Impact of Urea on Detergent Micelle Properties - Langmuir (ACS

Article pubs.acs.org/Langmuir

Impact of Urea on Detergent Micelle Properties Jana Broecker and Sandro Keller* Molecular Biophysics, University of Kaiserslautern, Erwin-Schrödinger-Straße 13, 67663 Kaiserslautern, Germany S Supporting Information *

ABSTRACT: Co-solvents, such as urea, can entail drastic changes in the micellization behavior of detergents. We present a systematic quantification of the impact of urea on the critical micellar concentration, the micellization thermodynamics, and the micelle size in three homologous series of commonly used non-ionic alkyl detergents. To this end, we performed demicellization experiments by isothermal titration calorimetry and hydrodynamic size measurements by dynamic light scattering on alkyl maltopyranosides, cyclohexyl alkyl maltopyranosides, and alkyl glucopyranosides at urea concentrations of 0−8 M. For all detergents studied, we found that the critical micellar concentration increases exponentially because the absolute Gibbs free energy of micellization decreases linearly over the entire urea concentration range, as does the micelle size. In contrast, the enthalpic and entropic contributions to micellization reveal more complex, nonlinear dependences on urea concentration. Both free energy and size changes are more pronounced for long-chain detergents, which bury more apolar surface area upon micelle formation. The Gibbs free energy increments per methylene group within each detergent series depend on urea concentration in a linear fashion, although they result from the entropic term for alkyl maltosides but are of enthalpic origin for cyclohexyl alkyl maltosides. We compare our results to transfer free energies of amino acid side chains, relate them to protein-folding data, and discuss how urea-induced changes in detergent micelle properties affect in vitro investigations on membrane proteins.

■

would resist cubic-phase crystallization.10 Moreover, the combined use of detergents and urea is common in many other applications and commercial products, such as liquid soaps, shampoos, hair conditioners, hair dyes, dye removers, and whitening toothpastes. The amphiphilicity of most detergents arises from the presence of a polar headgroup with an apolar chain.11 In aqueous environments, detergents exist primarily as monomers at low concentrations but self-associate to form micelles at higher concentrations, thereby minimizing the unfavorable exposure of their apolar moieties to water. The detergent concentration at which micellization sets in is called the critical micellar concentration (CMC). As a rule of thumb, the CMC decreases about 10-fold upon extension of the alkyl chain by two methylene groups within a homologous series of detergents carrying the same headgroup.2 Another important characteristic of detergent micelles is the mean aggregation number, which gives the average number of detergent monomers in a micelle. Typically, detergent micelles contain five to several hundred monomers and have sizes in the nanometer range.2 Among the many factors affecting these detergent micelle properties, urea is known to increase the CMC of the non-ionic detergents Triton X-100,12 n-octyl-β-D-maltopyranoside (OM),13 n-decyl-

INTRODUCTION Detergents are indispensable tools in membrane-protein research because their micellar assemblies mimic, to some extent, the anisotropic environment of lipid bilayers. In particular, detergents are crucial not only to extract a target membrane protein from its host membrane and keep it soluble during purification but also to maintain the purified protein in its functional, folded state during biophysical, biochemical, or structural-biological studies.1,2 For various applications in membrane-protein research, detergents are used in combination with urea as a co-solvent. (i) Refolding: A number of membrane proteins are recombinantly produced as inclusion bodies, which can be easily extracted from the host cell.3 After solubilization in urea or urea/sodium dodecyl sulfate mixtures, unfolded membrane proteins may then be refolded in the presence of a mild detergent by dilution.4 (ii) Stability measurements: The thermodynamics and kinetics of membrane insertion and concomitant folding of β-barrel outer membrane proteins can be quantified through their reversible unfolding out of detergent micelles by titration with urea.5−7 (iii) Reconstitution: The transfer of purified membrane proteins from detergent micelles into lipid bilayers may be facilitated by urea8 because the latter weakens the hydrophobic effect and promotes positive intrinsic membrane curvature.9 (iv) Crystallization: Urea swells lipidic cubic phases and increases the diameter of their aqueous channels. This supports the incorporation of bulky membrane proteins that otherwise © 2013 American Chemical Society

Received: April 11, 2013 Revised: June 7, 2013 Published: June 8, 2013 8502

dx.doi.org/10.1021/la4013747 | Langmuir 2013, 29, 8502−8510

Langmuir

Article

β-D-maltopyranoside (DM),14 n-dodecyl-β-D-maltopyranoside (DDM),14 n-octyl-β-D-glucopyranoside (OG),9,14 n-dodecylhexaethylene glycol monoether (C12E6),15 and n-dodecyloctaethylene glycol monoether (C12E8).9 Micelles composed of C12E8 have, for instance, also been shown to decrease in size from ∼10 nm in buffer to ∼4.5 nm in the presence of 2 M urea.9 To the best of our knowledge, however, the influence of urea on the CMC and micelle size of detergents routinely used in membrane-protein research has not been studied systematically thus far. We have employed isothermal titration calorimetry (ITC) and dynamic light scattering (DLS) to characterize the impact of urea on the micelle properties in three homologous series of non-ionic alkyl detergents, namely, n-alkyl-β-D-maltopyranosides, cyclohexyl-1-alkyl-β-D-maltopyranosides, and n-alkyl-β-Dglucopyranosides, henceforth referred to as (alkyl) maltosides, cyclohexyl alkyl maltosides, and (alkyl) glucosides, respectively (Figure 1). All three detergent series contain sugar-based

6, respectively; Figure 1B). Alkyl glucosides consist of a glucose headgroup and a linear alkyl chain. Among the members of this series, only n-octyl- and n-nonyl-β-D-glucoside (OG and NG, respectively; Figure 1C) are reasonably soluble in aqueous buffer.2

