (t ) from the coordinates r of a reference aqueous [C mim]Cl. The minimum structure of the peptide i 2 structure with N atoms, usually at the beginning of the resembles the structure found in pure water (see Fig. 6a simulation or the native configuration. and d). This finding is in agreement with the experiment in In the first MD simulation of a protein in an ionic Fig. 5b, where [C mim]Cl was characterized to be only a liquid by Micaelo ˆ and Soares mentioned above, the weak denaturant. authors analyzed the RMSD of cutinase in [C mim][PF ] In MD simulations of Shao (2013), the RMSD values 4 6 and [C mim][NO ] at different hydration levels and of the B domain of protein A from Staphylococcus aureus 4 3 temperatures (Micaelo ˆ and Soares 2008). In accordance decreased with increasing content of [C mim]Cl which was with experimental findings reporting that many enzymes interpreted as a stabilizing effect that may also be an effect retain activity in ILs with PF anions and lose activity of the increased viscosity. in ILs with NO anions (Lozano et al. 2001; Kaar et al. 2003; Persson and Bornscheuer 2003; Lou et al. 2006), they Root-mean-squared fluctuations observed generally larger RMSD values in [C mim][NO ]. 4 3 In contrast, Burney et al. did not find a significant The root-mean-squared fluctuation change in the RMSD of Candida rugosa lipase 1 in [C mim][PF ]and[C mim][NO ] (Burney and Pfaendtner 1 4 6 4 3 RMSF (i) = (r (t ) −r  ) (14) i i t 2013). Pfaendtner and co-workers (Burney et al. 2015) demonstrated that the RMSD is not only a function of the ionic liquid but also depends on the nature of the protein. For measures the fluctuation of atom i around its mean position the enzyme Candida rugosa lipase 1, the stability decreases r  (Pitera 2014). Experimentally, atomic flexibility is i t in the following order of the solvent: water > [C mim]Cl described with the crystallographic B-factor (Bornot et al. > [C mim][EtSO ]. However, no RMSD trend is visible 2011). Amino acids with high RMSF values have an 2 4 for Bos taurus α-chymotrypsin in the very same solvents increased mobility which may also be an indicator for (Burney et al. 2015). interactions with the ionic liquid (Attri et al. 2015; Burney Biophys Rev (2018) 10:825–840 833 Fig. 6 Free energy landscape with global energetic minimum conformation of a β -hairpin. a Native state of the peptide in water. b Denaturated state in aqueous [C mim]. c Denaturated state in [C mim]. d Denaturated state in [C mim]Cl. R is the end-to-end distance of the peptide. (Diddens et al. 2017) - Published by the PCCP Owner Societies and Pfaendtner 2013; Jaeger and Pfaendtner 2013)which index j may also comprise all atoms (separated by at least breaks or at least disturbs intramolecular hydrogen bonding three bonds) in the very same molecule to obtain intramo- elec vdW of the biomolecule. lecular energies U (t ) and U (t ). Steinhauser and co- PP PP elec For the analysis of simulations of three different cellu- workers (Haberler and Steinhauser 2011) computed U PP vdW lases in aqueous mixtures of the ionic liquid [C mim][OAc], and U  for a zinc finger protein in aqueous solutions PP elec Jaeger et al. (2015)usedthe RMSD and an RMSF (aver- of [C mim][CF SO ], and found a rise in U (x ) with 2 3 3 IL PP vdW aged by residue) of the C atoms of the protein. They increasing ionic mole fraction x , whereas U (x ) α IL IL PP found that the three investigated cellulases (a cellulase from decreased with increasing mole fraction. Moreover, the Thrichderma viride, a cellulase 5A from Thermotoga mar- sum of electrostatic and van-der-Waals energies exhibited itima, and an endoglucanase from Pyrococcus horikoshii) a maximum at an IL mole fraction of x = 0.073. The IL exhibited remarkably different stabilities in the aqueous authors termed this mole fraction a “magic point” since ionic liquid [C mim][OAc], highlighting that specific fea- other secondary structure descriptors also showed extrema tures of the protein, such as surface characteristics, and not at this point. This behavior can be attributed to a change only the ionic liquid itself determine stability. from dipolar screening by water to charge screening by the ionic liquid ions when the IL mole fraction increases, which elec Electrostatic interactions is consistent with a rise of U (x ). IL PP Klahn ¨ et al. analyzed the electrostatic and van-der- Non-bonded energies in MD simulation are most generally Waals-interaction energy of hydrated Candida antarctica elec calculated by the electrostatic energy U (t ) lipase B (CAL-B) with a set of ionic liquids composed of imidazolium- and guadinium-based cations and nitrate, 1 q · q i j tetrafluoroborate or hexfluorophosphate anions (Klahn ¨ et al. elec U (t ) = (15) 4π r (t ) 2011a, b). Here, the second atomic index j of Eq. 15 0 ij i j contains all solvent atoms of the particular solvent species. They found that interaction energies are dominated by between the partial charges q and q of the atoms i and i j elec Coulombic interactions U  between the anions (A) PA j at a distance of r (t ) as well as van-der-Waals energies ij and the protein, while smaller cation (C) contributions vdW U (t ) vdW stem from van-der-Waals-interactions U . The charge PC on cations is more delocalized while the charge density 12 6 σ σ ij ij vdW on the smaller anions is higher, allowing them to form U (t ) = 4 − (16) ij r (t ) r (t ) ij ij stronger hydrogen bonds with the protein. In their study i j on xylanase II in [C mim][EtSO ]and [C mim][OAc], 2 4 2 based on the interatomic Lennard-Jones parameter σ and Jaeger and Pfaendtner also attributed the low affinity of ij . cations for negatively charged residues to the high charge ij Usually, the first atomic index i concerns all atoms in delocalization (Jaeger and Pfaendtner 2013). Klahn ¨ et al. vdW elec the biomolecule P under investigation. The second atomic reported the highest values of U +U  (strongest 834 Biophys Rev (2018) 10:825–840 hydrogen-bond interactions) for the NO anion, which is that these high solvation enthalpies were mainly due to consistent with experimental and simulation results that large cage formation energies. These large cage formation NO destabilizes CAL-B (Klahn ¨ et al. 2011a;Micaelo ˆ and energies are due to the strong interactions between Soares 2008). Weakest interactions were observed for the ionic liquid ions compared to the weaker interactions in PF anion, which stabilizes CAL-B. From experiment, it is water. An exception was the ionic liquid [C mim][PF ], 4 6 also known that [C mim][NO ] can dissolve solid CAL-B, in which the cage formation energy was even lower 4 3 while [C mim][PF ] cannot (Klahn ¨ et al. 2011a; Lau et al. than in water. Furthermore, upon transfer from water to 4 6 2004). IL, the internal energy of the protein increased, except in [C mim][PF ], corresponding to the experimentally 4 6 Solvation free energy observed stability increase in [C mim][PF ]. Another 4 6 approach to estimate the free energy landscape of a protein The general procedure of solvation free energy calculation is the use of metadynamics (Barducci et al. 2008), which is briefly described in the next section on small molecules. was already mentioned in the context of a small β - As already mentioned before, the simulations of proteins hairpin protein whose unfolding free energy landscape in in ionic liquids are hampered by the increased viscosity aqueous [C mim][OAc], [C mim][BF ], and [C mim][Cl] 2 2 4 2 in ionic liquids, requiring long simulation times for was calculated via metadynamics (Lesch et al. 2015; sufficient sampling (Tung and Pfaendtner 2016). Under Diddens et al. 2017). As mentioned earlier, the strong these circumstances, it is impossible to acquire sufficient denaturation of the peptide by [C mim][OAc] apparent from preferential binding parameters was also visible in the sampling for the calculation of solvation free energies (Klahn ¨ et al. 2011a). However, some other methods for free energy landscapes. obtaining thermodynamic information about proteins in ionic liquids have been employed. Klahn ¨ et al. (2011a) calculated the solvation free enthalpy H of CAL-B based Small rigid drugs solv on a linear response formalism (Klahn ¨ et al. 2011a), Ionic liquids are of special interest for the extraction of solute solvent compounds from biomass. They can dissolve a wide range H = U − U + U + PV , solv solv gas solute of biomatrices, even materials such as cellulose (Pinkert (17) et al. 2009;Wangetal. 2012; Zhu et al. 2006) or chitin (Qin et al. 2010), and can enhance solubility of hydrophobic solute where U is the potential of the system containing N solv compounds in aqueous solution by acting as hydrotropes solvent solute solvated solutes in N solvent molecules, and U gas (Claudio ´ et al. 2015) or surfactants. and U are the potential energies of the N solutes in solvent From the computational point of view, these molecules gas phase and the pure solvent, respectively. The pressure- are much more rigid compared to a large flexible protein. volume term is small and may be neglected, and when Consequently, observables concerning size and shape, solute N = 1 is assumed, Eq. 17 may be written as e.g., the radius of gyration as well as root-mean-squared solute solvent solute−solvent displacements and fluctuations, are of less use for the H = U + U + U solv solv solv small biomolecular drugs dissolved in the aqueous mixtures solute solvent − U + U (18) gas of ionic liquids. Hence, particular interactions like π -π - solute solute−solvent stacking or hydrogen bonding attract more attention. Also, = U + U + U . (19) cage free energy calculations, which are not feasible for complete solute Here, U is the potential energy of only the solute in proteins, may help to understand the biomolecular solvation solv solvent solvent, U is the potential energy of only the solvent of smaller solutes. However, the current knowledge on the solv solute−solvent in the system, and U is the solute-solvent interaction of small biomolecules with ILs is not as far solute solvent interaction energy. U = U −U is termed the progressed compared to protein solvation. cage solv cage formation energy and measures the energy required to solute solvent form a solute cavity in the solvent. U = U − Hydrotropic theory solv solute U is the change in internal energy of the solute upon gas insertion into solvent (Klahn ¨ et al. 2011a). It was found The hydrotropic effect describes the increased water that solvation enthalpies in ILs increased noticeably in solubility of hydrophobic organic compounds by the ionic liquids compared to water, which is in accordance addition of a co-solvent (Shimizu and Matubayasi 2014; with experimental observations that CAL-B is less