Article pubs.acs.org/Biomac
Lysozyme Solubility and Conformation in Neat Ionic Liquids and Their Mixtures with Water Stephen Strassburg, Harry Bermudez, and David Hoagland* Polymer Science and Engineering Department, University of MassachusettsAmherst, Amherst, Massachusetts 01003, United States S Supporting Information *
ABSTRACT: The room temperature solubility of a number of model proteins is assessed for a diverse set of neat ionic liquids (ILs). For two soluble protein−IL pairs, lysozyme in [C2MIM][EtSO4] (1-ethyl-3-methylimidazolium ethylsulfate) and in [C2,4,4,4P][Et2PO4] (tributyl(ethyl)phosphonium diethylphosphate), protein solubility and structure at various temperatures are probed by dynamic light scattering (assessing dissolved molecular size), turbidimetry (reflecting degree of solubility), and Fourier transform infrared spectroscopy (uncovering helical secondary structure). As compared to aqueous environments, [C2,4,4,4P][Et2PO4] thermally stabilizes protein size and secondary structure while [C2MIM][EtSO4] does the opposite. Lysozyme denatured in [C2MIM][EtSO4] does not aggregate, presumably due to an absence of hydrophobic interactions, and the denaturation appears thermally reversible. Both ILs at room temperature are miscible with water in all proportions, but to create the corresponding ternary mixtures with protein, the order of mixing is important. Mixed to avoid additions of water to IL-dissolved protein, stable solutions are obtained with [C2MIM][EtSO4] at all solvent compositions. When water is added to IL-rich solutions, liquid−liquid demixing is noted.
■
INTRODUCTION Ionic liquids (ILs) are attracting attention as vehicles for the long-term storage and process stabilization of dissolved proteins.1 Comprised of two components, cation and anion, neat ILs provide tunable solvents for a variety of proteins and other biopolymers, and when mixed with water or aqueous buffers, their tunability grows. According to much literature, the denaturing temperature of proteins dissolved in IL−water mixtures can be above the aqueous denaturing temperature,2−7 and several IL−water mixtures show promise as media for enzyme catalysis, encapsulation, recovery, and refolding.8−11 In this contribution, protein solubility and stability are studied in neat ILs and IL-rich water mixtures, environments little scrutinized in previous literature, and even then, only for a few IL-protein pairs. Despite significant past study of ILs as protein solvents and stabilizers, choosing an appropriate IL to achieve a desired protein outcome remains challenging; an IL suitable for one protein may be highly unsuitable for another, and the range of IL choices is vast. Attri et al., for example, found that lengthening the alkyl substituents on imidazolium and phosphonium IL cations weakened the IL-stabilization of chymotrypsin,12 and Yan et al. noted the same trend for bovine serum albumin in the 1-alky-3-methylimidazolium IL series [CnMIM][Br] (n = 4, 6, 8, 10).13 However, the latter also reported that reducing alkyl substituent length does not promote stabilization of all proteins. Many investigations of protein solubility/stability in IL−water mixtures demonstrated that the anion exerts a larger (and more complex) role than the cation.9,15,16 To identify and isolate impacts from cation and © XXXX American Chemical Society
anion on proteins dissolved in water−IL mixtures, Constantinescu et al. developed correlations between ion positions in the Hofmeister series and RNaseA stability,3 and Akdogan et al. and Fujita et al. developed similar Hofmeister connections for human serum albumin and cytochrome c, respectively.15,16 However, these correlations are far from perfect, and the Hofmeister series would seem of little relevance to the highly concentrated or neat ILs of focus for the current contribution. The concentration of IL in a protein−IL−water mixture strongly affects protein solubility/stability. Byrne and Angell, for example, demonstrated a protic IL’s ability at relatively high concentration to fibrilize lysozyme by destabilizing its helical structure.17 Differently, the alpha helical structure of Aβ(1−40) peptide transforms into amyloid fibrils at low concentration of triethylammonium methylsulfate IL, [Tea][MeSO4], but its helical structure is stabilized at high concentration of the same IL.18 Understanding how concentrated or neat ILs preserve or disrupt a protein’s structure is crucial to the development of ILs as protein or polypeptide storage media. Little attention has been directed at the dissolution of proteins in neat ILs (those with no detectable water) or IL-rich enviornments.19,20 Bihari et al. suggested that neat [C2MIM][EtSO4] (1-ethyl-3methylimidazolium ethylsulfate) disrupts the tertiary structure of dissolved cytochrome c but leaves its secondary structure mostly intact.20 Further, this study determined that, despite Received: March 31, 2016 Revised: May 4, 2016
A
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules Table 1. Room Temperature Solubility Matrix for Dilute Proteins in ILsa
bovine serum albumin human serum albumin ovalbumin myoglobin α-chymotrypsin lysozyme cytochrome c lactoferrin
[C6,6,6,14P] [Cl]
[C2,4,4,4P] [Et2PO4]
[C1,8,8,8N] [Cl]
[C1,4Pyr] [DCA]
[C2MIM] [EtSO4]
[C4MIM] [OAc]
[C4MIM] [SCN]
[C4MIM] [PF6]
[C4MIM] [BF4]
×
○
○
×
×
○
○
×
×
×
×
×
×
×
○
○
×
×
○
○ ○ ○ ×b ○ ×
×
×
○ ○ ○ ○ ○ ○
○ ○ ○ ○ ○ ○
○ ○ ○ ○ ○ ○
× × × × × ×
× × × × × ×
× × × ×
× × × ×
○ ×
All proteins were tested at 10 mg/mL and 20 °C. A ○ indicates solubility, while × indicates insolubility. The blank boxes reflect unattempted experiments. bLysozyme does not dissolve at room temperature but does dissolve above 35 °C.
