Behavior of Most Widely Spread Lipids in Isotropic Bicelles - Langmuir

Jun 20, 2018 - Isotropic bicelles are a widely used membrane mimetic for structural studies of membrane proteins and their transmembrane domains. Simp...
0 downloads 0 Views 2MB Size
Download PDF
Article Cite This: Langmuir 2018, 34, 8302−8313

pubs.acs.org/Langmuir

Behavior of Most Widely Spread Lipids in Isotropic Bicelles E. F. Kot,†,‡ A. S. Arseniev,†,‡ and K. S. Mineev*,†,‡ †

Shemyakin−Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences RAS, str. Miklukho-Maklaya 16/10, Moscow 117997, Russian Federation ‡ Moscow Institute of Physics and Technology, Institutsky per., 9, Dolgoprudnyi 141700, Russian Federation

Langmuir 2018.34:8302-8313. Downloaded from pubs.acs.org by UNIV OF SOUTH DAKOTA on 09/20/18. For personal use only.

S Supporting Information *

ABSTRACT: Isotropic bicelles are a widely used membrane mimetic for structural studies of membrane proteins and their transmembrane domains. Simple and cheap in preparation, they contain a patch of lipid bilayer that reproduces the native environment of membrane proteins. Despite the obvious power of bicelles in reproducing the various kinds of environments, the vast majority of structural studies employ the single lipid/detergent system. On the other hand, even if the alternative bicelle composition is used, the properties of mixtures are not characterized, and the mere presence of lipid bilayer and discoidal shape of bicelle particles is not confirmed. Here we present an extensive investigation of various bicellar mixtures and describe the behavior of bicelles with lipids other than classical DMPC, namely sphingomyelins (SM), phosphatidylethanolamines (PE), phosphatidylglycerols (PG), phosphatidylserines (PS), and cholesterol. These lipids are rarely used in modern structural biology, but can help a lot in understanding the influence of the membrane composition on the properties of both integral and peripheral membrane proteins. Additionally, the ability of diheptanoylphosphatidylcholine (DH7PC) to serve as a rim-forming agent was investigated. We followed the phase transitions as revealed by 31P NMR and size of particles measured by 1H NMR diffusion as the criteria of the proper morphology and structure of bicelles. As an outcome, we state that SM exclusively, and PG/PS in mixtures with zwitterionic lipids can form small isotropic bicelles, which reproduce the key features of lipid behavior in bilayers. Mixtures, containing exclusively the anionic lipids, fail to reveal the lipid phase transition and do not follow the size predicted for the ideal bicelle particles. PE and DH7PC are the unwanted components of bicellar mixtures, and cholesterol can be added to bicelles, however, with certain precautions. In combination with our several most recent works, this study provides a practical guide for the preparation of small isotropic bicelles.

■

INTRODUCTION Structure determination of membrane proteins is one of the most challenging tasks of modern structural biology. To obtain the native fold and functionally relevant state of the membrane protein one needs to first place it in the environment that adequately mimics the properties of cell membranethe membrane mimetic. Cell membranes have a great range of lipid compositions, and the intrinsic physical properties of lipid compounds can differ substantially, resulting in complicated averaged behavior of the bulk bilayer. Natural membranes usually contain both saturated and unsaturated, bilayer-forming and nonbilayer-forming lipids, providing in total the necessary fluidity, curvature and permeability for different agents. Finally, cell membranes are asymmetriclipid composition is different in the inner and outer leaflets of the membrane,1 and, furthermore, the cell membrane is inhomogeneous and may contain the microdomains with the altered lipid and protein composition and physical properties.1 Thus, one would like to have at hand a number of various membrane mimetics that could reproduce the properties of membranes of different kinds of cells or even different compartments of cell membrane. This set of conditions may be provided by phospholipid bicelles. © 2018 American Chemical Society

Bicelles are formed in the mixtures of lipids with specific surfactants, and are thought to have a discoidal shape of particles.2 However, the phase diagram of bicelles is very complex and at certain conditions the hexagonal phase, perforated lamellae or the particles of other morphology (e.g., rod-like micelles) can be observed in such mixtures, depending on the temperature and lipid concentration.3 Here and below only those particles that are disc-like and contain the plane patch of lipid bilayer are referred to as bicelles. Size of bicelles can be controlled easily in a wide range by varying the lipid-to-detergent ratio (q). Large bicelles, formed with q > 2−3 for the most common compositions, can spontaneously orient in strong magnetic fields, and are referred to as “anisotropic”.4,5 In turn, mixtures with smaller particles, which are not capable of such orientation, are named “isotropic bicelles” (IsoBs). The present work is focused on the smallest IsoBs with radii of 2.5−4.5 nm, which are a widely used membrane mimetic for protein structural studies with solution NMR spectroscopy.6−8 A recent work has revealed9 that even Received: May 3, 2018 Revised: June 20, 2018 Published: June 20, 2018 8302

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

observed earlier23 in large IsoBs (R > 3.5 nm) above the critical temperature, close to phase transition point of model lipid. Its magnitude depends on the q ratio, concentration and the type of detergent used. In small bicelles TIG does not take place. Recently9 we showed that TIG depends on collision rate and is likely to be due to the reversible bicelle fusion. TIG is an unwanted effect, since it introduces the additional factor (temperature) in bicelle preparations that needs to be taken into account. As we hope, the present work provides the information, necessary to understand the properties of an arbitrary IsoBs preparation and adequately model various properties of the cell membrane. Unlike the previous studies, reporting the use of nonstandard IsoBs mixtures without their characterization, we describe the various parameters of bicelle particles. It allows defining the set of lipid compositions that form IsoBs properly and reproduce the key aspects of the bilayer membrane behavior, and allows finding the lipid components that should not be used in bicelle preparations or could be used only with certain precautions.

the small IsoBs, providing the most narrow peaks in solution NMR spectra, contain a patch of lipid bilayer. This gives IsoBs a huge advantage over the detergent micelles, amphipols and organic solvents. Besides, IsoBs are cheaper and easier to prepare than the other type of bilayer-containing membrane mimeticlipid−protein nanodiscs,10 which are larger and, consequently, provide the worse resolved NMR spectra. In general, there are several membrane properties that may affect the behavior of integral or peripheral membrane proteins and need to be reproduced in mimetic systems. First, the bilayer thickness can influence both the structure and stability of the membrane proteins due to the hydrophobic mismatch or lipophobic effects.11,12 This parameter may be mimicked by altering the length of the acyl chains of bicelle-forming lipids. Second, the electrostatic properties of the bilayer surface are important, especially for the juxtamembrane regions of the proteins.13−15 This parameter may be reproduced by combining the lipids with various headgroups in bicelles, or preparing the bicelles from the native lipid extracts.16,17 Last, some proteins can interact with specific and low-abundant components of the cell membrane, such as gangliosides, cholesterol, inositol phosphates, etc.18,19 Such compounds can be simply added to the bicellar mixtures in necessary amounts.20−22 Despite the obvious power of bicelles in reproducing the various kinds of environments, the vast majority of structural studies employ the single system, where the most convenient bilayer lipiddimyristoylphosphatidylcholine (DMPC)is mixed with various detergents.6 On the other hand, even if the alternative bicelle composition is used, the properties of lipid/detergent mixtures are not characterized and the mere presence of lipid bilayer and discoidal shape of bicelle particles is not confirmed. Characterization of bicellar mixture is a necessary step for the structural study, because the researcher would want to understand the size, shape and lipid packing properties of IsoBs for any given composition. In our previous works we described the behavior of isotropic bicelles composed of various phosphatidylcholine lipids and several rim-forming agents.9,23,24 Here we focus on bicelles composed of the less frequently used lipids: cholesterol, phosphatidylglycerols (PG), phosphatidylethanolamines (PE), phosphatidylserines (PS), sphingomyelins (SM) in both pure state and in mixtures with phosphatidylcholines (PC). We use [(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) as a rim-forming surfactant. While the majority of CHAPS and DHPC properties as rim-forming agents for small IsoBs are equivalent, CHAPS is much cheaper, more chemically stable and performs better in supporting the native structure and activity of transmembrane proteins.23 Additionally, particles formed by diheptanoylphosphatidylcholine (DH7PC) as a rim-forming agent are investigated. In our work we address three major aspects of the IsoBs behavior. First, we study, whether the size of bicelles follows the “ideal model”23,25 with reasonable parameters. The improper character of the size on q-ratio dependence may indicate the altered morphology of particles in the mixture under investigation. Second, we investigate the lipid phase transitions, which are the essential property of bilayer lipid membranes. The technique to detect the phase transitions in IsoBs was reported recently and is based on the 31P NMR measurements.9 The absence of such a transition may be indicative of the absence of the lipid bilayer. Last, we investigate the temperature-induced growth phenomenon (TIG). TIG was