■

MATERIALS AND METHODS

Materials. All chemicals were purchased in the highest available purity. Ultrapure urea and cyclohexyl alkyl maltosides were purchased from Affymetrix (High Wycombe, U.K.). Alkyl maltosides and alkyl glucosides were from Glycon Biochemicals (Luckenwalde, Germany). HCl and tris(hydroxymethyl)aminomethane (Tris) were obtained from Carl Roth (Karlsruhe, Germany), and NaCl (AnalaR Normapure) was from VWR International (Darmstadt, Germany). Tris buffer (50 mM Tris and 50 mM NaCl at pH 7.4) was used in all experiments. Urea Solutions. A ∼10 M urea stock solution was prepared in triple-distilled water by weight and treated with Amberlite IRN-150L deionizing resin (GE Healthcare, Uppsala, Sweden) at room temperature for ∼1 h. After sterile filtration, the urea concentration was determined on an Abbemat 500 refractometer (Anton Paar, Ostfildern-Scharnhausen, Germany) using a third-order polynomial approximation.17 Stock solutions were either stored at −80 °C or prepared freshly and were supplemented with buffer components and adjusted to pH 7.4 prior to an experiment. Detergent Solutions. Detergent powders were weighed on a high-precision XP205 Delta Range microbalance (Mettler Toledo, Greifensee, Switzerland). To prepare samples for DLS measurements, buffer containing the desired concentration of urea was centrifuged at 20800g at 25 °C for 30 min to remove dust particles. The weighed detergent powder was dissolved in 1 mL buffer containing the desired urea concentration. For DLS measurements, detergent concentrations were chosen to reach micellar concentrations (i.e., detergent concentrations above the CMC) of 150 mM for OM, 125 mM for NM, 100 mM for DM, 75 mM for UM, 50 mM for DDM, 75 mM for CyMal-3 and CyMal-4, and 50 mM for CyMal-5 and CyMal-6. For ITC experiments, detergent stock concentrations were 750 mM for OM, 500 mM for NM, 250 mM for DM, 150 mM for UM, 50 mM for DDM, 1000 mM for CyMal-3, 750 mM for CyMal-4, 250 mM for CyMal-5 and CyMal-6, 1000 mM for OG, and 750 mM for NG. DLS. Detergent solutions were mixed by pipetting and transferred to 1-mL disposable polystyrene cuvettes (Sarstedt, Nümbrecht, Germany). Measurements were carried out at 25 °C on a Nano Zetasizer ZS90 (Malvern, Worcestershire, U.K.) with a detection angle of 90°. The light source was a He−Ne laser with a wavelength of 633 nm. The attenuator position was determined automatically by the instrument software and was typically 11. Temperature equilibration was performed for 2 min prior to each measurement. Repetition of representative experiments demonstrated high reproducibility. Data analysis included accounting for the changes in viscosity and refractive index resulting from the addition of urea. Because all distribution functions were unimodal, symmetric, and narrow, their maxima corresponded approximately to the Z-average diameter (Z) obtained from cumulants analysis.18 ITC. Detergent samples were freshly prepared from stock solutions in buffer containing the desired urea concentration. ITC experiments were performed at 25 °C on a VP-ITC (GE Healthcare) fitted with Kalrez perfluoroelastomer O-rings (ERIKS, Bielefeld, Germany) resistant to high urea concentrations. Detergent at a concentration above its CMC (typically, 10 × CMC) was loaded into the injection syringe, whereas the reference and sample cells were filled with buffer containing the same concentration of urea. Experimental settings included an injection volume of 3−10 μL, a time spacing of 360 s, a stirring speed of 310 rpm, a reference power of 58.6−125.5 μJ/s, and a filter period of 6−20 s. Repetition of some representative experiments demonstrated high reproducibility. When demicellization gave rise to excessively strong heat signals (i.e., at very high detergent concentrations in the syringe), the sample cell was filled with buffer containing detergent at a concentration below its expected CMC, to

Figure 1. Structures of the three non-ionic sugar-based detergent series used in this study. (A) Alkyl maltosides consist of a maltose headgroup and a linear alkyl chain. (B) Cyclohexyl alkyl maltosides contain a maltose headgroup and a cyclic alkyl chain. (C) Alkyl glucosides comprise a glucose headgroup and a linear alkyl chain.

headgroups, are frequently used in the purification, reconstitution, and crystallization of membrane proteins,2,16 and are commercially available in high amounts and purities. Alkyl maltosides comprise a maltose headgroup and a linear alkyl chain. We investigated alkyl maltosides with chain lengths in the range of 8−12 C atoms, that is, n-octyl-, n-nonyl-, n-decyl-, n-undecyl-, and n-dodecyl-β-D-maltoside (OM, NM, DM, UM, and DDM, respectively; Figure 1A). All of these detergents are well-soluble in water and have CMCs in a concentration range that is readily accessible to ITC and DLS experiments. Cyclohexyl alkyl maltosides also contain a maltose headgroup but a cyclic rather than a linear alkyl chain. Therefore, they possess a smaller solvent-accessible apolar surface area (SASA) and, thus, a higher CMC than their counterparts, that accommodate the same number of C atoms in linear alkyl chains. Members that are well-soluble and have readily accessible CMC values are found in the range of 9−12 C atoms in the alkyl chain, that is, 3-cyclohexyl-1-propyl-, 4cyclohexyl-1-butyl-, 5-cyclohexyl-1-pentyl-, and 6-cyclohexyl-1hexyl-β-D-maltoside (CyMal-3, CyMal-4, CyMal-5, and CyMal8503


Langmuir

Article

Figure 2. Demicellization of alkyl maltosides at increasing urea concentrations studied by ITC at 25 °C. (A) Differential heating power, Δp, versus time, t, for the demicellization of 42 mM DM (C10) into buffer containing 4 M urea. The injection volume was 5 μL. (B−F) Normalized heats of reaction, Q, versus average detergent concentration in the cell during the injection, cdet, at urea concentrations of 0−8 M. Experimental data (open circles) and sigmoid fits (solid lines) according to eq 1 are shown for detergents with chain lengths of (B) 12, (C) 11, (D) 10, (E) 9, and (F) 8 C atoms. (Inset in panel D) Analysis of a demicellization isotherm at 4 M urea, where CMC = 4.54 mM. The CMC is defined as the detergent concentration at the point of inflection (filled circle), and the heat of demicellization, Qdemic, is given by the difference between the pre- and posttransition baselines (dotted lines) at the CMC (double-headed arrow).