soluble Eastoe et al. 2011;Claudio ´ et al. 2015; Zeindlhofer et al. in ionic liquids compared to water. It was also observed 2017, 2018) which usually consists of small, amphiphilic Biophys Rev (2018) 10:825–840 835 molecules (Booth et al. 2012). In principle, the following (Zeindlhofer et al. 2018) on the solvation of coffee ingredi- mechanisms are discussed: ents in [C mim][OAc] mixtures. The cations seem to repel water in general and also anions from the surface of the 1. The solute forms a stable complex with few hydrotropic solute at higher concentration. Even in the second solvation molecules via hydrogen bonding or π -π -stacking. The shell determined by Voronoi tessellation, the concentration complex has an increased polarity, hence increased of water is below bulk density which argues against mech- solubility in water. anism 1. Nevertheless, water still has access to the solute 2. The hydrotrope assembles in micelle-like structures in surface, even at higher concentrations, and micellar cationic which the solute can be accommodated. structures are not observed at the surface. Also, the last 3. The water-immiscible solute induces a particular struc- hydrotropic hypothesis 3 could not be proven. The inclusion ture of the hydrotrope in the aqueous solution. This new of acetate in the water network is usually not a problem and structure promotes the solubility of the solute. does not change the water structure in the proximity of the Imidazolium-based ILs seem to be promising candidates solute very much. However, a Voronoi shell-wise computa- for hydrotropy. Their charged, hydrophilic ring is accom- tion of the dielectric permittivities reveals a subtle transition panied by more or less hydrophobic alkyl side chains (Tan from the solute permittivity to the permittivity of the bulk et al. 2009;Claudio ´ et al. 2015). We recently reported solution (Zeindlhofer et al. 2018). Fig. 7 Distribution functions P(n ) of aggregate sizes in pure aqueous ionic liquid mixtures (a, b) and with vanillin (d, e). c depicts a vanillin-water mixture. Reproduced from Claudio ´ et al. (2015) with permission of The Royal Society of Chemistry 836 Biophys Rev (2018) 10:825–840 Furthermore, hydrogen bonding between the water and visible in Fig. 7c,whichisbrokenupintosmalleraggregates the solute is not changed by the cations (as suggested by in the aqueous IL solution by formation of larger vanillin- hypothesis 3) but in competition with the anions. Interest- cation clusters and smaller vanillin-anion clusters in Fig. 7d ingly, the solvation efficiency drops from tosylate to chlo- and e. − − ride for gallic acid (Claudio ´ et al. 2015): Cl <[MeSO ] − − < [CF SO ] < [N(CN) ] < [Tos ]. A similar ranking Solvation free energy 3 2 was found in vanillin. Although the cations are the most prominent species at the solute surface, particular anion Enthalpic and entropic effects of the dissolution of a solute interactions are very important for the solvation behav- in a solvent can be best characterized by the computation ior of the ionic liquid, i.e., hydrogen bonding. Cationic of solvation free energies. In contrast to large proteins, π -π -stacking is of minor importance. the amount of sampling needed for these calculations is feasible for small compounds. A common approach is Solute-solvent interactions and solute aggregation thermodynamic integration. During several simulations, a coupling parameter λ is stepwise increased from zero to The extraction of shikimic acid from star anise pods one corresponding to a switching off of solute-solvent solute−solvent using [C mim]-based ionic liquids with a set of different interactions U − − − − − − anions (PF ,NTf ,BF ,Cl ,OTf ,OAc )was 6 2 4 solute−solvent solvent U(λ) = (1 − λ) U + λ · U (20) reported in an experimental study (Zirbs et al. 2013). Accompanying MD simulations were performed to identify solvent U corresponds to the total energy of the solvent only. the relevant molecular interactions in the extraction process: The free energy difference between the initial and final No correlation of the experimental extraction yield with λ-state is given by the coordination number or the contact surface obtained from simulations could be established (Zirbs et al. 2013). ∂U (λ) G = dλ (21) solv Coordination numbers around shikimic acid are higher for ∂λ cations, but anions seem to interact stronger since they are To study molecular details of extraction of amino acids able to form hydrogen bonds, and their residence time in the first solvation shell around shikimic acid is longer than from aqueous solution using ILs (Absalan et al. 2010;Tome ´ et al. 2012), Seduraman et al. (2012) calculated the sol- for cations. The number of anion hydrogen bonds correlates with the extraction yield, highlighting the importance of the vation free energy of L-tryptophan in pure and aqueous [C mim][PF ], [C mim][PF ], and [C mim][BF ]. In con- anions in the extraction process. 