a
h for [C2MIM][EtSO4] and 72 h for [C2,4,4,4P][Et2PO4]. All sample preparation was designed to minimize contact with atmosphere; Protein-IL samples were capped and sealed after filtration or placed in a vacuum oven at 60 °C until needed for characterization, as reported by Bihari et al.20 After dissolution, solutions were cooled to room temperature before subsequent characterizations. Aqueous protein reference solutions were prepared in 0.05 M phosphate buffer at pH ∼ 7.2 (for DLS) and 0.05 M NaCl in D2O at pD ∼ 7.1 (for FTIR). Methods. DLS. DLS measurements of hydrodynamic radius Rh were performed using an ALV (model: SP-125) or Brookhaven (model: BI-200SM) scattering instrument operated at an angle between 30° and 120° and a temperature T between 20 and 90 °C. A thermocouple immersed in the scattering vat allowed T to be controlled within ±0.5 °C, and across the T range, the IL viscosities and refractive indices needed for data interpretation were obtained from literature.21,22 Error bars attached to Rh correspond to 95% confidence intervals of the student t-distribution as calculated from three replicate measurements. All Rh distributions were narrow as calculated by CONTIN and second-order cumulants analyses, and the reported Rh is that determined via the CONTIN analysis of the correlation function. FTIR. A PerkinElmer Spectrum ATR FTIR spectrometer provided room temperature spectra for lysozyme in D2O salt solution and ILs; 128 scans at 2 cm−1 resolution were averaged. For spectra collected above ambient T, samples were preheated to the desired T and then dropped directly onto the scanner, where T was monitored by a thermocouple. Because T drops due to heat loss, only 16 scans at 2 cm−1 resolution could be taken at higher T. UV−Vis Spectroscopy. Expressed as percent transmittance for 500 nm light, the turbidity of protein solutions placed in 0.2 cm path length quartz cells was evaluated using a Hitachi U-3010 spectrophotometer. To prevent settling, samples were agitated just prior to measurement.
denaturation, cytochrome c molecularly dissolves in this IL, displaying no tendency to aggregate or precipitate.20 Here, after addressing the solubility of numerous proteins in neat examples picked from common IL classes, one protein, lysozyme, is selected for further investigations of protein thermal behavior in two neat ILs, 1-ethyl-3-methylimidzolium ethylsulfate, [C 2 MIM][EtSO 4 ] and tributyl(ethyl)phosphonium diethylphosphate, [C2,4,4,4P][Et2PO4]. In both, lysozyme’s dissolved size and secondary structure are determined as functions of temperature by dynamic light scattering (DLS) and Fourier transform infrared spectroscopy (FTIR), respectively. Unfortunately, performing circular dichroism in ILs to determine protein secondary structure is difficult due to the intrinsic UV absorbance of most ILs. By turbidimetry, the solubility of lysozyme is examined for mixtures of the two ILs with water, spanning the range from pure IL to pure water while exploring the reversibility of phase behavior at each intervening composition. Lysozyme was chosen for the study simply as a well-characterized model protein that conveniently dissolves in several common neat ILs, while [C2MIM][EtSO4] was chosen to extend the considerable existing literature for proteins dissolved in this IL and its mixtures with water. By providing comparison and contrast to results from [C2MIM][EtSO4], [C2,4,4,4P][Et2PO4] adds insight into the universality of protein−IL solution behaviors.