The heat of demicellization, Qdemic, was taken as the difference between the pre- and post-transition baselines A1 and A2, respectively, at the CMC (Figure 2D). From normalization of Qdemic from the total detergent concentration in the syringe, csyr, to the concentration of micellar detergent, (csyr − CMC), the standard molar micellization enthalpy, ΔH°,mic, is given by csyr ΔH °,mic = −Q demic csyr − CMC (2)

ensure that the heat signals stayed within the dynamic range of the instrument. For CMC values >50 mM, two consecutive ITC titrations were performed. After completing the first titration, we retained the sample in the calorimeter cell, removed the expelled volume accumulated on the ledge, refilled the syringe with detergent, and started the second titration. The two data files were then concatenated for data analysis. Automated baseline adjustment, peak integration, and normalization of reaction heats with respect to the molar amount of detergent injected were accomplished using NITPIC.19 Data were fitted to a generic sigmoid function (see eq 1) by nonlinear leastsquares fitting20 using the Solver add-in (Frontline Systems, Incline Village, NV).

■

Determination of Kmic, ΔG°,mic, and −TΔS°,mic. The mole fraction partition coefficient of detergent from the aqueous (aq) phase into micelles (mic), Kmic, is defined as the ratio of the detergent mole fraction in micelles, Xmic, to that in the aqueous phase, Xaq.

THEORETICAL BASIS Demicellization by ITC. In an ITC demicellization titration,21,22 a micellar solution of detergent at a concentration above its CMC is titrated into buffer. Initially, detergent dilution below the CMC leads to micelle disintegration, which is accompanied by either exothermic or endothermic heats. As the titration proceeds and the cell concentration reaches the CMC, the injected micelles no longer dissociate; hence, only heat of dilution is recorded. Demicellization isotherms do not reveal a sharp transition from the monomeric to the monomer/ micelle coexistence range but rather a smooth, gradual decrease in the absolute heat of reaction, Q (Figure 2D). Therefore, we fitted all demicellization isotherms in terms of a generic sigmoid function with linear pre- and post-transition baselines denoted as A1 and A2, respectively, according to Q=

A1 − A 2 1 + e(cdet − CMC)/ Δ

+ A2

K mic ≡

X mic = X aq

1 CMC c H2O + CMC

(3)

Because the CMC is very small compared with the molar concentration of water, cH2O = 55.5 M, Kmic takes the form K mic ≈

1 CMC c H2O

=

55.5 M CMC

(4)

Accordingly, the CMC is determined by and inversely proportional to Kmic. The standard molar Gibbs free energy change upon micellization, ΔG°,mic, is calculated as ⎛ CMC ⎞ ⎟ ΔG°,mic ≡ −RT ln K mic = RT ln⎜ ⎝ 55.5 M ⎠

(1)

where cdet is the average detergent concentration during the injection and Δ reflects the width of the transition region. Thus, the CMC is taken as the detergent concentration at the inflection point of the sigmoid transition.

(5)

with R being the universal gas constant and T being the absolute temperature. The Gibbs−Helmholtz relation gives the standard molar entropic contribution to micellization as 8504


Langmuir

Article

Figure 3. Influence of urea on the CMC values of the three homologous detergent series studied: (A) alkyl maltosides, (B) cyclohexyl alkyl maltosides, and (C) alkyl glucosides with 8−12 C atoms in their apolar chains. Experimental CMC values (open circles) and exponential fits (solid lines).

Figure 4. Influence of urea on the thermodynamics of micelle formation: (A and D) alkyl maltosides, (B and E) cyclohexyl alkyl maltosides, and (C and F) alkyl glucosides with 8−12 C atoms in their apolar chains. (Top) Standard molar Gibbs free energy change upon micellization, ΔG°,mic. Experimental data (open circles) from eq 5 and linear fits (solid lines). (Bottom) Standard molar changes (denoted as ΔY°,mic) in enthalpy, ΔH°,mic (open circles), and entropy, −TΔS°,mic (open diamonds), obtained from eqs 2 and 6, respectively. Dotted lines are merely guides to the eye.

−T ΔS°,mic = ΔG°,mic − ΔH °,mic

enthalpy change upon micellization of ΔH°,mic = 1.3 kJ/mol. Normalization takes into account that 4.54 mM (11%) of the detergent in the syringe was already monomeric before the injection and, therefore, did not contribute to the demicellization heat. Demicellization isotherms for cyclohexyl alkyl maltosides and alkyl glucosides are given in Figures S1 and S2 of the Supporting Information, respectively. For all detergents studied, the addition of urea leads to a broadening of the transition region. CMC values and corresponding mole fraction partition coefficients, Kmic, are listed in Tables S1 and S2 of the Supporting Information, respectively. As exemplified in Figure 3, the CMC increases exponentially, and, consequently, Kmic decreases exponentially with urea concentration. The effect is more pronounced for long-chain detergents than for their short-chain counterparts. For instance, the CMC of DDM (C12) in the presence of 8 M urea is 7.2 times higher than the CMC in urea-free buffer, whereas the corresponding factor for NM (C9) amounts to only 3.6. Thermodynamics of Micelle Formation. In Figure 4, ΔG°,mic (A−C) as well as ΔH°,mic and −TΔS°,mic (D−F) are plotted as functions of curea for alkyl maltosides (A and D), cyclohexyl alkyl maltosides (B and E), and alkyl glucosides (C