4 6 8 6 8 4 In agreement with electrostatic analysis of the IL ion trast to experimental studies, they found only negligible with proteins (Klahn ¨ et al. 2011a;Klahn ¨ et al. 2011b), differences in the free energies in the pure ILs. However, elec U  of the anions (A) with the solute (S) is much including residual water in the IL simulations, which is SA elec larger compared to U  of the cations (C) (Zeindlhofer higher in [C mim][PF ], the experimental trend is repro- 4 6 SC duced illustrating the importance of water for the extraction et al. 2017;Tomee ´ tal. 2012). Furthermore, the van-der- vdW − Waals interaction U  between cations and the solute process. The solvation in BF -based ILs is more favorable, SC elec probably due to the fact that the water density is 7.5 times exceeds the corresponding Coulomb interaction U SC for C mim at the surface of various coffee ingredients higher compared to the hexafluorophosphate systems. Fur- thermore, electrostatic interactions of tryptophan with BF (Zeindlhofer et al. 2017). Cations are mainly found parallel to the aromatic solute rings in a stacked orientation, while are stronger than with PF . This anion dependency was anions preferred positions around the ring that allowed observed for other amino acids as well (Absalan et al. 2010). hydrogen bonding to ring substituents as visible from elec U . The same finding was observed for five different SA amino acids (glycine, alanine, valine, isoleucine, glutamic Conclusion acid) in aqueous solutions of [C mim][NTf ](Tomee ´ tal. 4 2 After the initial hype in the beginning of the twenty-first 2012). Another experimental study complemented by MD simu- century, expectations on ionic liquids decreased because of issues with the price, viscosity, the stability in water, lations concerned the extraction of vanillin in aqueous solu- tions of [C mim][N(CN) ]and[C mim][SCN] (Claudio ´ the toxicity, and biodegradability. Often, the unique selling 4 2 4 point of ionic liquids was missing in applications to et al. 2015). In MD simulations, the ionic liquid ions form several filamentous ionic strands and interact with water justify the higher cost compared to conventional solvents. in an anion-water hydrogen bond network. The hydropho- Moreover, exorbitant promises during the hype prompted bic vanillin molecules form a large droplet in pure water as several authors to question the current status of knowledge Biophys Rev (2018) 10:825–840 837 and the usefulness of IL in general (Kunz and Hackl ¨ Applying these suggestions to future simulations on 2016). Consequently, to prevent making mistakes again, biomolecular solvation in ionic liquids should help to reach which were made during the hype, good practice routines the “Plateau of Productivity” within reasonable time. should go without saying. In the current review, we briefly summarized meaningful computational methods to Acknowledgments Open access funding provided by Austrian Sci- investigate the solvation of large and small biomolecules in ence Fund (FWF). aqueous ionic liquid mixtures. The increasing number of publications concerning ionic Compliance with Ethical Standards liquids demonstrates their unbowed interest. However, in This article does not contain any studies with human participants or order to make real progress during the implementation of animals performed by any of the authors. ionic liquids and their computational analysis, several ideas should be taken into account: Conflict of interests Veronika Zeindlhofer declares that she has no conflict of interest. Christian Schroder ¨ declares that he has no conflict of interest. 1. The vast majority of computational works deal with aprotic imidazolium-based ionic liquids. Maybe, protic Open Access This article is distributed under the terms of the imidazoliums as well as other cations like choline Creative Commons Attribution 4.0 International License (http:// may be more biodegradable and cheaper. Therefore, creativecommons.org/licenses/by/4.0/), which permits unrestricted juxtaposition of computational results of these “new” use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a ionic liquids to the aprotic imidazolium-based ionic link to the Creative Commons license, and indicate if changes were liquid are interesting. made. 2. The same is true for the anions. Tetrafluorobo- rate and hexafluorophosphate can easily be coarse- grained and much experimental data exists, but these References anions are of less interest for the experimenter. Therefore, other anions such as dihydrogenphosphate Absalan G, Akhond M, Sheikhian L (2010) Partitioning of acidic, (dhp), methyl- and ethylsulfate, hydrophobic bis- basic and neutral amino acids into imidazolium-based ionic (trifluoromethylsulfonyl)-imide NTf , or hydrophilic liquids. Amino Acids 39(1):167–174 Anderson JL, Ding J, Welton T, Armstrong DW (2002) Characterizing dicyanamide N(CN) should also be included in the ionic liquids on the basis of multiple solvation interactions. J Am portfolio of the theoreticians. Even deep eutectic sol- Chem Soc 124(47):14247–14254 vents are interesting but currently still mostly neglected Anugwom I, Maki-Arv ¨ ela P., Virtanen P, Willfor ¨ S., Sjoholm ¨ in MD simulations on solvation of biomolecules. R., Mikkola JP (2012) Selective extraction of hemicelluloses 3. Many solvation effects depend on the concentration of from spruce using switchable ionic liquids. Carbohydr Polym 87(3):2005–2011 the co-solvent and should be given much more room Armand M, Endres F, MacFarlane DR, Ohno H, Scrosati B (2009) in current theoretical investigations as these simulations Ionic-liquid materials for the electrochemical challenges of the can be simply parallelized. future. Nat Mater 8(8):621–629 4. Particular effects of ionic liquids should be tested on Attri P, Sarinont