■
EXPERIMENTAL SECTION
Materials. Lysozyme from hen egg white was purchased from Sigma-Aldrich and used without further purification, and [C2MIM][EtSO4] and [C2,4,4,4P][Et2PO4] were respectively purchased from Sigma-Aldrich and donated by Cytec of Canada. Additional proteins as well as the ILs, [C1,8,8,8P][Cl] (trioctylmethylammonium chloride), [C 6 , 6 , 6 , 1 4 P][Cl] (trihexyltetradecylphosphonium chloride), [C6,6,6,14P][Br] (trihexyltetradecylphosphonium bromide), [C1,4Pyr][DCA] (1-butyl-1-methylpyrrolidinium dicyanamide), [C4MIM][OAc] (1-butyl-3-methylimidazolium acetate), [C4MIM][SCN] (1butyl-3-methylimidazolium thiocyanate), [C4MIM][BF4] (1-butyl-3methylimidazolium tetrafluoroborate), and [C4MIM][PF6] (1-butyl-3methylimidazolium hexafluorophosphate), employed in solubility tests were purchased from Sigma-Aldrich. Each IL was first heated above 60 °C in vacuum for 3 days to remove excess water, which was assayed to four decimal place accuracy by refractometry. Refractive indices after vacuum drying matched values reported in literature at 25 °C, 1.479 for [C2MIM][EtSO4],21 and 1.467 for [C2,4,4,4P][Et2PO4],22 establishing that the level of residual water was below the detection limit, ∼0.3 wt % (see the Supporting Information for more discussion of refractometry). Lysozyme was dissolved at 10 mg/mL in neat ILs heated to 60 °C; dissolution, as determined by visual inspection, was achieved within 24
■
RESULTS AND DISCUSSION Protein-IL Solubility Matrix. As opposed to the vast literature on proteins in IL−water mixtures, the literature on proteins in neat ILs, or in aqueous mixtures with an IL as the majority component, remains scant.17,18,20 Nearly all of the previous investigations of proteins in IL−water mixtures did not continue to IL concentrations beyond those causing the test protein to separate from an aqueous mixture, which viewing the IL as a salt, might be characterized as “salting out”. Thus, the presence of a window of solubility at higher IL concentration and extending to neat IL was missed. This window is a key focus of the current study. To discern factors that influence protein dissolution in neat ILs, the protein−IL solubility matrix shown as Table 1 was prepared using a span of well-studied model proteins (e.g., myoglobin, lysozyme, and albumins) mixed with various B
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules
temperature, and the two are thus suitable for the mixed solvent investigations. Lysozyme Dissolution and Solubility in Two ILs. Dissolving lysozyme in neat [C2,4,4,4P][Et2PO4] and [C2MIM][EtSO4] at room temperature proved problematic, and these protein−IL mixtures were consequently heated to above 60 °C for up to 72 h to prepare homogeneous, transparent solutions; no ill effects of this heating for lysozyme or ILs were noted. After return to room temperature and storage at this temperature for about a week, [C2,4,4,4P][Et2PO4] solutions separated. [C2MIM][EtSO4] solutions, on the other hand, remained stable at room temperature for at least several months. A major concern for heating was the possible hydrolysis of [EtSO4]− and [Et2PO4]−. According to Jacquemin et al., hydrolysis of [C2MIM][EtSO4] in mixtures with 1.8−18 vol % water could not be detected by several analytical methods for samples held at 80 °C for 48 h.28 Even in a 1:1 (mol:mol) mixture held at 150 °C for 72 h, hydrolysis went undetected. Likewise, using data from Wolfenden et al., the first order rate constant for the hydrolysis of [Et2PO4]− at room temperature is 1.6 × 10−13 s−1, equivalent to a half-life for the hydrolysis reaction of 105 years.29 Given that our water content in neat ILs from refractometry was below 0.3 vol %, all water contents, temperatures, and times of our study were much less extreme than necessary for detectable anion hydrolysis. At higher vol % of water, hydrolysis will be more rapid, and therefore, heating of protein−IL−water mixtures was completely avoided. Based on Jacquemin et al. and Wolfenden et al., even at high water content, room temperature hydrolysis will be negligible. Lysozyme Size by DLS. As presented in Figure 1, for T near and above 35 °C, Rh of lysozyme in [C2,4,4,4P][Et2PO4]