(6)

■

RESULTS To systematically study the influence of urea on detergent micelle properties, we investigated three homologous series of non-ionic alkyl detergents to quantify the thermodynamics of micelle formation and to determine their hydrodynamic size using ITC and DLS, respectively. Effect of Urea on the CMC. Figure 2 depicts demicellization isotherms for the alkyl maltoside series in the absence and presence of increasing urea concentrations and, as an example, shows the thermogram (A) and corresponding isotherm (inset in panel D) recorded upon injection of 5-μL aliquots of 42 mM DM (C10) into buffer at a urea concentration of 4 M. Initially, micelle disintegration was accompanied by an exothermic heat of Q = −1.5 kJ/mol. As the DM concentration in the calorimeter cell approached the CMC, Q drastically decreased in magnitude and finally leveled off to a dilution heat of Qdil = 0.02 kJ/mol. The CMC, defined as the DM concentration at the inflection point, was found to be 4.54 mM. By normalizing Qdemic to the molar amount of micellar detergent injected, we obtained a standard molar 8505


Langmuir

Article

and F). ΔH°,mic and −TΔS°,mic were not calculated for OM (C8) because the demicellization isotherm lacks well-defined pre- and post-transition baselines (Figure 2F). For all detergents, ΔG°,mic is negative at all urea concentrations and, as a corollary of the exponential decrease in Kmic, linearly increases with curea. In contrast, ΔH°,mic and −TΔS°,mic change in a nonlinear, counteracting fashion: ΔH°,mic is initially positive and decreases with curea, whereas −TΔS°,mic is negative and increases with curea. For the long-chain detergents UM (C11), DDM (C12), and CyMal-6 (C12), ΔH°,mic changes its sign as curea is increased (panels D and E of Figure 4). Thermodynamic Increments Per Methylene Group. ΔG°,mic, ΔH°,mic, and −TΔS°,mic linearly depend on the number of C atoms, n, in the apolar moieties of alkyl maltosides and cyclohexyl alkyl maltosides (see Figure S3 of the Supporting Information). The incremental contributions of a methylene group given by the slopes of these straight lines are shown in Figure 5 as functions of urea concentration. For both alkyl Figure 6. Influence of urea on the micelle size as assessed by DLS at 25 °C: (A) DDM and (B) CyMal-6. Normalized autocorrelation functions, C(τ), versus delay time, τ, obtained in the presence of 0−8 M urea with micellar detergent concentrations of 50 mM (i.e., cdet = CMC + 50 mM). (Insets) Mean values of the Z-average diameter from triplicate measurements as functions of curea. Experimental data (open circles) and linear fits (solid lines) for (A) alkyl maltosides and (B) cyclohexyl alkyl maltosides with 8−12 C atoms in their apolar chains.

■

DISCUSSION With increasing concentrations of urea, the CMC values of the non-ionic alkyl detergents investigated increase (Figure 3), while their micelle sizes decrease (Figure 6). Although the enthalpic and entropic terms change nonlinearly, the resulting Gibbs free energies of micellization are linear functions of urea concentration (Figure 4). Moreover, changes in the Gibbs free energy increments per methylene group result from the entropic contribution for alkyl maltosides but are of enthalpic origin for cyclohexyl alkyl maltosides. In both cases, however, they depend on urea concentration in a linear manner (Figure 5). Micellization in the Presence of Urea. To a first approximation, the effect of urea on the chemical potential of detergent in micelles, μmic, may be assumed to be small in comparison with its influence on the chemical potential of detergent monomers in the aqueous phase, μaq. This assumption is supported by fluorescence studies demonstrating that the interior of Triton X-100 micelles is not affected by urea and that urea/detergent interactions take place only at the micelle surface.12,23 In contrast, urea is expected to substantially reduce μaq by solvating the apolar moieties of detergent monomers and weakening the hydrophobic effect, thus depressing the water-to-micelle partition coefficient and raising the CMC (Figure 3 and see Table S2 of the Supporting Information). All CMC values that we determined in the presence of urea are in fair agreement with those reported in the literature, most of which have been derived from fluorescence-based approaches (Table 1). Urea causes a substantial broadening of the transition region from the monomeric concentration range to the monomer/ micelle coexistence range (Figure 2 and see Figures S1 and S2 of the Supporting Information) and leads to a decrease in the micelle size (Figure 6). The latter observation can be explained

Figure 5. Influence of urea on the increments in ΔG°,mic, ΔH°,mic, and −TΔS°,mic (denoted as ΔY°,mic) per methylene group, ∂ΔY°,mic/∂n: (A) alkyl maltosides and (B) cyclohexyl alkyl maltosides. Experimental data (open circles) and linear fits (solid lines).

maltosides and cyclohexyl alkyl maltosides, ∂ΔG°,mic/∂n increases linearly with curea; for instance, each additional methylene group reduces ΔG°,mic by −3.1 kJ/mol in buffer but only by −2.6 kJ/mol in the presence of 8 M urea. The effect of urea originates solely from the entropic term, −∂TΔS°,mic/ ∂n, for alkyl maltosides but exclusively from the enthalpic term, ∂ΔH°,mic/∂n, for cyclohexyl alkyl maltosides. Influence of Urea on the Micelle Size. Changes in the micelle size resulting from the addition of urea were monitored with the aid of DLS. Figure 6 exemplary depicts DLS autocorrelation functions obtained for DDM and CyMal-6 micelles at different urea concentrations. With increasing curea, the autocorrelation functions became slightly broader but decayed at shorter correlation times. The former observation is due to a slight increase in polydispersity, whereas the latter reflects a decrease in hydrodynamic size as expressed by the Zaverage diameter (insets in Figure 6). In the case of DDM (C12), for instance, Z = 7.2 nm in buffer but is reduced to Z = 5.8 nm in the presence of 8 M urea. 8506