T, Kim M, Amano T, Koga K, Cho AE, Ha Choi E, Shiratani M (2015) Influence of ionic liquid and ionic salt more than one protein as the observation made may be on protein against the reactive species generated using dielectric a peculiarity of the investigated protein. barrier discharge plasma. Sci Rep 5:17781 5. Polarizable force fields to include electronic effects Barducci A, Bussi G, Parrinello M (2008) Well-tempered metadynam- may also be an option. ics: a smoothly converging and tunable free-energy method. Phys Rev Lett 100(2):020603 6. The computational setup is of utmost importance for the Benedetto A, Ballone P (2016) Room temperature ionic liquids meet reliability of the MD results. Hence, simulation system biomolecules: a microscopic view of structure and dynamics. ACS sizes should be large enough to prevent artifacts from Sustain Chem Eng 4(2):392–412 self-interactions. For example, at least five complete Bhargava BL, Klein ML (2009) Aqueous solutions of imida- solvation shells around the biomolecule should be zolium ionic liquids: molecular dynamics studies. Soft Matter 5(18):3475–3480 considered. The simulation period should be at least Booth JJ, Abbott S, Shimizu S (2012) Mechanism of hydrophobic 50 ns in low viscous systems, i.e., mixtures with high drug solubilization by small molecule hydrotropes. J Phys Chem water content, and even much longer with increasing B 116(51):14915–14921 ionic liquid concentration. Of course, the simulation Bornot A, Etchebest C, De Brevern AG (2011) Predicting protein flexibility through the prediction of local structures. Proteins period has also to be prolonged when analyzing 79(3):839–852 folding/unfolding processes. Here, smaller proteins Brandt A, Grasvik ¨ J., Hallett JP, Welton T (2013) Deconstruction may be more useful than larger proteins to cover the of lignocellulosic biomass with ionic liquids. Green Chem complete process or at least major parts of it. 15(3):550–583 838 Biophys Rev (2018) 10:825–840 Burney P, Nordwald EM, Hickman K, Kaar JL, Pfaendtner J (2015) Gutowski KE, Broker GA, Willauer HD, Huddleston JG, Swatloski Molecular dynamics investigation of the ionic liquid/enzyme RP, Holbrey JD, Rogers RD (2003) Controlling the aqueous misci- interface: application to engineering enzyme surface charge. bility of ionic liquids: aqueous biphasic systems of water-miscible Proteins: Struct Funct Bioinf 83(4):670–680 ionic liquids and water-structuring salts for recycle, metathesis, Burney P, Pfaendtner J (2013) Structural and dynamic features of and separations. J Am Chem Soc 125(22):6632–6633 Candida rugosa lipase 1 in water, octane, toluene, and ionic liquids Haberler M, Steinhauser O (2011) On the influence of hydrated BMIM-PF6 and BMIM-NO3. J Phys Chem B 117(9):2662– ionic liquids on the dynamical structure of model proteins: a compu- 2670 tational study. Phys Chem Chem Phys 13(40):17994–18004 Haberler M, Schroder ¨ C, Steinhauser O (2011) Solvation studies of Chang TM, Dang LX, Devanathan R, Dupuis M (2010) Structure a zinc finger protein in hydrated ionic liquids. Phys Chem Chem and dynamics of N, N-Diethyl-N-methylammonium triflate ionic Phys 13(15):6955–6969 liquid, neat and with water, from molecular dynamics simulations. Hallett JP, Welton T (2011) Room-temperature ionic liquids: solvents J Phys Chem A 114(48):12764–12774 for synthesis and catalysis. 2. Chem Rev 111(5):3508–3576 Chen X, Liu J, Wang J (2010) Ionic liquids in the assay of proteins. Heid E, Schroder ¨ C (2016) Computational solvation dynamics Anal Methods 2(9):1222–1226 of oxyquinolinium betaine linked to trehalose. J Chem Phys Chevrot G, Fileti EE, Chaban VV (2015) Enhanced stability of the 145(16):164507 model mini-protein in amino acid ionic liquids and their aqueous Heid E, Schroder C (2017) Effect of a tertiary butyl group on solutions. J Comput Chem 36(27):2044–2051 polar solvation dynamics in aqueous solution: a computational Claudio ´ AFM, Freire MG, Freire CS, Silvestre AJ, Coutinho JA (2010) approach. J Phys Chem B 121(41):9639–9646 Extraction of vanillin using ionic-liquid-based aqueous two-phase Heid E, Schroder C (2018) Solvation dynamics in polar solvents and systems. Sep Purif Technol 75(1):39–47 imidazolium ionic liquids: failure of linear response approxima- Claudio ´ AFM, Neves MC, Shimizu K, Lopes JNC, Freire MG, tions. Phys Chem Chem Phys 20(7):5246–5255 Coutinho JAP (2015) The magic of aqueous solutions of ionic Holbrey JD, Seddon KR (1999) Ionic liquids. Clean Technol Environ liquids: ionic liquids as a powerful class of catanionic hydrotropes. Policy 1(4):223–236 Green Chem 17(7):3948–3963 Itoh T (2017) Ionic liquids as tool to improve enzymatic organic Constantinescu D, Weingartner ¨ H, Herrmann C (2007) Protein synthesis. Chem Rev 117(15):10567–10607 denaturation by ionic liquids and the hofmeister series: a case Jaeger VW, Pfaendtner J (2013) Structure, dynamics, and activity of study of aqueous solutions of ribonuclease–A. Angew Chem Int xylanase solvated in binary mixtures of ionic liquid and water. Ed 46(46):8887–8889 ACS Chem Biol 8(6):1179–1186 Cull S, Holbrey J, Vargas-Mora V, Seddon K, Lye G (2000) Room- Jaeger V, Burney P, Pfaendtner J (2015) Comparison of three temperature ionic liquids as replacements for organic solvents in ionic liquid-tolerant cellulases by molecular dynamics. Biophys J multiphase bioprocess operations. Biotechnol Bioeng 