■

EXPERIMENTAL SECTION

Materials. Lipids and detergents 1,2-dimyristoyl-sn-glycero-3phosphocholine (DMPC), 1,2-dimyristoyl-sn-glycero-3-phosphatidylethanolamine (DMPE), 1,2-dimyristoyl-sn-glycero-3-phosphatidylglycerol (DMPG), 1,2-dimyristoyl-sn-glycero-3-phosphatidylserine (DMPS), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2dihexanoyl-sn-glycero-3-phosphocholine (DHPC), 1,2-diheptanoylsn-glycero-3-phosphocholine (DH7PC), 3-[(3-cholamidopropyl)dimethylammonio]- 1-propanesulfonate (CHAPS) were the product of Avanti Polar Lipids Inc., USA. Cholesterol, cholesteryl hemisuccinate (CHS) and sphingomyelin (SM) from egg yolk were provided by Sigma-Aldrich Inc., Germany. D2O was the product of Cambridge Isotope Laboratory, USA. Sample Preparations. Here and below q denotes the effective lipid/detergent ratio, taking into account the presence of the detergent in the monomeric form. Unless otherwise stated, mol/ mol % with respect to the whole amount of lipids are used to characterize the contents of certain lipids in IsoBs. Bicelles with q ∼ 1.5 were used to characterize the phase transitions and TIG, such solutions form particles which are large enough to experience these effects in a wide range of ambient temperatures. To record the radius on q dependencies, q of CHAPS-based bicelles was varied in the range 0.3−1.5, which corresponds to the conditions, potentially applicable in the structural studies with solution NMR spectroscopy. To prepare the bicelles, ∼10 mg of lipids in a dry powder were dissolved by the necessary amount of 10% w/w stock solution of CHAPS, DH7PC or DHPC, and then the total volume of the sample was adjusted to 500 μL by the addition of 30 mM NaPi or TRIS buffer, pH 7.0 ± 0.1, containing 10% of D2O. In case of no-salt samples, necessary pH was obtained by NaOH titration. The resulting mixture was frozen in liquid nitrogen and then heated to 40−50 °C in the ultrasonic bath for 3−4 times. To vary the q ratio, the mixture was either diluted by the stock solution of detergent or added to the necessary amount of lipid in dry powder and put into the ultrasonic bath for 5−6 min at 40 °C. For the dilution experiments, 10% w/w solution of bicelles at necessary q was prepared. To prepare the bicelles with cholesterol, lipids were mixed with cholesterol in chloroform, then dried on the concentrator for several hours and under high vacuum overnight. The resulting powder was dissolved by the 500 μL of the solution, containing the rim-forming detergent at necessary concentration, 30 mM NaPi buffer, pH 7.0 at 40 °C. No freeze−thaw cycles and sonification were appliedthese procedures resulted in the precipitation. According to our data (Figure S1), 10 min shaking at 40 °C is enough to prepare the sample of DMPC/CHAPS q = 1.5 IsoBs, with no need in additional 8303

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir procedures. Unlike cholesterol, CHS did not require the chloroform and was added dissolved in pH 7.0 0.1 M NaPi buffer at concentration of 100 mM, followed by 3−4 freeze−thaw cycles. No samples with cholesterol contents above 10% were prepared, due to the precipitation of lipids. New sample was prepared for each cholesterol concentration. Contents of cholesterol are given with respect to the total quantity of lipids (including the cholesterol but not rim-forming detergents) in the sample. The concentration of components was additionally controlled by 1D NMR spectroscopy. Suspension of DPPC and DMPE with CHAPS in aqueous solution resulted in DMPE precipitation. Applying the technique of mixing in chloroform (described above for cholesterol) made it possible to obtain the transparent sample with no precipitation at 50 °C; however, embedding DMPE appeared to be unstable at lower temperatures. 31 P NMR Experiments. 31P NMR spectra were recorded on the Bruker Avance 600 MHz spectrometer equipped with the broadband double resonance probe. A single pulse experiment was employed using the π/6 pulse of 4 us and relaxation delay of 3 s, acquisition time of 2.7 s with broadband WALTZ16 proton decoupling at the field strength of 1.42 kHz. A preacquisition delay of 10 min was used before each experiment to ensure the equilibrium state of the system. Spectra were referenced indirectly with respect to the proton signal of trimethylsilylpropanoic acid (TSP) at 0.0 ppm using the gyromagnetic ratio. To determine the chemical shifts and to measure the fractions of gel/liquid-crystalline bicelles, the 31P spectra were deconvoluted using the Lorentzian lineshapes and nonlinear regression in Wolfram Mathematica software. The temperature was varied in either direction to exclude the hysteresis and control the equilibrium state of the sample. The occurrence of phase transition upon cooling of bicelles was verified by 31P spectra, as described in our previous work.9 The peak, corresponding to the gel phase demonstrates the much greater temperature coefficient of the chemical shift, than the one for fluid phase. In case of fractional phase transition, when gel- and fluid-state particles coexist in solution, the phase transition is manifested by the splitting of phosphate group peak in the NMR spectrum into the narrow signal for the fluid state and the broad one for the gel state. If the transition is not fractional, it is expressed by the fracture of the temperature dependence of the chemical shift. 1 H NMR Experiments. All diffusion 1H NMR spectra were recorded on the Bruker Avance 700 spectrometer with the working frequency of protons equal to 700 MHz, equipped with the roomtemperature triple resonance TXI probe. Diffusion was measured using the double stimulated echo pulse sequence with the efficient suppression of convection effects and of the intense solvent signal.26 Diffusion experiments were recorded with 32 scans and 32 increments of the encoding gradient pulse (gradient pulse strength was varied linearly from 12 to 52 G/cm). Summary length of the encoding gradient pulse was 4.0 ms, duration of the diffusion delay was 300− 400 ms. The gradients were calibrated using the diffusion coefficient of residual water in 100% D2O at 25 °C. The at least 30 min interval was used after the change of the temperature to ensure the equilibrium state of the system. In order to observe the TIG or its absence, DOSY measurements were performed at least at two temperature points, above and below the phase transition temperature of the corresponding lipid. The whole range of measurements was 278−323 K. To minimize the effects of signal overlap and magnetic field inhomogeneity, the integral of the most intense and narrow peak at 3.2 ppm (303 K), corresponding to N-(CH3)n moieties, was used to determine the diffusion coefficients of DPPC, DMPC, SM and CHAPS in 1H DOSY spectra. In case of DMPG we utilized the peaks, corresponding to the terminal glycerol CH2 moiety between 3.70 and 3.6 ppm. For DMPS/CHAPS bicelles, the peak of acyl chain (CH2)n at 1.28 ppm was used. For DMPC/DH7PC particles the DMPC and DH7PC diffusion was measured using the integrals of acyl chain terminal CH3 group peaks at 0.84 and 0.91 ppm, respectively. This approach appeared to be the most error-resistant in comparison to averaging the diffusion coefficients, measured for several peaks of the