representative ILs so as to sample the broadest range of behaviors. The cations include ammonium, phosphonium, pyrrolidinium, and imidazolium, and the anions include ethylsulfate, diethylphosphate, dicyanamide, thiocyanate, acetate, tetrafluoroborate, hexafluorophosphate, and chloride. These ions cover a wide range of hydrophobicity/hydrophilicity, Lewis acidity/basicity, ion size, dielectric constant, and so on.23−25 Together, the cation and anion determine the balance of forces that is ultimately responsible for protein solubility, to be discussed below. To achieve dissolution, all of these protein−IL mixtures were heated at 60 °C for up to 3 days. When the samples looked homogeneous, they were immediately cooled to room temperature for storage; except for lysozyme in [C2,4,4,4P][Et2PO4], homogeneity was either attained in less than a day or not at all, the latter indicating insolubility. Table 1 indicates homogeneity, i.e., solubility, with a ○ and insolubility with an ×. The exceptional case for initial dissolution, lysozyme in [C2,4,4,4P][Et2PO4], also displayed subsequent slow room temperature precipitation. Reheating to 35 °C caused redissolution. The table establishes two important results, that (i) many proteins dissolve in appropriate neat ILs and (ii) different neat IL classes can support protein dissolution. Both conclusions are apparently new and somewhat unexpected. Also, the table suggests that neat hydrophilic ILs−those miscible with water−seem better disposed to protein dissolution than those that are hydrophobic. For lysozyme in three ILs, [C2,4,4,4P][Et2PO4], [C2MIM][EtSO4], and [C4MIM][SCN], the maximum solubility (i.e., saturation concentration) was explored by visual inspection to higher concentrations than the 10 mg/mL standard established for Table 1, with results for three cases reported in Table S1. Above 15 mg/mL, lysozyme-[C2,4,4,4P][Et2PO4] mixtures at 35 °C became turbid. (All concentrations at 20 °C were turbid). However, for lysozyme in [C2MIM][EtSO4] and [C4MIM][SCN] at 20 °C, visual homogeneity was maintained up to 60 mg/mL (the highest concentration tested.) Each lysozyme solution was also examined by DLS to determine Rh, with results given in Table S1. For [BMIM][SCN], Rh ∼ 3.2 ± 0.8 nm, consistent with the protein’s denatured size in buffer.26 DLS results for [C2MIM][EtSO4] and [C2,4,4,4P][EtPO4] are discussed in next section. Lysozyme’s high solubility and denatured size in [C4MIM][SCN] is anticipated, as the thiocyanate anion is strongly denaturing27 in aqueous buffers; therefore, proteins dissolved in neat [C4MIM][SCN] likely lose their native conformations, favoring dissolution. The protein−solvent interactions necessary for dissolution are essentially the same as those responsible for denaturation, so dissolution coincident with structure retention requires interactions of narrowly defined strength. Fortunately, thousands of ILs are known, spanning a range of protein−solvent interaction strengths, and as importantly, many ILs are mutually miscible, allowing easy preparation of ILs of intermediate solvent strength. In this first investigation, guided by Table 1 and to avoid denaturation, [C2,4,4,4P][Et2PO4] and [C2MIM][EtSO4] were chosen as prototype ILs for subsequent conformation analysis. Although not solubilizing all proteins, [C2,4,4,4P][Et2PO4] and [C2MIM][EtSO4] have a higher DLS scattering contrast (i.e., large refractive index increment) with proteins than [C 2MIM][OAc], which produced no correlation function in DLS, and significantly, presented no peaks in the amide I region of FTIR. Both of the chosen prototype ILs are fully miscible with water at room
Figure 1. Hydrodynamic radius Rh vs temperature T for lysozyme in [C2,4,4,4P][Et2PO4], [C2MIM][EtSO4], and pH 7 aqueous phosphate buffer.
tracks closely with Rh of lysozyme in aqueous buffer, and in both cases, Rh ≈ 2.0 ± 0.4 nm, in accord with a literature value of Rh for the protein’s native state.26 Below 35 °C, where solubility in the aqueous buffer persists with Rh unchanged, lysozyme in this IL separates and settles. And at higher T, near and above 70 °C, the protein’s thermal denaturation in the aqueous buffer induces aggregation and a resulting abrupt upturn in Rh.26 Trends in this higher T range for [C2,4,4,4P][Et2PO4] are much different, with solubility maintained absent aggregation even as T rises well above the aqueous denaturation temperature of 69−72 °C.30 C
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules
helical structure. In D2O salt solution at 80 °C, where the protein is known to adopt a random coil conformation,34 the peak also lies at 1644 cm−1. Trends in IL and buffer resemble effects noted at high IL concentration in IL−water mixtures as lysozyme loses its helical structure and forms fibrils.17 Rather surprisingly, Figure 2B indicates that at all T up to 80 °C the amide I peak in [C2,4,4,4P][Et2PO4] remains at 1650 cm−1, suggesting that the IL not only supports but also stabilizes alpha helical structure. Both FTIR and DLS data point to Tindependent retention of secondary structure in neat [C2,4,4,4P][Et2PO4] and T-dependent loss of secondary structure in neat [C2MIM][EtSO4]. Of course, the amide I peak position is only an overall indicator of protein secondary structure, and denaturation in tertiary structure might occur independently, as seen previously for cytochrome c in [C2MIM][EtSO4].20 Factors Affecting Solubility, Conformation, and Aggregation. For a protein to be soluble, a sufficiently favorable