69(2):227– 108(4):880–892 Kaar JL, Jesionowski AM, Berberich JA, Moulton R, Russell AJ Diddens D, Lesch V, Heuer A, Smiatek J (2017) Aqueous ionic (2003) Impact of ionic liquid physical properties on lipase activity liquids and their influence on peptide conformations: denaturation and stability. J Am Chem Soc 125(14):4125–4131 and dehydration mechanisms. Phys Chem Chem Phys 19:20430– Kelkar MS, Maginn EJ (2007) Effect of temperature and water content on the shear viscosity of the ionic liquid 1-ethyl-3- Dreyer S, Kragl U (2008) Ionic liquids for aqueous two-phase methylimidazolium bis(trifluoromethanesulfonyl)imide as studied extraction and stabilization of enzymes. Biotechnol Bioeng by atomistic simulations. J Phys Chem B 111(18):4867–4876 99(6):1416–1424 Kilpelainen ¨ I, Xie H, King A, Granstrom M, Heikkinen S, Earle MJ, Seddon KR (2009) Ionic liquids. Green solvents for the Argyropoulos DS (2007) Dissolution of wood in ionic liquids. J future. Pure Appl Chem 72(7):1391–1398 Agric Food Chem 55(22):9142–9148 Eastoe J, Hatzopoulos MH, Dowding PJ (2011) Action of hydrotropes King AWT, Asikkala J, Mutikainen I, Jarvi ¨ P, Kilpelainen ¨ I and alkyl-hydrotropes. Soft Matter 7(13):5917–5925 (2011) Fort DA, Remsing RC, Swatloski RP, Moyna P, Moyna G, Rogers RD Distillable acid–base conjugate ionic liquids for cellulose (2007) Can ionic liquids dissolve wood? Processing and analysis dissolution and processing. Angew Chem Int Ed 50(28):6301– of lignocellulosic materials with 1-n-butyl-3-methylimidazolium 6305 chloride. Green Chem 9(1):63–69 Kirkwood JG, Buff FP (1951) The statistical mechanical theory of Freemantle M (1998) Designer solvents. Chem Eng News 76(13):32– solutions. I. J Chem Phys 19(6):774–777 37 Klahn ¨ M, Lim GS, Wu P (2011a) How ion properties determine the Freire MG, Claudio ´ AFM, Araujo ´ JMM, Coutinho JAP, Marrucho stability of a lipase enzyme in ionic liquids: a molecular dynamics IM, Lopes JNC, Rebelo LPN (2012) Aqueous biphasic systems: study. Phys Chem Chem Phys 13(41):18647–18660 a boost brought about by using ionic liquids. Chem Soc Rev Klahn ¨ M, Lim GS, Seduraman A, Wu P (2011b) On the different roles 41(14):4966–4995 of anions and cations in the solvation of enzymes in ionic liquids. Fujita K, MacFarlane DR, Forsyth M, Yoshizawa-Fujita M, Murata Phys Chem Chem Phys 13(4):1649–1662 K, Nakamura N, Ohno H (2007) Solubility and stability of Kragl U, Eckstein M, Kaftzik N (2002) Enzyme catalysis in ionic cytochrome c in hydrated ionic liquids: effect of oxo acid residues liquids. Curr Opin Biotechnol 13(6):565–571 and kosmotropicity. Biomacromolecules 8(7):2080–2086 Kufareva I, Abagyan R (2012) Methods of protein structure Gao WW, Zhang FX, Zhang GX, Zhou CH (2015) Key factors comparison. Methods Mol Biol 857:231–257 affecting the activity and stability of enzymes in ionic liquids and Kunz W, Hackl ¨ K (2016) The hype with ionic liquids as solvents. novel applications in biocatalysis. Biochem Eng J 99:67–84 Chem Phys Lett 661:6–12 Ghosh S, Parui S, Jana B, Bhattacharyya K (2015) Ionic liquid Lai JQ, Li Z, Lu ¨ YH, Yang Z (2011) Specific ion effects of ionic liquids induced dehydration and domain closure in lysozyme: FCS and on enzyme activity and stability. Green Chem 13(7):1860–1868 MD simulation. J Chem Phys 143(12):125103 Lau RM, Sorgedrager MJ, Carrea G, Rantwijk FV, Secundo F, Gupta KM, Hu Z, Jiang J (2013) Molecular insight into cellulose Sheldon RA (2004) Dissolution of Candida antarctica lipase B regeneration from a cellulose/ionic liquid mixture: effects of water in ionic liquids: effects on structure and activity. Green Chem concentration and temperature. RSC Adv 3(13):4425–4433 6(9):483–487 Biophys Rev (2018) 10:825–840 839 Lee SY, Khoiroh I, Ooi CW, Ling TC, Show PL (2017) Recent Persson M, Bornscheuer UT (2003) Increased stability of an esterase advances in protein extraction using ionic liquid-based aqueous from Bacillus stearothermophilus in ionic liquids as compared to two-phase systems. Sep Purif Rev 46(4):291–304 organic solvents. J Mol Catal B: Enzym 22(1-2):21–27 Lesch V, Heuer A, Tatsis VA, Holm C, Smiatek J (2015) Peptides in Pierce V, Kang M, Aburi M, Weerasinghe S, Smith PE (2008) Recent the presence of aqueous ionic liquids: tunable co-solutes as denaturants applications of Kirkwood-Buff theory to biological systems. Cell or protectants? Phys Chem Chem Phys 17(39):26049–26053 Biochem Biophys 50:1–22 Liang M, Zhang X, Kaintz A, Ernsting NP, Maroncelli M (2014) Pinkert A, Marsh KN, Pang S, Staiger MP (2009) Ionic liquids and Solvation dynamics in a prototypical ionic liquid + dipolar aprotic their interaction with cellulose. Chem Rev 109(12):6712–6728 liquid mixture: 1-butyl-3-methylimidazolium tetrafluoroborate + Pitera JW (2014) Expected distributions of root-mean-square positio- acetonitrile. J Phys Chem B 118(5):1340–1352 nal deviations in proteins. J Phys Chem B 118(24):6526–6530 Lou WY, Zong MH, Smith TJ, Wu H, Wang JF (2006) Impact of Plechkova NV, Seddon KR (2008) Applications of ionic liquids in the ionic liquids on papain: an investigation of structure–function chemical industry. Chem Soc Rev 37(1):123–150 relationships. Green Chem 8(6):509–512 Qin Y, Lu X, Sun N, Rogers RD (2010) Dissolution or extraction Lozano P, de Diego T, Guegan JP, Vaultier M, Iborra JL of crustacean shells using ionic liquids to obtain high molecular (2001) Stabilization of