compounds. Additionally, the use of the most narrow peak allows the more adequate averaging of the diffusion coefficients between the particles of various size. To measure the diffusion of free monomeric DH7PC, 1H DOSY spectra of a separate 0.5 mM DH7PC sample were obtained at 313 and 293 K. The diffusion coefficient at 313 K was equal to 6.5 × 10−6 cm2/sec, which provides no evidence in favor of micelle formation. Data Analysis. To analyze the diffusion coefficients and convert them to the radii of particles, we used the formalism, described in details in our previous work.23 The analysis included the corrections for the diffusion obstruction due to the high concentration of particles and due to the discoidal shape of bicelles.27−30 Experimental errors were determined using the Monte Carlo analysis, where the normal distribution of diffusion coefficient with standard deviation, equal to approximation error, was generated. For lipids with myristoyl acyl chains (DMPC, DMPG, DMPS) the steric thickness of bilayer was considered to be 4.37 nm. For DPPC, POPC and SM the thickness was taken to be 4.65 nm. These values were obtained by adding 1.8 nm to the hydrophobic thickness, Dc, corresponding to the lipid acyl chains.31,32 Particles, utilizing DHPC and DH7PC, were described using the model of a bicelle with an elliptic semitoroidal rim:25 É ÅÄÅ 1/2 Ñ Ñ rq ÅÅÅÅ jij 2 32λ zyz ÑÑÑÑ zz ÑÑ R=r+ ÅÅπ + jjjπ + 4λ ÅÅÅ 3q z{ ÑÑÑ k ÅÇ ÑÖ

(1)

Here r is the length of detergent molecule, R is the radius of bicelle and λ is the molecular volume ratio, detergent over lipid. For CHAPS detergent, the cylindrical rim model23 was used: ÅÄÅ ÅÅ q R = rÅÅÅÅ1 + + ÅÅ λ ÅÇ

ÑÉ q(q + λ) ÑÑÑÑ ÑÑ ÑÑ λ ÑÑÖ

(2)

■

RESULTS AND DISCUSSION Selection of the Experimental Approach. In the present work, we investigate two parameters of small IsoBs particlessize and lipid phase transitions. To conduct the study, we implement an approach, based on the solution NMR spectroscopy. Our work is focused mainly on the particles, having radii in the range 2.5−4.5 nm. Such objects could be visualized directly only using the negative stain electron microscopy, which implies the drying of sample and strong interaction between the object and the stain. Bicelles are not stable enough for us to be confident, that such an approach would not distort the data. Thus, we utilize the NMR spectroscopy to determine the diffusion coefficients of IsoBs, which can be converted into the radii of bicelles by applying several corrections, which are listed in the Experimental Section. NMR has certain significant advantages over the dynamic light scattering (DLS) which is a widely used technique with similar capabilities. Both methods provide the diffusion coefficients, which are in mutual agreement;23 however, NMR performs better in the case of polydisperse solutions. NMR permits to treat separately the individual compounds in the mixture and even the different structural states of the same compound. Using NMR allows the determination of the monomeric detergent concentration in solution, which is an important parameter for the bicelle preparation. Finally, NMR allows to simultaneously measure the necessary hydrodynamic parameters and control the sample quality, bicelle q factor and presence of impurities. Thus, we consider NMR diffusion as the method of choice for the characterization of small IsoBs suspensions. 8304

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

Figure 1. (A) Radii of bicelle particles are plotted as a function of q ratio: DMPC/DHPC (green) 300 K (empty squares), 313 K (empty circles) and theoretical ideal bicelle model (dashed line); DMPC/DH7PC 293 K (blue) experimental radii (squares), ideal bicelle model with predicted r and λ for DH7PC (dotted line), ideal bicelle model fit of first 4 points (dashed line), 313 K (red) experimental radii (circles). Out-of-range R values are shown in parentheses. (B) 31P NMR spectra of DMPC/DHPC (left) and DMPC/DH7PC (right) q = 0.6 bicelles, recorded at various temperatures, as indicated. Peaks from the liquid-crystalline and gel phase DMPC are denoted by * and **, respectively.

Figure 2. (A) Superposition of 31P SM spectra in q = 1.3 and 0.5 SM/CHAPS bicelles at 293 K. Liquid crystalline and gel phase peaks are indicated by * and **, respectively. (B) The percentage of SM in the gel phase at different temperatures and q. (C) Temperature dependence of 31P chemical shifts at different q. Average linear approximation for liquid-crystalline (LC) and gel (G) SM are shown by dashed and dotted lines, respectively. (D) SM/CHAPS bicelles radii R at different q at 323 (red circles) and 308 (blue squares) K. Dashed lines show ideal bicelle model approximation of the R(q) data at 323 K (red) and 308 K (blue). Vertical gray dashed line delineates the q range where TIG takes place.

Using DH7PC as a Rim-Forming Agent. DH7PC is a short-chain phospholipid, known to form micelles with critical concentration (CMC) of about 1.4 mM33 and is used as a detergent for membrane protein solubilization.34,35 Besides, in several recent works36,37 DH7PC was used as rim-forming agent for DMPC IsoBs, however, this choice had no significant reason except for the reduced CMC of the detergent and was based mostly on the minor structural difference between the DH7PC and DHPC; the latter is known to form the bicelle rim properly. There is a work, reporting the size of DH7PC-based bicelles at several q,38 however, the detailed investigation of the

particles structure was not performed and the methodology of the measurements was not well established that time, several additional corrections need to be introduced to the reported data to do the proper analysis. Here we examined DMPC/ DH7PC mixtures of several q and compared the data with DMPC/DHPC IsoBs, studied earlier.23 We varied the q of the DMPC/DH7PC mixture in the range 0.84−0.18 and obtained the R(q) dependencies at two temperatures: 293 and 313 K (Figure 1). As one can observe, the behavior of DH7PC-based mixtures is completely different from the one of the DHPC-based. At higher temperature, the 8305