mixing interaction must be maintained with the solvent. Here, both IL anions are Lewis bases that act as hydrogen-bond acceptors (HBA) for the protein amide groups.35,36 The relative strength of this interaction is revealed by the anion pKa, with ethylsulfate being a better HBA than diethylphosphate (see the Supporting Information). Second, IL cations such as [C2MIM]+ can act as hydrogen-bond donors (HBD), thereby interacting with the protein backbone carbonyl groups. The [C2,4,4,4P]+ cation, however, has no such HBD ability. These interactions are consistent with the larger (thermal) solubility window for [C2MIM][EtSO4] as compared to [C2,4,4,4P][Et2PO4]. Overall, the solubility results of Table 1 are rationalized by consideration of anion basicity, cation acidity, and hydrophobicity. As a general rule, proteins do not dissolve in hydrophobic ILs such as [C4MIM][BF4] and [C6,6,6,14P][Cl]. On the other hand, if the total interaction strength between IL ions and protein is too high, as typical with thiocyanate,27 the protein not only dissolves but also unfolds. Lysozyme in [C2MIM][EtSO4] shows thermally induced unfolding (Figures 1 and 2) similar to its behavior in buffer, but importantly and differently, the protein does not aggregate afterward. The unfolding can be explained in terms of the enthalpy loss associated with an increased number of protein− solvent contacts (e.g., favorable acidity/basicity), the conformational entropy gain as the protein disorders, and possibly, the entropy gain due to changes in protein solvation. For lysozyme in [C2,4,4,4P][Et2PO4], the protein mostly, or fully, retains its native size and secondary structure as T increases (Figure 1 and Figure 2). While [C2,4,4,4P]+ cations have weak, favorable interactions with newly exposed hydrophobic residues, we speculate that this bulky cation displaces or otherwise obstructs the interaction of [Et2PO4]− anions with protein amides. The net result is the exchange of a strong favorable interaction (anion-amide) with a weak favorable interaction (cationhydrophobic R group), shifting the equilibrium toward the folded state. Solubility of Lysozyme in IL−Water Mixtures. While the preceding experiments elucidated aspects of lysozyme’s behavior in neat ILs, elucidating these aspects in IL−water (or buffer) mixtures is just as important, as in any IL−protein technology, the protein must eventually be transferred to buffer, a step exposing the protein to mixed solvent at stage(s) during the transfer. Many studies investigated the impacts of ILs on proteins in dominantly water (or buffer) environments, but in the majority, experiments were only conducted at IL concentrations below the onset of phase separation (or
The same data show constant Rh for lysozyme in [C2MIM][EtSO4] up to 55 °C, a small increase in protein size with subsequent heating, and finally, a higher and constant Rh above 65 °C. Lysozyme thus appears to heat denature in [C2MIM][EtSO4], but this event is not accompanied by aggregation; also, unlike aqueous buffer, a clear T onset for denaturation is not evident. Even at 20 °C, Rh is slightly higher than in buffer or [C2,4,4,4P][Et2PO4], suggesting a slight denaturation has already occurred. The samples in IL required heating for protein dissolution, a possible explanation for the initial denaturation. In the high T plateau, Rh ≈ 3.2 ± 0.2 nm, consistent with this study’s finding that Rh ≈ 2.9 ± 0.4 nm when lysozyme is denatured in a pH 12 buffer and a literature report that Rh ≈ 3.4 nm for the denatured protein.26 Interestingly, with cooling to room temperature of an IL solution heated above 65 °C, Rh returns to its initial value. Such a recovery may indicate a refolding to the native state, or equally, a misfolding to a compact state of the same overall size. Notwithstanding trends of Rh with T, key outcomes of the DLS experiments are that (i) lysozyme can molecularly dissolve in the two ILs (at least within appropriate ranges of T) and (ii) the T-dependent solubilities of lysozyme in the two ILs are different. These ILs are disparate, varying in ionic strength, hydrophobicity, cation identity, and anion identity, and so their interactions with lysozyme are clearly different. Trends in solubility and structure are obviously not universal (and probably complex), a conclusion anticipated by the complicated pattern of protein solubilities in Table 1. Secondary Structure of Lysozyme in ILs by FTIR. Although no aggregation in the two prototype ILs was observed even above the aqueous denaturing temperature, notable differences in Rh were seen, and for [C2,4,4,4P][Et2PO4], Rh across the full T range was consistent with the protein in its native conformation. Lysozyme consists of ∼45% α helix, ∼ 19% β sheet, and ∼36% random coil.31 To look for IL-induced changes in secondary structure, the amide I peak position in the ILs can be compared to the literature peak position at 1650 cm−1.32 At 35 °C, the lowest T at which lysozyme dissolves in [C2,4,4,4P][Et2PO4], Figure 2A shows the amide I peak at 1650 cm−1, and at 20 °C, the lowest T tested with [C2,MIM][EtSO4], the same figure displays the amide I peak located at 1648 cm−1. While subtle, the shift to lower wavenumber in the latter IL is consistent with a difference (loss) of helical structure.33,34 At 80 °C, the amide I peak in [C2MIM][EtSO4] shifts even more, to 1644 cm−1, indicating an even larger loss of