α-chymotrypsin by ionic liquids in weight purified chitin and direct production of chitin films and transesterification reactions. Biotechnol Bioeng 75(5):563–569 fibers. Green Chem 12(6):968 Maki-Arv ¨ ela P, Anugwom I, Virtanen P, Sjoholm ¨ R, Mikkola JP Raut DG, Sundman O, Su W, Virtanen P, Sugano Y, Kordas K, (2010) Dissolution of lignocellulosic materials and its constituents Mikkola JP (2015) A morpholinium ionic liquid for cellulose using ionic liquids—a review. Ind Crops Prod 32(3):175–201 dissolution. Carbohydr Polym 130:18–25 Mandal PK, Samanta A (2005) Fluorescence studies in a pyrrolidinium Roy D, Maroncelli M (2012) Simulations of solvation and solvation ionic liquid: polarity of the medium and solvation dynamics. J dynamics in an idealized ionic liquid model. J Phys Chem B Phys Chem B 109(31):15172–15177 116(20):5951–5970 McCammon JA, Gelin BR, Karplus M (1977) Dynamics of folded Sajadi M, Berndt F, Richter C, Gerecke M, Mahrwald R, Ernsting NP proteins. Nature 267(5612):585–590 (2014) Observing the hydration layer of trehalose with a linked Micaelo ˆ NM, Soares CM (2008) Protein structure and dynamics molecular terahertz probe. J Phys Chem Lett 5(11):1845–1849 in ionic liquids. Insights from molecular dynamics simulation Samanta A (2010) Solvation dynamics in ionic liquids: what we have studies. J Phys Chem B 112(9):2566–2572 learned from the dynamic fluorescence stokes shift studies. J Phys Mikkola JP, Kirilin A, Tuuf JC, Pranovich A, Holmbom B, Kustov Chem Lett 1(10):1557–1562 LM, Murzin DY, Salmi T (2007) Ultrasound enhancement of Sasaki H (2015) Simulating hype cycle curves with mathematical cellulose processing in ionic liquids: from dissolution towards functions : some examples of high-tech trends in Japan. Int J Inf functionalization. Green Chem 9(11):1229–1237 Technol 7(2):1–12 Moniruzzaman M, Nakashima K, Kamiya N, Goto M (2010) Recent Schmollngruber M, Schroder ¨ C, Steinhauser O (2013) Polarization advances of enzymatic reactions in ionic liquids. Biochem Eng J effects on the solvation dynamics of coumarin C153 in ionic 48(3):295–314 liquids: components and their cross-correlations. J Chem Phys Naushad M, ALOthman ZA, Khan AB, Ali M (2012) Effect of ionic 138(20):204504 liquid on activity, stability, and structure of enzymes: a review. Int Schofer ¨ SH, Kaftzik N, Wasserscheid P, Kragl U (2001) Enzyme J Biol Macromol 51(4):555–560 catalysis in ionic liquids: lipase catalysed kinetic resolution of 1- Neumayr G, Rudas T, Steinhauser O (2010) Global and local phenylethanol with improved enantioselectivity. Chem Commun Voronoi analysis of solvation shells of proteins. J Chem Phys 0(5):425–426 133(8):084108 Schroder ¨ C (2017) Proteins in ionic liquids: current status of Omta AW, Kropman MF, Woutersen S, Bakker HJ (2003) Negligible experiments and simulations. Top Curr Chem 375(2) effect of ions on the hydrogen-bond structure in liquid water. Schroder ¨ C, Rudas T, Neumayr G, Benkner S, Steinhauser O (2007) Science 301(5631):347–349 On the collective network of ionic liquid/water mixtures. I. Park S, Kazlauskas RJ (2003) Biocatalysis in ionic liquids— Orientational structure. J Chem Phys 127(23):234503 advantages beyond green technology. Curr Opin Biotechnol Schroder ¨ C, Hunger J, Stoppa A, Buchner R, Steinhauser O (2008) 14(4):432–437 On the collective network of ionic liquid/water mixtures. II. Parviainen A, King AWT, Mutikainen I, Hummel M, Selg C, Hauru Decomposition and interpretation of dielectric spectra. J Chem LKJ, Sixta H, Kilpelainen ¨ I (2013) Predicting cellulose solvating Phys 129(18):184501 capabilities of acid—base conjugate ionic liquids. ChemSusChem Schroder ¨ C, Neumayr G, Steinhauser O (2009) On the collective 6(11):2161–2169 network of ionic liquid/water mixtures. III. Structural analysis of Parvulescu ˆ VI, Hardacre C (2007) Catalysis in ionic liquids. Chem ionic liquids on the basis of Voronoi decomposition. J. Chem. Phys Rev 107(6):2615–2665 130(19):194503 Patel R, Kumari M, Khan AB (2014) Recent advances in the Schroder ¨ C, Sega M, Schmollngruber M, Gailberger E, Braun applications of ionic liquids in protein stability and activity: a D, Steinhauser O (2014) On the collective network of ionic review. Appl Biochem Biotechnol 172(8):3701–3720 liquid/water mixtures. IV. Kinetic and rotational depolarization. J. Pei Y, Wang J, Wu K, Xuan X, Lu X (2009) Ionic liquid-based aqueous Chem. Phys 140(20):204505 two-phase extraction of selected proteins. Sep Purif Technol Schroder ¨ C, Steinhauser O, Sasisanker P, Weingartner ¨ H (2015) Ori- 64(3):288–295 entational alignment of amyloidogenic proteins in pre-aggregated Pereira JFB, Lima AS, Freire MG, Coutinho JAP (2010) Ionic liquids solutions. Phys Rev Lett 114(12). https://doi.org/10.1007/s41061- as adjuvants for the tailored extraction of biomolecules in aqueous 017-0110-2 biphasic systems. Green Chem 12(9):1661–1669 Seduraman A, Wu P, Klahn ¨ M (2012) Extraction of tryptophan with Pereira JFB, Ventura SPM, e Silva FA, Shahriari S, Freire MG, ionic liquids studied with molecular dynamics simulations. J Phys Coutinho JAP (2013) Aqueous biphasic systems composed of Chem B 116(1):296–304 ionic liquids and polymers: a platform for the purification of Senske M, Constantinescu-Aruxandei D, Havenith M, Herrmann biomolecules. Sep Purif Technol 113:83–89 C, Weingartner ¨ H, Ebbinghaus S (2016) The temperature 840 Biophys Rev (2018) 10:825–840 dependence of the Hofmeister series: thermodynamic fingerprints liquids: a review of molecular dynamics simulations. J Mol Liq of cosolute—protein interactions. Phys Chem Chem Phys 210:178–188 18(43):29698–29708 Ventura SPM, e Silva FA, Quental MV, Mondal D, Freire Shao Q (2013) On the influence of hydrated imidazolium-based MG, Coutinho JAP (2017) Ionic-liquid-mediated extraction and ionic liquid on protein structure stability: a molecular dynamics separation processes for bioactive compounds: past, present, and simulation study. J Chem Phys 139(11):115102 future trends. Chem Rev 117(10):6984–7052 Shimizu S, Matubayasi N (2014) Hydrotropy: Monomer–Micelle Ventura SPM, Neves CMSS, Freire MG, Marrucho IM, Oliveira equilibrium and minimum hydrotrope concentration. J Phys Chem J, Coutinho JAP (2009) Evaluation of anion influence on the B 118(35):10515–10524 formation and extraction capacity of ionic-liquid-based aqueous Smiatek J (2017) Aqueous ionic liquids and their effects on protein biphasic systems. J Phys Chem B 113(27):9304–9310 structures: an overview on recent theoretical and experimental Wang H, Gurau G, Rogers RD (2012) Ionic liquid processing of results. J Phys: Condens Matter 29(23):233001 cellulose. Chem Soc Rev 41(4):1519–1537 Smith PE (2004) Cosolvent interactions with biomolecules: relating Wasserscheid P (2017) Talk at the 655th Heraeus symposium: Surfaces computer simulation data to experimental thermodynamic data. J and Interfaces of Ionic Liquids Phys Chem B 108:18716–18724 Weingartner ¨ H, Cabrele C, Herrmann C (2012) How ionic liquids Steudte S, Neumann J, Bottin-Weber U, Diedenhofen M, Arning can help to stabilize native proteins. Phys Chem Chem Phys J, Stepnowski P, Stolte S (2012) Hydrolysis study of fluo- 14(2):415–426 roorganic and cyano-based ionic liquid anions—consequences Welton T (1999) Room-temperature ionic liquids. Solvents for for operational safety and environmental stability. Green Chem synthesis and catalysis. Chem Rev 99(8):2071–2084 14(9):2474–2483 Yang Z, Pan W (2005) Ionic liquids: green solvents for nonaqueous Summers CA, Flowers RA (2000) Protein renaturation by the liquid biocatalysis. Enzyme Microb Technol 37(1):19–28 organic salt ethylammonium nitrate. Protein Sci 9(10):2001–2008 Zeindlhofer V, Khlan D, Bica K, Schroder ¨ C (2017) Computational Swatloski RP, Spear SK, Holbrey JD, Rogers RD (2002) Dissolution analysis of the solvation of coffee ingredients in aqueous ionic of cellose with ionic liquids. J Am Chem Soc 124(18):4974–4975 liquid mixtures. RSC Adv 7(6):3495–3504 Tan SSY, MacFarlane DR, Upfal J, Edye LA, Doherty WOS, Patti AF, Zeindlhofer V, Berger M, Steinhauser O, Schroder ¨ C (2018) A shell- Pringle JM, Scott JL (2009) Extraction of lignin from lignocel- resolved analysis of preferential solvation of coffee ingredients in lulose at atmospheric pressure using alkylbenzenesulfonate ionic aqueous mixtures of the ionic liquid 1-ethyl-3-methylimidazolium liquid. Green Chem 11(3):339–345 acetate. J Chem Phys 148(19):193819 Terranova ZL, Corcelli SA (2013) On the mechanism of solvation Zhang Y, Cremer P (2006) Interactions between macromolecules and dynamics in imidazolium-based ionic liquids. J Phys Chem B ions: the Hofmeister series. Curr Opin Chem Biol 10(6):658–663 117(49):15659–15666 Zhang Y, Cremer PS (2009) The inverse and direct Hofmeister Thuy Pham TP, Cho CW, Yun YS (2010) Environmental fate and series for lysozyme. Proc Natl Acad Sci USA 106(36):15249– toxicity of ionic liquids: a review. Water Res 44(2):352–372 15253 Tome ´ LIN, Jorge M, Gomes JRB, Coutinho JAP (2012) Molecular Zhang J, Wu J, Yu J, Zhang X, He J, Zhang J (2017) Application dynamics simulation studies of the interactions between ionic of ionic liquids for dissolving cellulose and fabricating cellulose- liquids and amino acids in aqueous solution. J Phys Chem B based materials: state of the art and future trends. Mater Chem 116(6):1831–1842 Front 1(7):1273–1290 Tung HJ, Pfaendtner J (2016) Kinetics and mechanism of ionic-liquid Zhao H (2010) Methods for stabilizing and activating enzymes in ionic induced protein unfolding: application to the model protein HP35. liquids—a review. J Chem Technol Biotechnol 85(7):891–907 Mol Syst Des Eng 1(4):382–390 Zhao H (2016) Protein stabilization and enzyme activation in ionic Ulbert O, Belafi-Bak ´ o ´ K, Tonova K, Gubicza L (2005) Thermal liquids: specific ion effects: protein stabilization and enzyme acti- stability enhancement of Candida rugosa lipase using ionic liquids. vation in ionic liquids. J Chem Technol Biotechnol 91(1):25–50 Biocatal Biotransform 23(3–4):177–183 Zhu S, Wu Y, Chen Q, Yu Z, Wang C, Jin S, Ding Y, Wu G (2006) van Rantwijk F, Sheldon RA (2007) Biocatalysis in ionic liquids. Dissolution of cellulose with ionic liquids and its application: a Chem Rev 107(6):2757–2785 mini-review. Green Chem 8(4):325–327 Varela LM, Mendez-Morales ´ T, Carrete J, Gomez-Gonz ´ alez ´ V, Zirbs R, Strassl K, Gaertner P, Schroder ¨ C, Bica K (2013) Exploring Docampo-Alvarez B, Gallego LJ, Cabeza O, Russina O (2015) ionic liquid-biomass interactions: towards the improved isolation Solvation of molecular cosolvents and inorganic salts in ionic of shikimic acid from star anise pods. RSC Adv 3(48):26010

Journal

Biophysical ReviewsSpringer Journals

Published: Apr 23, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off