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

= 1.18 ± 0.26, r = 1.15 ± 0.13 nm at 323 K, which is close to the parameters of DPPC/CHAPS particles.23 Concentration of monomeric detergent was also of the same same order in SM/ CHAPS and PC/CHAPS bicelles, revealing the similar dependence on the q (Figure S2). 31P spectra of SM/ CHAPS mixture clearly revealed the fractional phase transition, which was observed for all considered q values (Figure S3), reduction of q decreases the effective phase transition temperature (Figure 2A). As seen on Figure 2B, the fraction of lipids in the gel phase is strongly diminished upon CHAPS titration. 31P chemical shift analysis at various q shows only the minor changes in the headgroup environment for both gel and fluid phases in small and large bicelles (Figure 2C). All these results are in agreement with those obtained earlier for DPPC/ CHAPS mixtures,9 indicating the formation of bicelles. We also made an attempt to reproduce the uniform liquidordered phase of lipids, by mixing the SM with cholesterol in IsoBs. However, even in the presence of small amounts (5− 10%) of cholesterol, the bicelles were not formed and lipids precipitated. Most likely, the liquid-ordered phase that is formed in SM/cholesterol mixture40 is not compatible with the bicellar structure and the lamellar phase is formed instead. Similarly, the samples containing DPPC/SM/cholesterol/ CHAPS 1:1:1:2 (q = 1.5) mixture were highly opaque, which suggests the presence of very large particles (>100 nm). Moreover, particles which are visible in NMR spectra are characterized by the effective q ratio of 0.3 instead of initially prepared 1.5. Therefore, the sample segregates and two types of particles are formed: very large bicelles or lamellae, which are visible by the naked eye and invisible in NMR spectra and small IsoBs. Two kinds of particles are in the equilibrium, and first kind, likely, contains the lipids in liquid-ordered phase. This is actually expected, because the SM/cholesterol enriched membrane microdomains are known to be not dissolved by mild detergents and form large patches of bilayer being mixed with CHAPS.41 This property of lipid rafts was initially in the basis of the method of their first identification. Thus, we can conclude here that while the use of DMPE should be avoided, SM is a suitable lipid for the bicelle preparation; however, our attempts to reproduce the liquidordered phase of lipids using the SM and cholesterol in small isotropic bicelles were unsuccessful. PG/PS in Isotropic Bicelles. Lipids with negatively charged headgroups are the other important component of lipid membranes. Cell membranes usually contain up to 20% of anionic lipids.1 They introduce the charge to the surface of lipid bilayer, which is necessary for some essential lipid− protein interactions, accomplished by the peripheral proteins and juxtamembrane regions.15,42 Most integral membrane proteins contain the positively charged and aromatic amino acids on the water/lipid interface, the presence of anionic lipids may be an important factor, determining the stability of such proteins and the orientation of TM domains.13 Large anisotropic bicelles were shown to be capable of including the fraction of negatively charged lipids,43 and there are structural and biophysical studies, performed in isotropic bicelles of anionic lipids,44,45 but the structure of such particles was never reported. Here we studied the mixtures of DMPG and DMPS with CHAPS to clarify the subject. While the lipid packing properties and phase transition temperatures of both DMPG and DMPS are similar to those of DMPC, in bicelles we observe a completely different behavior of anionic and zwitterionic lipids. First of all, the R(q)

size of bicelles is affected by the dramatic TIG and reaches 19 nm at q = 0.84. On the other hand, at 293 K the radii of bicelles are relatively small, but the character of R(q) dependence is distorted: above q = 0.6 the particles start to decrease their size upon increasing the q. We approximated the first four points of the R(q) dependence with the ideal bicelle model and obtained the following parameters: λ= 1.61 ± 0.14 and r = 2.0 ± 0.05 nm. Based on the properties of DH7PC (chain length and molecule volume) one could expect the rim thickness of 1.3 nm and λ equal to 0.65, if the bicelles are formed properly. According to 31P NMR spectra (Figure 1B), DMPC/DH7PC particles experience the lipid phase transition, however at lower temperatures than DMPC/DHPC mixtures of the same q. Besides, above 293 K, peaks corresponding to lipids and detergents merge into one broad signal. The coalescence of cross-peaks at high temperatures and decreased lipid phase transition point indicate the enhanced lipid/ detergent mixing in DMPC/DH7PC solutions. Thus, we can conclude that DMPC/DH7PC mixtures do not follow the ideal bicelle model with reasonable parameters and experience the enhanced TIG and lipid-detergent mixing. While below 283 K the presence of lipid bilayer is observed in such IsoBs, most likely, the particles of altered morphology are formed above the critical temperature of DMPC. Use of DH7PC as a bicelle rimforming agent should be avoided. Bicelles of SM/PE Lipids. In our previous studies, various PC bicelles were thoroughly examined under different conditions and using several detergents.9,23 Another zwitterionic headgroup, PE, lacks three methyl groups on the positively charged nitrogen atom. This difference leads to the conical shape of PE molecules, that form the inverted hexagonal phase in the aqueous environment. In cell membranes PE molecules play an important role in the curvature maintenance and are found mainly on the inner bilayer leaflet. Distinct behavior of DMPE was also observed in the lipid/detergent mixtures: pure DMPE in CHAPS or DHPC at various q lost transparency at temperatures lower than 313 K, showing no bicelle formation below phase transition and the intrinsic inability of DMPE to form the bilayer. We managed to embed as much as 20 mol % of DMPE into the DPPC/CHAPS and DMPC/CHAPS bicelles, but the particles were unstable at temperatures below 313 K and such mixed bicelles precipitated. Thus, DMPE is hardly suitable for the bicelle preparation both alone and in mixtures with other lipids. We continued our study with the other widespread zwitterionic membrane lipidSM. Lipid rafts are enriched with SM,39 which makes this lipid a prospective compound to mimic the properties of specific compartments of cell membrane. Different extracts of SM were already used in DHPC-based bicelles to study the Notch protein, but the structure of bicelle particles in such mixtures was never characterized.20 To study the behavior of SM in bicelles, we prepared several egg yolk SM/CHAPS mixtures with q ranging from 0.5 to 1.5 and measured the size of particles, concentration of monomeric detergent and 31P NMR spectra at several temperatures. According to our data, SM reproduced all key features of PC in isotropic bicelles. The R(q) dependence was almost identical to the one of DPPC/CHAPS mixture with pronounced TIG at q > 1.0 (Figure 2D). Ideal bicelle model approximation of the low-q points (with no TIG) with eq 2 resulted in λ= 1.66 ± 0.04, r = 1.40 ± 0.02 nm at 308 K and λ 8306

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

Figure 3. (A) Particle radii as functions of q ratio for DMPS (red, blue), DMPG (brown, yellow) and DMPC (black). Ideal bicelle model approximation is plotted in dashed. (B) 31P spectra of DMPG/DHPC (red) q = 0.6 (analogous to q = 1.3 for CHAPS), DMPG/CHAPS (green) q = 1.5, DMPS/CHAPS (blue) q = 1.5 and DMPC/CHAPS (black) q = 1.5 bicelles, recorded at 278 K. DMPC fluid and gel phases are indicated by * and **, respectively. Original spectra are plotted in dashed, solid curves show the result of approximation. (C) Particle radii for DMPC/DMPG/ CHAPS mixtures at 300 K. (D) Particle radii for DPPC/DMPG/CHAPS mixtures at 313 K. Data for DPPC/CHAPS bicelles were taken from the work.23

DMPC/CHAPS samples, containing 20 and 40% of DMPG and measured the R(q) dependencies at 300 and 313 K. As one could observe in Figures S4 and 3C, bicelles with 20% of DMPG almost perfectly reproduce the radii of DMPC bicelles at 300 K and are slightly greater at 313 K. For such particles the TIG is present but its magnitude is reduced for high q. The reduction of TIG may be explained by the electrostatic repulsion between particles, which prevents their fusion. On the other hand, the 40% DMPG bicelles reveal the average behavior between the 100% DMPG and 100% DMPC particles with the absence of TIG. This may indicate the changes in particle morphology that occur at higher PG contents. Thus, while at low abundance, DMPG can clearly form bicelles being mixed with PC lipid, at PG contents above 30−40% either the lipid packing properties change substantially, or the lipid/ detergent mixing starts to occur. As for the monomeric detergent concentration in solution, this property of bicelles changes gradually from 0 to 100% of DMPG (Figure S2), revealing the absence of any abrupt change in the free energy of CHAPS in the rim of anionic particles. The reduced concentration of CHAPS in solution, observed for the charged bicelles, is likely to be the consequence of the more favorable interaction between the CHAPS and PG/PS headgroups. To further investigate the properties of PG/PC bicelles, we studied the phase transitions with 31P NMR spectroscopy. For

dependencies are far from predicted by the ideal bicelle model. The particles are formed with size, much greater than observed for DMPC/CHAPS bicelles and no TIG is observed (Figure 3A). Instead of an almost linear dependence, the concave curve is revealed. Moreover, none of the prepared mixtures reveals the presence of lipid phase transition down to 278 K, including both bicelles based on CHAPS and on DHPC (Figure 3B). Addition of salt to the solution (up to 100 mM) does not change the behavior of particles. Thus, lipids with PS and PG headgroups do not form bicelles being taken alone, but rather associate in some kind of particles with the pronounced mixing between the lipids and detergents. This is supported by the concentration of monomeric detergent, which is reduced substantially in PG/PS bicelles in comparison to the PC ones (Figure S2). Reduction of the detergent concentration in bulk solution implies the altered mode of interaction between the lipid and detergent. Most likely, the electrostatic repulsion between the charged headgroups does not allow the tight packing of lipids into bilayer in the absence of the lateral pressure. This problem could be overcome by mixing the anionic lipids with zwitterionic ones, and we investigated mixtures of such kind to find out, whether the anionic lipids can be incorporated into the small isotropic bicelles. Mixing Charged Lipids with PCs. To study the behavior of mixed zwitterionic/anionic bicelles, we first prepared the 8307