Langmuir

Article

for a strong increase in the entropic contribution −TΔS°,mic = −23.7 kJ/mol, thus rendering micellization less favorable with ΔG°,mic = −26.9 kJ/mol. Although the ΔG°,mic increments per methylene group are almost identical for alkyl maltosides and cyclohexyl alkyl maltosides, they completely differ with respect to their enthalpic and entropic components (Figure 5). Exothermic micellization enthalpies have been attributed to interchain van der Waals interactions in the micellar state not observed in liquid hydrocarbons,27 and a so-called “nonclassical hydrophobic effect” characterized by a negative enthalpy change rather than an entropy gain has been invoked to rationalize the partitioning of various apolar compounds into lipid bilayer membranes.28 Hence, it appears tempting to speculate that the invariance of the micellization enthalpy of alkyl maltosides is indicative of the absence of urea-induced disruption of enthalpically favorable interactions within the apolar micelle core, because the linear alkyl chains are flexible enough to accommodate the above-mentioned changes in headgroup hydration and micelle size. Then, urea would weaken micelle formation exclusively by attenuating the entropic gain associated with the burial of apolar surface. Conversely, the bulky cyclohexyl alkyl chains most likely cannot fully adapt to urea-induced changes, resulting in both an enthalpic penalty as well as a concomitant gain in chain entropy that virtually offsets the entropy loss due to attenuation of the hydrophobic effect. However, it should be noted that other scenarios might be equally plausible, because enthalpic and entropic contributions to the solvation of linear and cyclic alkanes are complex and often cancel each other to a considerable extent.29 Classification of an intra- or intermolecular process as being hydrophobic on the basis of its enthalpic and entropic components alone may be misleading. A large negative heat capacity change accompanying the dehydration of apolar surfaces describes the hydrophobic effect best, regardless of the entropic and enthalpic contributions.30,31 Comparison to Transfer Free Energies of Amino Acid Side Chains. Interactions between urea and apolar moieties, such as detergent alkyl chains or amino acid side chains, can be quantified on the basis of transfer free energies. According to Tanford’s group transfer model, the overall Gibbs free energy accompanying the transfer of a molecule from water to urea, ΔG°,urea/H2O, is the sum of the transfer free energies of its constituent solvent-exposed moieties.32 Water-to-urea transfer free energies have been determined for a number of amino acid side chains.33 For example, the difference between asparagine and glutamine in the transfer free energies from water to 8 M urea suggests an increment per methylene group of −0.83 kJ/ mol, whereas the difference between alanine and leucine corresponds to an increment of −0.43 kJ/mol. The latter value is in excellent agreement with the differences in ∂ΔG°,mic/∂n between buffer and 8 M urea solutions, which we found to be −0.46 and −0.49 kJ/mol for alkyl maltosides and cyclohexyl alkyl maltosides, respectively (Figure 5). Comparison to Urea-Induced Protein Unfolding. In vitro assays relying on chemical denaturants, such as urea, are invaluable in the elucidation of the forces that contribute to protein stability. To obtain reliable stability values, the folded and unfolded states of a protein need to be populated simultaneously at measurable levels. However, under physiological conditions, the unfolded state is usually present in such a small fraction that it can hardly be detected by most experimental methods.34 Thus, to quantify the Gibbs free

Table 1. Comparison of CMC Values in the Presence of Urea for Alkyl Maltosides and Alkyl Glucosides with Chain Lengths of 8, 10, or 12 C Atoms CMC (mM) curea (M) maltosides

C8

C10 C12 glucosides

C8

this study

2 4 6 8 2 4 2 4 2

30.9 41.3 53.9 ∼75b 3.07 4.54 0.256 0.453 29.9

4

38.6

literature value (reference)

methoda

∼42 ∼6413 ∼8013 ∼11013 2.6914 3.5414 0.2814 0.4114 ∼309 26.914 33.114

pyrene pyrene pyrene pyrene ANS ANS ANS ANS ITC ANS ANS

13

Pyrene, pyrene fluorescence; ANS, 1-anilino-8-naphthalene sulfonic acid fluorescence. bEstimate only because of a very shallow demicellization isotherm (Figure 2F). a

by the fact that urea accumulation at the polar/apolar interface enhances headgroup solvation23 and, thus, increases the effective size of the headgroup, which, in turn, results in a more pronounced intrinsic curvature.9 Moreover, accumulation of urea in the headgroup region decreases the interfacial tension and, hence, reduces the free energy penalty incurred upon exposure of apolar moieties to the aqueous solvent.9 This attenuation of the hydrophobic effect favors the formation of smaller micelles above the CMC and facilitates the population of premicellar detergent assemblies just below the CMC, both of which are expected to be less effective than larger micelles in shielding detergent alkyl chains from the solvent. Obviously, the phase-separation model24 cannot account for micelle shrinkage and transition broadening because it considers micellar assemblies as a thermodynamic phase without any reference to size and structure and because it implies a sharp transition in the demicellization isotherm at the CMC. Hence, we attempted to analyze our data in terms of the cooperativeaggregation model,25 which assumes both a defined maximum micelle size and population of smaller detergent assemblies, that gives rise to a finite transition width. Although this analysis qualitatively supported the conclusion that the urea-induced broadening effect reflects a decrease in the micelle size, overparametrization of the fitting model resulted in excessive parameter correlations that impaired a reliable quantification and meaningful interpretation (data not shown). It should also be noted that the micelle sizes reported herein (Figure 6) represent hydrodynamic diameters, while the geometries of alkyl maltoside and alkyl glucoside micelles are more appropriately described by oblate rather than spherical shapes.26 Influence of Urea on the Thermodynamics of Micellization. With increasing curea, micellization becomes less endothermic (or even exothermic for some long-chain detergents) but also less favorable in terms of entropy (Figure 4). As an example, the thermodynamics of DDM micellization in the absence and presence of urea shall be considered. In buffer, an unfavorable ΔH°,mic = 1.1 kJ/mol and a favorable −TΔS°,mic = −32.9 kJ/mol result in a strongly favorable ΔG°,mic = −31.8 kJ/mol. At 8 M urea, a moderate decrease in the enthalpic term ΔH°,mic = −3.2 kJ/mol only partly compensates 8507