Figure 2. FTIR of lysozyme in (A) D2O at 20 °C, [C2,4,4,4P][Et2PO4] at 35 °C, and [C2MIM][EtSO4] at 20 °C; and (B) D2O, [C2,4,4,4P][Et2PO4], and [C2MIM][EtSO4], all at 80 °C. D
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules precipitation), usually noted at an IL volume fraction in the range of 10−30%.3,4,13−15,37 Knowledge of solubility across the full composition space, neat buffer to neat IL, has not been addressed, even in the limit of low protein concentration, perhaps due to an expectation that the protein will remain insoluble at all compositions above the onset of insolubility. Phase behavior of proteins in IL−water (buffer) mixtures is potentially a rich topic, and recognizing the significant number of unknowns, this report addresses lysozyme’s solubility at only one protein concentration, 10 mg/mL, as a precursor to future examinations of the full T-dependent IL−water−protein ternary phase diagram. Initial mixed solvent experiments with [C2MIM][EtSO4] invoked an “incremental” or “stepped” dialysis procedure, devised to evade the precipitation inevitably noted when IL− protein solutions were dialyzed directly against buffer. In this procedure, lysozyme dissolved in the neat IL was initially dialyzed across a 4−8 kDa molecular-weight-cutoff membrane against a bath of 90 vol % IL and 10 vol % water. After 24 h, the bath was adjusted to 70 vol % IL and 30 vol % water, and the plan was to repeat this rise of water content in subsequent steps. However, lysozyme separated at ∼30 vol % water, terminating the protocol. The protein was made insoluble by adding water to salt rather than the reverse, the usual course of “salting out”. As a consequence, three alternative and simpler methods for changing solvent composition were pursued to clarify the observed phase separation. In the first, lysozyme was directly added to homogeneous IL−water mixtures (“premixed”). In the second, lysozyme was dissolved initially in water, and IL was subsequently added to fix the final mixed solvent composition (“IL-added”). And in the third, lysozyme was dissolved initially in neat IL and water was subsequently added to fix the final mixed solvent composition (“water-added”). In all three approaches, rather than impose incremental additions, a final composition was obtained by adding the requisite volume of the second solvent at once, offering a discrete composition jump rather than the continuous composition adjustment of stepped dialysis. The latter is made impractical by the slow rate of protein dissolution. After 24 h equilibration under stirring, protein solubility in the final mixtures was ascertained by turbidimetry. As demonstrated in Figure 3 (top and middle panels), no lysozyme insolubility is observed for [C2MIM][EtSO4] in premixed or IL-added mixtures. However, in water-added mixtures (bottom panel) insolubility is noted at all compositions below 4.7 M IL (stated differently, above 10 vol % water). The latter explains the precipitation observed during regular and stepped water dialysis of IL-dissolved protein. The insolubility of protein in IL-rich [C2MIM][EtSO4] mixtures as water is added can be rationalized by DLS and FTIR data that exhibit slight denaturation in the neat IL; this denaturation will expose hydrophobic regions that induce aggregation as water is introduced. The denaturation/ aggregation of lysozyme in water-added ILs appears irreversible, as mixtures remain turbid for days or more. Intriguingly, a literature investigation by DSC of lysozyme in water-dominated [C2MIM][EtSO4]-buffer mixtures suggested that the IL thermally stabilizes lysozyme, and the same investigation showed by activity assays that adding 1 M IL refolds in water the thermally denatured protein.9 Figure 4 shows that solubility trends for lysozyme in water[C2,4,4,4P][Et2PO4] mixtures are different than those just
Figure 3. Dissolution of lysozyme in water−[C2MIM][EtSO4] mixtures at 20 °C as determined by turbidimetry. Top panel, premixed; middle panel, IL-added; and bottom panel, water-added. Arrows indicate the direction of dilution.
Figure 4. Dissolution of lysozyme in water-[C2,2,2,4P][Et2PO4] mixtures at 20 °C as determined by turbidimetry. Top panel, premixed; middle panel, IL-added; and bottom panel, water-added. Arrows indicate the direction of dilution.