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

Figure 4. (A) Fraction of DPPC in the gel phase in DMPG/DPPC/CHAPS q = 1.5 IsoBs, with different contents of DMPG in bicelles. Reference plot for 100% DPPC bicelles is shown in black. (B) The ratio between the normalized fluid phase fractions of DPPC and DMPG. Data for 10% is not shown due to the big errors. (C) The bicelle radius R as a function of temperature at different DMPG contents. (D) 31P spectra of 20% DMPG/DPPC/CHAPS IsoBs, q = 1.5, pH 7.0 at 303 K: at 50 mM NaCl (red), at 10 mM CaCl2 (green), without salt (Blue). Peaks corresponding to DMPG, fluid and gel, DPPC, fluid and gel are noted with ° and °°, * and **, respectively. Original spectra are plotted in dashed, solid curves show the result of approximation.

that purpose, we mixed the DPPC and DMPG with CHAPS at several ratios and recorded the 31P spectra at temperatures ranging from 278 to 333 K. DPPC is much more convenient than DMPC due its higher critical temperature, which allows the measurement in the broader range of temperatures. 31P spectra showed two well-separated peaks of PG and PC headgroups, both split into the narrow signal of the fluid phase and the broad one of the gel phase upon cooling below 303 K (Figure S5). While phase transition of pure DMPG bilayer takes place at 296 K, already at 298 K the gel phase DMPG peak is observed. The presence of DMPG phase transition in mixed bicelles, unlike the pure ones, is the evidence of bicelle formation. Deconvolution of the spectra allowed us to perform the quantitative analysis of the phase behavior. Adding DMPG decreases the effective phase transition temperature for DPPC, this is clearly seen in Figure 4A. This is expected, because the critical temperature of DMPG is much lower than the one of DPPC. The ratio between fluid PC fraction and fluid PG

fraction counted as (DPPC fluid /DPPC all )/(DMPG fluid / DMPGall) gradually changes from 1 at 310 K, when both lipids are entirely in the fluid state, to almost 0 at 278 K, when all of DPPC is in gel state while some part of DMPG still remains fluid (Figure 4B). This observation suggests the decrease of lipid miscibility or some extent of segregation between DMPG and DPPC upon cooling. It is important, that the fluidPC/fluidPG ratio depends on the temperature identically for all tested bicelle compositions, which implies that inclusion of up to 50% of DMPG to DPPC/CHAPS bicelles does not change substantially the profile of PC/PG mixing. 1H diffusion measurements showed that the addition of DMPG suppresses the TIG additively, though the critical temperature of TIG does not change substantially and is around 305 K (Figure 4C). Apparently, the presence of charged PG headgroups can make the bicelle collisions more rigid or decrease the collision frequency due to electrostatic repulsion, which can explain the observed effect. 8308

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

Figure 5. Spectra of DPPC/CHAPS IsoBs with 20% (red) and 40% (blue) of POPC. (A) 31P spectra, 293 K, fluid and gel phase peaks are marked by * and **, respectively. (B) 1H spectra, 293 K, peaks are numbered as indicated. Original spectra are plotted in dashed, solid curves show the result of approximation. (C) 1H DOSY radii, corresponding to peaks 1 (red), 2 (green) and 3 (blue) of 40% (rhombuses) and 20% (circles) POPC. P value ranges are noted to the left for 40% and to the right for 20% in gray. Red p value bars show the significant difference between radii for peak 1 at 20% and 40% POPC contents. (D) Populations of 1H peaks 1 (red circles), 2 (green squares) and 3 (blue rhombuses), and fluid phase 31 P peaks (gray empty squares) at 20% and 40% POPC contents. (E) Hydrodynamic radii of DPPC bicelles with 0% (black), 20% (red) and 40% (blue) POPC as a function of temperature. (F) The fraction of gel phase in DPPC/POPC bicelles plotted as a function of temperature.

10 mM CaCl2 reveals the emerging gel-phase peak of DMPG, the increase of DPPC gel phase ratio up to ca. 60% and the shift of broad peaks of ca. 0.4 ppm upfield (Figure 4D). The narrow peaks are not shifted substantially, indirectly showing no significant changes in the properties of fluid phase. The addition of Ca2+ also resulted in the lipid precipitation below 293 K. This suggests that calcium ions are the unwanted component of bicelle solutions, when negatively charged lipids are present. We need to point out here that DMPC/DMPG and DPPC/ DMPG mixtures were studied by different approaches and the data could not be compared directly, since the interaction mode between DMPG and DMPC/DPPC may differ substantially. To investigate this subject, we recorded the R(q) dependence for DPPC/DMPG/CHAPS bicelles, containing 20% and 40% of the charged lipid (Figure 3D). This experiment revealed that the size of DPPC/DMPG and DMPC/DMPG mixed IsoBs behaves identically. At 20% of DMPG the curve reproduces the dependence for bicelles, prepared exclusively of the zwitterionic lipid, while at 40% of

Anionic lipids in bilayers are known to be affected by both the monovalent and divalent cations.46,47 The divalent cations reduce the phase transition temperature of bilayers, while the monovalent ions increase it. The proper formation of DMPG/ DPPC/CHAPS IsoBs gave us an opportunity to observe the effect of monovalent (Na+) and divalent (Ca2+) cations on the behavior of the mixed bilayer in particles with 20% DMPG in the DPPC bilayer. We compared the following samples: in 30 mM TRIS buffer (as in the previous step); with no salts added, with 50 and 100 mM NaCl; with 1, 10, and 20 mM CaCl2; with 20 mM CaCl2 and 50 mM NaCl. 50 and 100 mM of Na+ ions did not influence the phase behavior significantly (Figure 4D), though they increased the TIG sharply (Figure S5). In the agreement with the theory of bicelle fusion as a source of the TIG, the effect of cations on bicelle size can be explained by shielding the negative headgroup charge of DMPG, which lowers the electrostatic repulsion and increases the effective collision rate. Ten mM of Ca2+ had a similar and slightly higher effect on TIG, though, unlike Na+, Ca2+ influenced the phase condition of both lipids greatly. The 31P spectrum at 303 K and 8309

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

Figure 6. (A) Radius of DMPC/CHAPS q = 1.5 bicelles is plotted as a function of temperature and cholesterol contents. (B) Radius of POPC/ CHAPS q = 1.6 bicelles is plotted as a function of temperature and cholesterol contents. (C) Radius of DMPC/CHAPS q = 1.5 bicelles is plotted as a function of temperature and contents of CHS. (D) Radius of DMPC/CHAPS q = 1.5 (blue), q = 1.4 (green) and q = 1.5 with 6.4% of cholesterol are plotted as a function of temperature.