Langmuir

Article

energy change upon folding, ΔG°,fold, solution conditions have to be found that shift the equilibrium toward the unfolded state, which may be achieved by changes in pH or temperature or the addition of chemical denaturants. Gibbs free energy values measured at finite urea concentrations are then extrapolated back to physiological conditions in the absence of denaturant. This is usually performed according to the linear-extrapolation model (LEM),35 which assumes a dependence of the form ΔG°,fold(curea) = ΔG°,fold(H2O) + mcurea. Here, ΔG°,fold(H2O) is the sought measure of protein stability in the absence of a denaturant, while the so-called m value corresponds to the slope of the linear change in the Gibbs free energy of folding with urea concentration. Although the LEM is by far the most popular model for the quantification of protein stabilities, its validity has been demonstrated in only a few cases.36 In contrast, demicellization experiments provide direct experimental access to ΔG°,mic values over the entire range of urea concentrations without any need for extrapolation, demonstrating a linear dependence upon curea (Figure 4). Thus, ΔG°,mic(H2O) values obtained by linear regression across all urea concentrations are identical to those measured directly in the absence of denaturant. As expected and borne out by Figure 7A, all Gibbs free energies of micellization are linear

amount to 0.058 kJ/(mol M) for alkyl maltosides and to 0.062 kJ/(mol M) for cyclohexyl alkyl maltosides. Assuming a SASA increment per methylene group of 29 Å2,38,39 we obtain a ∂m/ ∂ΔSASA value of about 2.0 J/(mol M Å−2). This is in broad agreement with ∂m/∂ΔSASA > 1.1 J/(mol M Å−2) estimated for the micellization of the non-ionic detergent C12EO89 but is considerably greater than (0.63 ± 0.50) J/(mol M Å−2) reported for the solvation of apolar protein surface.37 The larger ∂m/∂ΔSASA observed for detergent micelles compared to proteins may reflect improved packing and water exclusion in micelles formed by long-chain detergents,27 but alternative explanations, such as an overestimation of the SASA change accompanying protein unfolding as a result of residual structure in the unfolded state, cannot be ruled out. Extrapolation of the data in Figure 7B suggests that the m values for alkyl maltosides and cycloalkyl maltosides vanish at 1.8 and 2.5 methylene groups, respectively, which thus seem to experience no change in solvation upon micellization. This is in good accordance with values of 2.2 and 3.1 methylene groups previously determined27 by extrapolation of heat capacities accompanying the micellization of diacylphosphatidylcholines and lysophosphatidylcholines, respectively. Interestingly, alkyl maltoside and cyclohexyl alkyl maltoside detergents having the same number of C atoms in their apolar tails display not only very similar increments but also virtually identical absolute m values (Figure 7B), although the latter might be expected to experience a smaller increase in SASA upon demicellization. In contrast, the size of the headgroup has a significant effect, which becomes obvious by comparing alkyl maltosides and alkyl glucosides. In the absence of urea, the thermodynamics of micellization are virtually identical for these two detergent classes (Figures 4 and 7A), but the addition of urea entails slightly larger changes in ΔG°,mic for alkyl maltosides than that for alkyl glucosides. This translates into somewhat lower m values for the latter (Figure 7B), which agrees with the expectation that the glucose headgroup shields the apolar micelle core less well from the solvent than the larger maltose headgroup. Implications for Membrane-Protein Research. The above findings are of immediate relevance to the purification, unfolding, reconstitution, and crystallization of membrane proteins. For instance, an increase in the CMC caused by urea may impair protein-stability studies, where membrane proteins solubilized in detergent micelles are unfolded or refolded over a wide range of urea concentrations.5,6 In particular, it is crucial to differentiate between changes reflecting membrane-protein unfolding and effects originating from the disintegration of detergent micelles upon the addition of urea.7 Thus, such experiments should be performed at detergent concentrations above the CMC value at the highest urea concentration used, which may be much larger than the buffer CMC (Figure 3). Conversely, a urea-induced increase in the CMC could facilitate the crystallization of membrane proteins in lipidic cubic phases because the latter are stabilized by a reduction in the concentration of micellar detergent,40 which is known to promote their transformation into lamellar structures.41 Hence, urea may tip this equilibrium in favor of conditions prone to crystallization, which adds to its proposed beneficial effect resulting from a swelling of the mesophase to accommodate larger membrane proteins.10 Finally, changes in the micelle size upon the addition or removal of urea might affect the hydrophobic match or mismatch42 and, consequently,

Figure 7. Summary of the effects of urea on the thermodynamics of micelle formation in three homologous series of detergents: (A) ΔG°,mic(H2O) and (B) m as functions of the number of C atoms in the detergent chain, n. Experimental data (open circles) and linear fits (solid lines) for alkyl maltosides, cyclohexyl alkyl maltosides, and alkyl glucosides.

functions of alkyl chain length but depend on the type of apolar moiety, because detergents with cyclic alkyl chains differ from their linear-chain counterparts in terms of ΔG°,mic(H2O) by 3.3 kJ/mol, which is somewhat larger than the contribution of one methylene group to micelle stability (Figure 5). Changes in Micelle Stability with Urea: The m Value. m values derived from protein-stability measurements have been found to correlate with the change in solvent-accessible apolar surface area (SASA) upon unfolding.37 Consistent with this observation, Figure 7B reveals that m values pertaining to micelle formation increase linearly with alkyl chain length and, thus, the expected SASA change within each homologous detergent series. The increments in m per methylene group 8508


Langmuir

■

the assembly, stability, or solubility of membrane-embedded proteins.