described for water-[C2MIM][EtSO4] mixtures, most notably with a narrow range of insolubility surfacing irrespective of mixing order at ∼0.7−0.8 M IL (∼70−75 vol % water). The close appearance of turbidity curves for premixed and IL-added mixtures suggests that both approaches lead to close-to equilibrium states across the full composition range. If so, the folding/unfolding and aggregation/disaggregation transitions suggested at intermediate compositions are reversible. The insoluble composition range, 0.7−0.8 M, falls approximately E
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules where lysozyme might be expected to “salt out” from the solvent mixture. The samples of Figures 3 and 4 were examined 24 h after protein dissolution (premixed) or final solvent addition (ILadded and water-added). In this time, turbidity in all cases was mostly, but not completely, stabilized. In particular, the composition ranges for demixing were well-identified even if the absolute values of turbidity were still changing slowly at all compositions. No attempt was made to quantify the rates of mixing/demixing, which certainly vary with this composition, and therefore the small fluctuations in turbidity for the IL- and water-added samples are not considered significant. In addition to the usual factors affecting mixing kinetics, the impact of high IL viscosity (9821 and 30622 cP for [C2MIM][EtSO4] and [C2,4,4,4P][Et2PO4], respectively at 25 °C) must not be discounted. The most important aspect of Figures 3 and 4 is the large composition span over which lysozyme dissolves at high IL content. The IL effectively “solvates” the protein, a phenomenon unexpectedly dominating the greater part of the IL−water composition space. The structure of the protein across this span remains unknown, but one anticipates structures comparable to that in the neat IL, with the secondary structure somewhat destabilized by the increased capacity of water to displace intraprotein hydrogen bonds. In all protein systems of high IL concentration, electrostatic interactions are so highly screened by dissociated IL cations and anions that their impact on protein structure is unlikely. Whether ILs and inorganic salts behave differently in regard to protein interactions is unclear, although in the limit as neat liquids, ILs are clearly distinct from ordinary inorganic salts in supporting protein dissolution. The term “salting out” is typically attached to the precipitation of a protein from an aqueous solution when the salt concentration exceeds a critical value; in this definition, precipitation is the phase separation of a homogeneous liquid into a solid phase (perhaps crystallized) and a less concentrated liquid phase. When ILs are mixed with water or buffers, dissolved proteins might be expected to salt out when the IL composition is high, the IL simply acting as a dissociated salt. However, in the systems described here, such a process has never been observed. Instead, optical microscopy (data not shown), shows the actual process is liquid−liquid phase separation, with the minority phase, dispersed as fine droplets providing an increased turbidity that is accentuated by a large difference in refractive index between water (∼1.333) and IL (∼1.479). Liquid−liquid phase separation is even observed when just 1 wt % of protein is added to a homogeneous, (30:70 vol:vol) mixture of IL and water (see top panel, Figure 4). This character suggests that, rather than reference to salting out, a more proper context are aqueous biphasic systems (ABS).38−40
to adopt its full native conformation in either a neat IL or an aqueous mixture concentrated in IL. This study discovered a large solubility window for aqueous mixtures at IL concentrations beyond the onset of phase separation, but essentially nothing is yet known about protein conformation in this concentrated regime. Information needed to develop tools and approaches for transferring proteins from highly concentrated and neat IL environments into aqueous buffers is needed, although first steps in this direction were garnered here by simple turbidity measurements. In particular, direct dialysis into aqueous environments at room temperature disappointingly seems unequal to the transfer task. Such practical issues connect to fundamental questions: what is the nature of the interactions responsible for solubility and how can IL−protein interactions be controlled to achieve retention of protein structure and easy transfer to buffer? Given the immense number of protein and IL candidates, whether such questions can be answered in a universal way remains uncertain. Hopefully, while demonstrating the scale of open issues, this study established a few guideposts for more comprehensive future studies.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biomac.6b00468. Refractometry for determining water content, Table of pKa values, and Table of maximum solubility and Rh for lysozyme in ILs (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS This work was financially supported by the NSF Materials Research Science and Engineering Center (DMR-0820506). REFERENCES
(1) Weingärtner, H.; Cabrele, C.; Herrmann, C. Phys. Chem. Chem. Phys. 2012, 14 (2), 415. (2) Kumar, A.; Venkatesu, P. RSC Adv. 2013, 3 (2), 362. (3) Constantinescu, D.; Weingärtner, H.; Herrmann, C. Angew. Chem., Int. Ed. 2007, 46 (46), 8887−8889. (4) Attri, P.; Venkatesu, P. J. Chem. Thermodyn. 2012, 52, 78−88. (5) Attri, P.; Jha, I.; Choi, E. H.; Venkatesu, P. Int. J. Biol. Macromol. 2014, 69, 114−123. (6) Weaver, K. D.; Vrikkis, R. M.; Van Vorst, M. P.; Trullinger, J.; Vijayaraghavan, R.; Foureau, D. M.; McKillop, I. H.; MacFarlane, D. R.; Krueger, J. K.; Elliott, G. D. Phys. Chem. Chem. Phys. 2012, 14 (2), 790. (7) Wei, W.; Danielson, N. D. Biomacromolecules 2011, 12, 290−297. (8) Summers, C. A.; Flowers, R. A. Protein Sci. 2000, 9, 2001−2008. (9) Buchfink, R.; Tischer, A.; Patil, G.; Rudolph, R.; Lange, C. J. Biotechnol. 2010, 150 (1), 64−72. (10) Morikawa, M. A.; Takano, A.; Tao, S.; Kimizuka, N. Biomacromolecules 2012, 13, 4075−4080. (11) Lu, X.; Hu, J.; Yao, X.; Wang, Z.; Li, J. Biomacromolecules 2006, 7, 975−980. (12) Attri, P.; Venkatesu, P.; Kumar, A. Phys. Chem. Chem. Phys. 2011, 13 (7), 2788−2796.