bicelles were used previously to characterize various membrane proteins, including the integrins and ion channels.48−51 Here we investigate the POPC/DPPC/CHAPS mixtures on the example of q = 1.5 DPPC/CHAPS bicelles with 20% and 40% POPC. 31 P spectra, obtained below DPPC phase transition, revealed two peaks, corresponding to the liquid crystalline and gel phase lipids. The fraction of lipids in the gel phase reaches 80−90% at 278 K, implying that the gel peak includes the signal from both POPC and DPPC. This clearly indicates that the lipids are mixed in both the liquid-crystalline and gel states, providing the common signals in NMR spectra (Figure 5A,F). The effective critical temperature depends on the POPC contents and is reduced in POPC-abundant samples, which corresponds to the expected behavior of a mixture, analogously to the lipid bilayers. Higher contents of POPC also increases the average radii of bicelles above the critical temperature of DPPC and enhances the TIG (Figure 5E). A close look at 1H spectra of these IsoBs revealed three peaks in the choline region at temperatures below 298 K (Figure 5B). Neither the pure DPPC or POPC in bicelles showed more than one 1H choline peak in the whole experimental temperature range. Such partitioning of peaks points out at the different environment of choline group protons; it also allows to measure the diffusion for each of the peaks separately. Using the 1H DOSY we determined the Stokes radii, corresponding to peaks 1, 2, and 3. The results were in most cases substantially different, according to the Mann−Whitney test (Figure 5C). Joint quantitative analysis of 31P and 1H

DMPG, the averaged curve is observed. Therefore, it is likely, that the conclusions that we make regarding the properties of mixed zwitterionic/anionic bicelles have a general character. To sum up, in DPPC/DMPG mixtures we observe the isotropic bicelle particles that undergo phase transitions and have the characteristic size. An important conclusion here is that the cation-dependent and temperature-dependent behavior of lipid bilayers is well reproduced in the mixed DPPC/ DMPG bicelles, with the DMPG fraction of up to 50%. This is a clear evidence in favor of the bilayer formation in such kind of particles, despite the altered character of R(q) dependence and partial lipid segregation that are observed. Thus, while the isotropic bicelles containing exclusively the anionic lipids cannot be obtained, the charge of the native membranes may be mimicked by adding 20−40% of anionic lipids to the bicelles composed of PCs. Mixing PCs with Various Fatty Residues. In previous experiments lipids with the different headgroups were studied, the acyl chains were not taken into account. However, the behavior of bicelles made of the lipids with saturated and unsaturated fatty acids is of the great interest. The latter play an important role in controlling the fluidity of biological membranes as they distort the order of acyl chain packing and lower the transition temperature.1 Inclusion of lipids with unsaturated chains to the bicelles may help to investigate the effect of membrane fluidity on the properties of membrane proteins. We have reported the properties of POPC/CHAPS bicelles recentlythey are characterized by the dramatic TIG and can be hardly used in structural studies.23 POPC/DHPC 8310

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir

size has decreased to 24 nm and was less than 15 nm upon the addition of 10% cholesterol (Figure 6B). The effect cannot be attributed to the reduced headgroup area of cholesterol in comparison to DMPC or POPC, which is an important parameter, determining the size of particles in solution.25 It appeared that bicelles, with DMPC being not substituted by cholesterol, but simply removed from the solution (q = 1.4), have much greater size than bicelles with the cholesterol and higher effective q (1.5) (Figure 6D). The polar derivative of cholesterol, cholesteryl hemisuccinate (CHS), is known to mimic in the certain aspects the cholesterol in model membranes.53 We observed that CHS as well as cholesterol reduced the temperature-induced growth of isotropic bicelles, however, the effect of CHS is much weaker: 20% of CHS were equivalent to 7% of cholesterol in DMPC/CHAPS bicelles (Figure 6C). The effects of CHS may be attributed to the negative charge of the succinate moiety and electrostatic repulsion, the behavior of mixed DMPC/ CHS particles resembles the one of the DPPC/DMPG. However, this is not the case for cholesterol, which is neutral. The influence of cholesterol on the size of isotropic bicelles may be related to the altered lipid packing properties in cholesterol-containing particles. A tighter lipid packing in the case of cholesterol-containing bicelles should decrease the probability of bicelle fusion and weaken the TIG. Thus, we suggest that addition of cholesterol to IsoBs makes lipids pack more tightly, which is in agreement with computer simulations of PC bilayers.54 It may happen, that some analogue of the liquid-ordered phase is formed, however, our data are too indirect to state this, and 31P spectra reveal no difference between the DMPC/CHAPS and DMPC/Cholesterol/ CHAPS bicelles at temperatures above 298 K (Figure S6). Surprisingly, we did not manage to develop a reliable and reproducible protocol for the bicelle preparation with cholesterol contents above 10% mol/mol, due to the partial precipitation of sample at both high and low q and with both CHAPS and DHPC taken as a rim-forming agent. This, and the fact that all samples were unstable below 293 K makes the usage of cholesterol rather challenging and tricky from the viewpoint of experiment methodology. DMPC/cholesterol bicelles need to be always stored at high temperature and the fraction of cholesterol needs to be controlled at any point.

spectra of DPPC/POPC IsoBs at 283−293 K showed, that the fraction of liquid crystalline phase, measured as relative intensity of the narrow 31P peak, matches the fraction of peak 2 area on 1H spectra at all examined temperatures (Figure 5D). This confirms indirectly that peak 2 corresponds to the fluid phase of both lipids, while peaks 1 and 3 refer to the gel state of either POPC or DPPC. Peak 2 gives the smallest hydrodynamic radii of 3.0−3.5 nm, and this is in agreement with our earlier work, where we show, that smaller particles are predominantly in the fluid phase.9 Fraction of peak 1 grows upon POPC addition, suggesting that peak 1 corresponds to POPC in the gel state (and peak 3, to the gel DPPC). Phase transitions did not occur in pure POPC bicelles (its critical temperature is 271 K); therefore, the presence of gel POPC implies the mixing between POPC and DPPC in bicelle particles. It is noteworthy that at low temperatures most of the POPC is in the gel state. For 20% POPC/DPPC IsoBs at 283 K the fraction of gel POPC is almost 20%. Finally, both the radii corresponding to peaks 1 and 3 characterize such gelphase mixed bicelle particles. Since these two radii are statistically different at all tested temperatures, and the radii, measured for POPC, are always greater than the ones for DPPC, we can conclude that the radius of the gel-state particle grows with the POPC content. This may be explained by the peculiarities in the packing of POPC unsaturated chain in the gel phase. Phase transition observations allow us to conclude that POPC and DPPC form bicelles with CHAPS and are mixed even below the phase transition. In protein NMR only the homogenic systems of fluid-state IsoBs are applicable. According to the present data, the size distribution of DPPC/POPC bicelles is relatively broad (as follows from the radii of bicelles, averaged over the whole population of either POPC or DPPC in the gel phase, Figure 5C), suggesting that polydisperse solutions are obtained. As well, the radii of mixed DPPC/POPC bicelles are much greater than the ones of pure DPPC and show an expressed TIG, which is unwanted in case of solution NMR spectroscopy. Thus, we can conclude here that POPC is not the best choice as a bicelle-forming lipid, both taken alone and mixed with the DPPC. Only the very low-q mixtures need to be considered for the structural studies. Effects of Cholesterol/CHS. Cholesterol is also an important component of cell membrane, and is known to modify the lipid packing properties in bilayers. There are studies, reporting the effects of cholesterol on the protein structure in IsoBs,20,52 however, the influence of cholesterol on the behavior of small bicelles was never investigated. In the present work, we prepared several DMPC/CHAPS q = 1.5 mixtures, with contents of cholesterol varied in the range from 1 to 7% with respect to the total amount of bilayer lipids. We did not manage to assess the influence of cholesterol on the phase transition properties of bicelles, because the cholesterolcontaining mixtures appeared to be unstable below the critical temperature of DMPC and partially precipitated. On the other hand, the influence of cholesterol on the temperature-induced growth of bicelles was dramatic. We observed that the addition of cholesterol to IsoBs gradually decreases the size of particles above the phase transition temperature and almost does not affect the size of bicelles close to the critical point (Figure 6A). A similar effect but with much greater magnitude was observed for the POPC/CHAPS bicelles with cholesterol. Bicelles with q = 1.60 had radii of 74 nm at 313 K in the absence of cholesterol, in the presence of 5% of this lipid the