■

ASSOCIATED CONTENT

S Supporting Information *

Demicellization isotherms for cyclohexyl alkyl maltosides and alkyl glucosides, tabular CMC and Kmic values, comparison of CMC values in buffer, and thermodynamic parameters and micelle sizes as functions of alkyl chain length. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Garavito, R. M.; Ferguson-Miller, S. Detergents as tools in membrane biochemistry. J. Biol. Chem. 2001, 276, 32403−32406. (2) Privé, G. G. Detergents for the stabilization and crystallization of membrane proteins. Methods 2007, 41, 388−397. (3) Mouillac, B.; Banères, J.-L. Mammalian membrane receptors expression as inclusion bodies in Escherichia coli. Methods Mol. Biol. 2010, 601, 39−48. (4) Banères, J.-L.; Popot, J.-L.; Mouillac, B. New advances in production and functional folding of G-protein-coupled receptors. Trends Biotechnol. 2011, 29, 314−322. (5) Buchanan, S. K. β-Barrel proteins from bacterial outer membranes: Structure, function and refolding. Curr. Opin. Struct. Biol. 1999, 9, 455−461. (6) Stanley, A. M.; Fleming, K. G. The process of folding proteins into membranes: challenges and progress. Arch. Biochem. Biophys. 2008, 469, 46−66. (7) Fiedler, S.; Broecker, J.; Keller, S. Protein folding in membranes. Cell. Mol. Life Sci. 2010, 67, 1779−1798. (8) Udho, E.; Jakes, K. S.; Buchanan, S. K.; James, K. J.; Jiang, X.; Klebba, P. E.; Finkelstein, A. Reconstitution of bacterial outer membrane TonB-dependent transporters in planar lipid bilayer membranes. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 21990−21995. (9) Patel, H.; Raval, G.; Nazari, M.; Heerklotz, H. Effects of glycerol and urea on micellization, membrane partitioning and solubilization by a non-ionic surfactant. Biophys. Chem. 2010, 150, 119−128. (10) Cherezov, V.; Clogston, J.; Papiz, M. Z.; Caffrey, M. Room to move: crystallizing membrane proteins in swollen lipidic mesophases. J. Mol. Biol. 2006, 357, 1605−1618. (11) Fisher, L. R.; Oakenfull, D. G. Micelles in aqueous solution. Chem. Soc. Rev. 1977, 6, 25−42. (12) Ruiz, C. C.; Sánchez, F. G. Effect of urea on aggregation behavior of Triton X-100 micellar solutions: A photophysical study. J. Colloid Interface Sci. 1994, 165, 110−115. (13) Andersen, K. K.; Wang, H.; Otzen, D. E. A kinetic analysis of the folding and unfolding of OmpA in urea and guanidinium chloride: Single and parallel pathways. Biochemistry 2012, 51, 8371−8383. (14) Walter, A.; Kuehl, G.; Barnes, K.; VanderWaerdt, G. The vesicleto-micelle transition of phosphatidylcholine vesicles induced by nonionic detergents: Effects of sodium chloride, sucrose and urea. Biochim. Biophys. Acta 2000, 1508, 20−33. (15) Briganti, G.; Puvvada, S.; Blankschtein, D. Effect of urea on micellar properties of aqueous solutions of nonionic surfactants. J. Phys. Chem. 1991, 95, 8989−8995. (16) Newstead, S.; Ferrandon, S.; Iwata, S. Rationalizing α-helical membrane protein crystallization. Protein Sci. 2008, 17, 466−472. (17) Warren, J. R.; Gordon, J. A. On the refractive indices of aqueous solutions of urea. J. Phys. Chem. 1966, 70, 297−300. (18) Koppel, D. E. Analysis of macromolecular polydispersity in intensity correlation spectroscopy: The method of cumulants. J. Chem. Phys. 1972, 57, 4814−4820. (19) Keller, S.; Vargas, C.; Zhao, H.; Piszczek, G.; Brautigam, C. A.; Schuck, P. High-precision isothermal titration calorimetry with automated peak-shape analysis. Anal. Chem. 2012, 84, 5066−5073. (20) Kemmer, G.; Keller, S. Nonlinear least-squares data fitting in Excel spreadsheets. Nat. Protoc. 2010, 5, 267−281. (21) Kresheck, G. C.; Hargraves, W. A. Thermometric titration studies of the effect of head group, chain length, solvent, and temperature on the thermodynamics of micelle formation. J. Colloid Interface Sci. 1974, 48, 481−493. (22) Paula, S.; Süs, W.; Tuchtenhagen, J.; Blume, A. Thermodynamics of micelle formation as a function of temperature: A high sensitivity titration calorimetry study. J. Phys. Chem. 1995, 99, 11742− 11751. (23) Raghuraman, H.; Pradhan, S. K.; Chattopadhyay, A. Effect of urea on the organization and dynamics of Triton X-100 micelles: A fluorescence approach. J. Phys. Chem. B 2004, 108, 2489−2496.

CONCLUSION We systematically characterized the impact of urea on the micelle properties of non-ionic detergents within the homologous series of alkyl maltosides, cyclohexyl alkyl maltosides, and alkyl glucosides. With increasing urea concentration, the absolute Gibbs free energy of micellization and the micelle size decrease linearly. Both effects are more pronounced for detergents with long alkyl chains, which bury more apolar surface area upon micelle formation. The resulting exponential increase in the CMC caused by the presence of urea has to be taken into account and may be exploited beneficially in various applications of membrane-protein research, including purification, unfolding, reconstitution, and crystallization trials.

■

Article

AUTHOR INFORMATION

Corresponding Author

*Telephone: +49-631-205-4908. Fax: +49-631-205-4895. Email: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Sebastian Fiedler and Dr. Carolyn Vargas (both from the University of Kaiserslautern) for fruitful discussions and helpful comments on the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) with Grant KE 1478/1-2 and the Stiftung Rheinland-Pfalz für Innovation with Grant 961-386261/969.