■
CONCLUSIONS Neat ILs, as well as their concentrated mixtures with water or aqueous buffer, provide a new class of protein solvents, one with considerable implications for technology. Here, it was demonstrated that many proteins and ILs can be partnered successfully, resulting in solutions of diverse behaviors. Unexpectedly, true molecular solubility, even in neat ILs, is found to be a fairly common attribute of folded enzymatic proteins, which may retain much secondary structure. However, while solubility and secondary structure in neat ILs is established for some partners, no protein has yet been proven F
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules (13) Yan, H.; Wu, J.; Dai, G.; Zhong, A.; Chen, H.; Yang, J.; Han, D. J. Lumin. 2012, 132 (3), 622−628. (14) Constatinescu, D.; Herrman, C.; Weingartner, H. Phys. Chem. Chem. Phys. 2010, 12 (8), 1648. (15) Fujita, K.; MacFarlane, D. R.; Forsyth, M.; Yoshizawa-Fujita, M.; Murata, K.; Nakamura, N.; Ohno, H. Biomacromolecules 2007, 8, 2080−2086. (16) Akdogan, Y.; Junk, M. J. N.; Hinderberger, D. Biomacromolecules 2011, 12 (4), 1072−1079. (17) Byrne, N.; Angell, C. A. J. Mol. Biol. 2008, 378 (3), 707−714. (18) Debeljuh, N.; Barrow, C. J.; Henderson, L.; Byrne, N. Chem. Commun. (Cambridge, U. K.) 2011, 47 (22), 6371−6373. (19) Takekiyo, T.; Yamazaki, K.; Yamaguchi, E.; Abe, H.; Yoshimura, Y. J. Phys. Chem. B 2012, 116 (36), 11092−11097. (20) Bihari, M.; Russell, T. P.; Hoagland, D. a. Biomacromolecules 2010, 11 (11), 2944−2948. (21) Gómez, E.; González, B.; Calvar, N.; Tojo, E.; Domínguez, Á . J. Chem. Eng. Data 2006, 51 (6), 2096−2102. (22) Deive, F. J.; Rivas, M. a.; Rodríguez, A. J. Chem. Thermodyn. 2013, 62, 98−103. (23) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39 (21), 3772−3789. (24) Greaves, T. L.; Drummond, C. J. Chem. Rev. 2008, 108 (1), 206−237. (25) Ueno, K.; Tokuda, H.; Watanabe, M. Phys. Chem. Chem. Phys. 2010, 12 (8), 1649−1658. (26) Wilkins, D. K.; Grimshaw, S. B.; Dobson, C. M.; Jones, J. A.; Smith, L. J. Biochemistry 1999, 38, 16424−16431. (27) Baldwin, R. L. Biophys. J. 1996, 71 (4), 2056−2063. (28) Jacquemin, J.; Goodrich, P.; Jiang, W.; Rooney, D. W.; Hardacre, C. J. Phys. Chem. B 2013, 117 (6), 1938−1949. (29) Wolfenden, R.; Ridgway, C.; Young, G. J. Am. Chem. Soc. 1998, 120 (97), 833−834. (30) Venkataramani, S.; Truntzer, J.; Coleman, D. R. J. Pharm. BioAllied Sci. 2013, 5 (2), 148−153. (31) Levitt, M.; Greer, J. J. Mol. Biol. 1977, 114, 181−293. (32) Byler, M.; Susi, H. Biopolymers 1986, 25 (3), 469−487. (33) Nevskaya, A. Biopolymers 1976, 15 (4), 637−648. (34) Jackson, M.; Mantsch, H. H. Crit. Rev. Biochem. Mol. Biol. 1995, 30 (2), 95−120. (35) Nayak, P. K.; Hathorne, A. P.; Bermudez, H. Phys. Chem. Chem. Phys. 2013, 15 (6), 1806−1809. (36) Ueki, T.; Watanabe, M. Bull. Chem. Soc. Jpn. 2012, 85 (1), 33− 50. (37) Heller, W. T.; O’Neill, H. M.; Zhang, Q.; Baker, G. a. J. Phys. Chem. B 2010, 114 (43), 13866−13871. (38) Abraham, M. H.; Zissimos, A. M.; Huddleston, J. G.; Willauer, H. D.; Rogers, R. D.; Acree, W. E., Jr. Ind. Eng. Chem. Res. 2003, 42 (3), 413−418. (39) Freire, M. G.; Cláudio, A. F. M.; Araújo, J. M. M.; Coutinho, J. a. P.; Marrucho, I. M.; Lopes, J. N. C.; Rebelo, L. P. N. Chem. Soc. Rev. 2012, 41 (14), 4966. (40) Pereira, M. M.; Pedro, S. N.; Quental, M. V.; Lima, Á . S.; Coutinho, J. a. P.; Freire, M. G. J. Biotechnol. 2015, 206, 17−25.
G
DOI: 10.1021/acs.biomac.6b00468 Biomacromolecules XXXX, XXX, XXX−XXX