■

SUMMARY AND CONCLUSIONS Here we present an extensive investigation of various bicellar mixtures, focused on the behavior of bicelles with lipids other than classical DMPC. These lipids are rarely used in modern structural biology, but can help a lot in understanding the influence of the membrane composition on the properties of both integral and peripheral membrane proteins. Our study allowed to reveal the lipid mixtures that are capable and incapable of proper formation of small IsoBs. According to our data, particles made of exclusively anionic lipids do not reveal any characteristic features of IsoBs. PE alone does not form the soluble small particles at all and, being mixed with PC, precipitates at temperatures, which are normally used in structural studies. Lipids with unsaturated acyl chains are also the bad option for the small IsoBs, even in the mixture with saturated PCs. Such mixtures form bicelles, but demonstrate the high polydispersity and enhanced TIG, which are the unwanted phenomena. DMPC/DH7PC mixtures as well do not provide the proper IsoBs solutions. SM, being mixed with cholesterol and PC in bicelles also provides highly 8311

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir polydisperse and opaque solutions with small IsoBs in the fluid phase coexisting with large particles in the liquid-ordered state. These examples are rather instructive, since they demonstrate that any lipid/detergent mixture should be characterized prior to its using in the structural investigation. Some of the listed mixtures were already utilized and referred to as bicelles,36,37 including the work from our group,45 while the mere presence of lipid bilayer in such particles is questionable. On the other hand, we show here that SM can form IsoBs properly, and anionic lipids form IsoBs, being mixed with PCs. Cholesterol can be included to the PC particles at relatively high amounts (up to 10% of the total lipids), and slightly affects the lipid packing parameters in IsoBs. One of the major findings of our study is that if IsoBs are formed, they in many aspects reproduce the behavior of lipid bilayers. This refers to the dependence of the lipid phase transition temperature on the bicelle composition and the presence of monovalent and divalent cations in mixed IsoBs. Like in bilayers, the addition of lipid with low critical temperature decreases the critical temperature of the mixture, while the addition of divalent cations increases the critical temperature and changes the lipid packing properties of the lipids. This is an additional point in favor of the ability of isotropic bicelles to adequately mimic the cell membranes. Finally, in combination with our several most recent works,9,23,24 this study provides a practical guide for the preparation of small IsoBs. Now the ability of almost arbitrary lipids to form IsoBs properly is established, and the size dependence on the q ratio is known, as well as the concentration of free detergent at any given temperature and concentration. This knowledge provides an opportunity for the researcher to control the properties of a bicelle solution during the structural study and to investigate the various lipid-related effects, ensuring the proper parameters of the membrane mimetic environment.

■

bicelle; DMPC, 1,2-dimyristoyl-sn-glycero-3-phosphocholine; DMPE, 1,2-dimyristoyl-sn-glycero-3-phosphatidylethanolamine; DMPG, 1,2-dimyristoyl-sn-glycero-3-phosphatidylglycerol; DMPS, 1,2-dimyristoyl-sn-glycero-3-phosphatidylserine; POPC, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine; DPPC, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine; DHPC, 1,2-dihexanoyl-sn-glycero-3-phosphocholine; DH7PC, 1,2-diheptanoyl-sn-glycero-3-phosphocholine; CHAPS, 3-[(3cholamidopropyl)dimethylammonio]-1-propanesulfonate; CHS, cholesteryl hemisuccinate; SM, sphingomyelin from egg yolk; PC, phosphatidylcholine; PG, phosphatidylglycerol; PE, phosphatidylethanolamine; PS, phosphatidylserines; TSP, trimethylsilylpropanoic acid

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.8b01454. Figures S1−S6 (PDF)

■

REFERENCES

(1) Luckey, M. Membrane Structural Biology: With Biochemical and Biophysical Foundations; Cambridge University Press: Cambridge, 2008. (2) Gabriel, N. E.; Roberts, M. F. Interaction of Short-Chain Lecithin with Long-Chain Phospholipids: Characterization of Vesicles That Form Spontaneously. Biochemistry 1986, 25 (10), 2812−2821. (3) van Dam, L.; Karlsson, G.; Edwards, K. Direct Observation and Characterization of DMPC/DHPC Aggregates under Conditions Relevant for Biological Solution NMR. Biochim. Biophys. Acta, Biomembr. 2004, 1664 (2), 241−256. (4) Sanders, C. R.; Prestegard, J. H. Magnetically Orientable Phospholipid Bilayers Containing Small Amounts of a Bile Salt Analogue, CHAPSO. Biophys. J. 1990, 58 (2), 447−460. (5) Sanders, C. R.; Schwonek, J. P. Characterization of Magnetically Orientable Bilayers in Mixtures of Dihexanoylphosphatidylcholine and Dimyristoylphosphatidylcholine by Solid-State NMR. Biochemistry 1992, 31 (37), 8898−8905. (6) Warschawski, D. E.; Arnold, A. A.; Beaugrand, M.; Gravel, A.; Chartrand, É .; Marcotte, I. Choosing Membrane Mimetics for NMR Structural Studies of Transmembrane Proteins. Biochim. Biophys. Acta, Biomembr. 2011, 1808 (8), 1957−1974. (7) Sanders, C. R.; Sönnichsen, F. Solution NMR of Membrane Proteins: Practice and Challenges. Magn. Reson. Chem. 2006, 44 (S1), S24−S40. (8) Mineev, K. S.; Nadezhdin, K. D. Membrane Mimetics for Solution NMR Studies of Membrane Proteins. Nanotechnol. Rev. 2017, 6 (1), 13−32. (9) Kot, E. F.; Goncharuk, S. A.; Arseniev, A. S.; Mineev, K. S. Phase Transitions in Small Isotropic Bicelles. Langmuir 2018, 34 (11), 3426−3437. (10) Denisov, I. G.; Sligar, S. G. Nanodiscs for Structural and Functional Studies of Membrane Proteins. Nat. Struct. Mol. Biol. 2016, 23 (6), 481−486. (11) Yano, Y.; Matsuzaki, K. Measurement of Thermodynamic Parameters for Hydrophobic Mismatch 1: Self-Association of a Transmembrane Helix. Biochemistry 2006, 45 (10), 3370−3378. (12) Mokrab, Y.; Stevens, T. J.; Mizuguchi, K. Lipophobicity and the Residue Environments of the Transmembrane Alpha-Helical Bundle. Proteins: Struct., Funct., Genet. 2009, 74 (1), 32−49. (13) van Klompenburg, W.; Nilsson, I.; von Heijne, G.; de Kruijff, B. Anionic Phospholipids Are Determinants of Membrane Protein Topology. EMBO J. 1997, 16 (14), 4261−4266. (14) Pöyry, S.; Vattulainen, I. Role of Charged Lipids in Membrane Structures  Insight given by Simulations. Biochim. Biophys. Acta, Biomembr. 2016, 1858 (10), 2322−2333. (15) Khan, H. M.; He, T.; Fuglebakk, E.; Grauffel, C.; Yang, B.; Roberts, M. F.; Gershenson, A.; Reuter, N. A Role for Weak Electrostatic Interactions in Peripheral Membrane Protein Binding. Biophys. J. 2016, 110 (6), 1367−1378. (16) Liebau, J.; Pettersson, P.; Zuber, P.; Ariöz, C.; Mäler, L. FastTumbling Bicelles Constructed from Native Escherichia Coli Lipids. Biochim. Biophys. Acta, Biomembr. 2016, 1858 (9), 2097−2105.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

K. S. Mineev: 0000-0002-2418-9421 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The work is supported by the Russian Science Foundation, grant #14-14-00573. Experiments were partially carried out using the equipment provided by the IBCH core facility (CKP IBCH, supported by Russian Ministry of Education and Science, grant RFMEFI62117X0018).