■

ABBREVIATIONS USED ANS, 1-anilino-8-naphthalene sulfonic acid; c, concentration; C, normalized autocorrelation function; CMC, critical micellar concentration; CyMal-3, 3-cyclohexyl-1-propyl-β-D-maltopyranoside; CyMal-4, 4-cyclohexyl-1-butyl-β-D-maltopyranoside; CyMal-5, 5-cyclohexyl-1-pentyl-β-D-maltopyranoside; CyMal6, 6-cyclohexyl-1-hexyl-β-D-maltopyranoside; DDM, n-dodecylβ-D-maltopyranoside; DLS, dynamic light scattering; DM, ndecyl-β-D-maltopyranoside; ΔG°, standard molar Gibbs free energy change; ΔH°, standard molar enthalpy change; ΔS°, standard molar entropy change; ITC, isothermal titration calorimetry; K, mole fraction partition coefficient; LEM, linear extrapolation model; m, m value; μ, chemical potential; n, number of C atoms in the apolar detergent chain; NG, n-nonylβ-D-glucopyranoside; NM, n-nonyl-β-D-maltopyranoside; OG, n-octyl-β-D-glucopyranoside; OM, n-octyl-β-D-maltopyranoside; p, power; Q, heat; R, universal gas constant; SASA, solventaccessible apolar surface area; t, time; T, absolute temperature; τ, delay time; Tris, tris(hydroxymethyl)aminomethane; UM, nundecyl-β-D-maltopyranoside; X, detergent mole fraction; ΔY°, ΔG°, ΔH°, or −TΔS°; Z, Z-average diameter 8509


Langmuir

Article

(24) Shinoda, K.; Hutchinson, E. Pseudo-phase separation model for thermodynamic calculations on micellar solutions. J. Phys. Chem. 1962, 66, 577−582. (25) Beck, A.; Li-Blatter, X.; Seelig, A.; Seelig, J. On the interaction of ionic detergents with lipid membranes. Thermodynamic comparison of n-alkyl-+N(CH3)3 and n-alkyl-SO4−. J. Phys. Chem. B 2010, 114, 15862−15871. (26) Oliver, R. C.; Lipfert, J.; Fox, D. A.; Lo, R. H.; Doniach, S.; Columbus, L. Dependence of micelle size and shape on detergent alkyl chain length and head group. PLoS One 2013, 8, No. e62488. (27) Heerklotz, H.; Epand, R. M. The enthalpy of acyl chain packing and the apparent water-accessible apolar surface area of phospholipids. Biophys. J. 2001, 80, 271−279. (28) Seelig, J.; Ganz, P. Nonclassical hydrophobic effect in membrane binding equilibria. Biochemistry 1991, 30, 9354−9359. (29) Gallicchio, E.; Kubo, M. M.; Levy, R. M. Enthalpy−entropy and cavity decomposition of alkane hydration free energies: Numerical results and implications for theories of hydrophobic solvation. J. Phys. Chem. B 2000, 104, 6271−6285. (30) Baldwin, R. L. Temperature dependence of the hydrophobic interaction in protein folding. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 8069−8072. (31) Fernández-Vidal, M.; White, S. H.; Ladokhin, A. S. Membrane partitioning: “Classical” and “nonclassical” hydrophobic effects. J. Membr. Biol. 2011, 239, 5−14. (32) Bolen, D. W.; Rose, G. D. Structure and energetics of the hydrogen-bonded backbone in protein folding. Annu. Rev. Biochem. 2008, 77, 339−362. (33) Nozaki, Y.; Tanford, C. The solubility of amino acids and related compounds in aqueous urea solutions. J. Biol. Chem. 1963, 238, 4074−4081. (34) Hilser, V. J.; García-Moreno, E. B.; Oas, T. G.; Kapp, G.; Whitten, S. T. A statistical thermodynamic model of the protein ensemble. Chem. Rev. 2006, 106, 1545−1558. (35) Greene, R. F., Jr.; Pace, C. N. Urea and guanidine hydrochloride denaturation of ribonuclease, lysozyme, α-chymotrypsin, and βlactoglobulin. J. Biol. Chem. 1974, 249, 5388−5393. (36) Santoro, M. M.; Bolen, D. W. A test of the linear extrapolation of unfolding free energy changes over an extended denaturant concentration range. Biochemistry 1992, 31, 4901−4907. (37) Myers, J. K.; Pace, C. N.; Scholtz, J. M. Denaturant m values and heat capacity changes: Relation to changes in accessible surface areas of protein unfolding. Protein Sci. 1995, 4, 2138−2148. (38) Hermann, R. B. Use of solvent cavity area and number of packed solvent molecules around a solute in regard to hydrocarbon solubilities and hydrophobic interactions. Proc. Natl. Acad. Sci. U.S.A. 1977, 74, 4144−4145. (39) Kresheck, G. C. Isothermal titration calorimetry studies of neutral salt effects on the thermodynamics of micelle formation. J. Phys. Chem. B 2009, 113, 6732−6735. (40) Sennoga, C.; Heron, A.; Seddon, J. M.; Templer, R. H.; Hankamer, B. Membrane-protein crystallization in cubo: Temperaturedependent phase behaviour of monoolein−detergent mixtures. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2003, 59, 239−246. (41) Misquitta, Y.; Caffrey, M. Detergents destabilize the cubic phase of monoolein: Implications for membrane protein crystallization. Biophys. J. 2003, 85, 3084−3096. (42) Killian, J. A. Hydrophobic mismatch between proteins and lipids in membranes. Biochim. Biophys. Acta 1998, 1376, 401−415.

8510


Impact of Urea on Detergent Micelle Properties - Langmuir (ACS

Recommend Documents