■

ABBREVIATIONS NMR, nuclear magnetic resonance; TIG, temperature-induced growth; CMC, critical micelle concentration; IsoB, isotropic 8312

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313

Article

Langmuir (17) Struppe, J.; Whiles, J. A.; Vold, R. R. Acidic Phospholipid Bicelles: A Versatile Model Membrane System. Biophys. J. 2000, 78 (1), 281−289. (18) Pawar, A. B.; Sengupta, D. Effect of Membrane Composition on Receptor Association: Implications of Cancer Lipidomics on ErbB Receptors. J. Membr. Biol. 2018, DOI: 10.1007/s00232-018-0015-1. (19) Fantini, J.; Di Scala, C.; Baier, C. J.; Barrantes, F. J. Molecular Mechanisms of Protein-Cholesterol Interactions in Plasma Membranes: Functional Distinction between Topological (Tilted) and Consensus (CARC/CRAC) Domains. Chem. Phys. Lipids 2016, 199, 52−60. (20) Deatherage, C. L.; Lu, Z.; Kroncke, B. M.; Ma, S.; Smith, J. A.; Voehler, M. W.; McFeeters, R. L.; Sanders, C. R. Structural and Biochemical Differences between the Notch and the Amyloid Precursor Protein Transmembrane Domains. Sci. Adv. 2017, 3 (4), e1602794. (21) Barbosa-Barros, L.; de la Maza, A.; López-Iglesias, C.; López, O. Ceramide Effects in the Bicelle Structure. Colloids Surf., A 2008, 317 (1−3), 576−584. (22) Hamilton, P. J.; Belovich, A. N.; Khelashvili, G.; Saunders, C.; Erreger, K.; Javitch, J. A.; Sitte, H. H.; Weinstein, H.; Matthies, H. J. G.; Galli, A. PIP2 Regulates Psychostimulant Behaviors through Its Interaction with a Membrane Protein. Nat. Chem. Biol. 2014, 10 (7), 582−589. (23) Mineev, K. S.; Nadezhdin, K. D.; Goncharuk, S. A.; Arseniev, A. S. Characterization of Small Isotropic Bicelles with Various Compositions. Langmuir 2016, 32 (26), 6624−6637. (24) Mineev, K. S.; Nadezhdin, K. D.; Goncharuk, S. A.; Arseniev, A. S. Façade Detergents as Bicelle Rim-Forming Agents for Solution NMR Spectroscopy. Nanotechnol. Rev. 2017, 6 (1), 93−103. (25) Triba, M. N.; Warschawski, D. E.; Devaux, P. F. Reinvestigation by Phosphorus NMR of Lipid Distribution in Bicelles. Biophys. J. 2005, 88 (3), 1887−1901. (26) Zheng, G.; Price, W. S. Simultaneous Convection Compensation and Solvent Suppression in Biomolecular NMR Diffusion Experiments. J. Biomol. NMR 2009, 45 (3), 295−299. (27) Chou, J. J.; Baber, J. L.; Bax, A. Characterization of Phospholipid Mixed Micelles by Translational Diffusion. J. Biomol. NMR 2004, 29 (3), 299−308. (28) Martinez, V. A.; Thijssen, J. H. J.; Zontone, F.; van Megen, W.; Bryant, G. Dynamics of Hard Sphere Suspensions Using Dynamic Light Scattering and X-Ray Photon Correlation Spectroscopy: Dynamics and Scaling of the Intermediate Scattering Function. J. Chem. Phys. 2011, 134 (5), 054505. (29) Ortega, A.; García de la Torre, J. Hydrodynamic Properties of Rodlike and Disklike Particles in Dilute Solution. J. Chem. Phys. 2003, 119 (18), 9914−9919. (30) Jóhannesson, H.; Halle, B. Solvent Diffusion in Ordered Macrofluids: A Stochastic Simulation Study of the Obstruction Effect. J. Chem. Phys. 1996, 104 (17), 6807−6817. (31) Kučerka, N.; Nieh, M.-P.; Katsaras, J. Fluid Phase Lipid Areas and Bilayer Thicknesses of Commonly Used Phosphatidylcholines as a Function of Temperature. Biochim. Biophys. Acta, Biomembr. 2011, 1808 (11), 2761−2771. (32) Nagle, J. F.; Tristram-Nagle, S. Structure of Lipid Bilayers. Biochim. Biophys. Acta, Rev. Biomembr. 2000, 1469 (3), 159−195. (33) Hauser, H. Short-Chain Phospholipids as Detergents. Biochim. Biophys. Acta, Biomembr. 2000, 1508 (1−2), 164−181. (34) Chen, W.; Dev, J.; Mezhyrova, J.; Pan, L.; Piai, A.; Chou, J. J. The Unusual Transmembrane Partition of the Hexameric Channel of the Hepatitis C Virus. Structure 2018, 26 (4), 627−634. (35) Shivanna, B. D.; Rowe, E. S. Preservation of the Native Structure and Function of Ca2+-ATPase from Sarcoplasmic Reticulum: Solubilization and Reconstitution by New Short-Chain Phospholipid Detergent 1,2-Diheptanoyl-Sn-Phosphatidylcholine. Biochem. J. 1997, 325 (2), 533−542. (36) Lenoir, M.; Grzybek, M.; Majkowski, M.; Rajesh, S.; Kaur, J.; Whittaker, S. B.-M.; Coskun, Ü .; Overduin, M. Structural Basis of

Dynamic Membrane Recognition by Trans-Golgi Network Specific FAPP Proteins. J. Mol. Biol. 2015, 427 (4), 966−981. (37) Morgado, L.; Zeth, K.; Burmann, B. M.; Maier, T.; Hiller, S. Characterization of the Insertase BamA in Three Different Membrane Mimetics by Solution NMR Spectroscopy. J. Biomol. NMR 2015, 61 (3−4), 333−345. (38) Lu, Z.; Van Horn, W. D.; Chen, J.; Mathew, S.; Zent, R.; Sanders, C. R. Bicelles at Low Concentrations. Mol. Pharmaceutics 2012, 9 (4), 752−761. (39) Simons, K.; Sampaio, J. L. Membrane Organization and Lipid Rafts. Cold Spring Harb. Cold Spring Harbor Perspect. Biol. 2011, 3 (10), a004697−a004697. (40) Ipsen, J. H.; Karlström, G.; Mouritsen, O. G.; Wennerström, H.; Zuckermann, M. J. Phase Equilibria in the PhosphatidylcholineCholesterol System. Biochim. Biophys. Acta, Biomembr. 1987, 905 (1), 162−172. (41) Schuck, S.; Honsho, M.; Ekroos, K.; Shevchenko, A.; Simons, K. Resistance of Cell Membranes to Different Detergents. Proc. Natl. Acad. Sci. U. S. A. 2003, 100 (10), 5795−5800. (42) Güler, G.; Gärtner, R. M.; Ziegler, C.; Mäntele, W. LipidProtein Interactions in the Regulated Betaine Symporter BetP Probed by Infrared Spectroscopy. J. Biol. Chem. 2016, 291 (9), 4295−4307. (43) Marcotte, I.; Dufourc, E. J.; Ouellet, M.; Auger, M. Interaction of the Neuropeptide Met-Enkephalin with Zwitterionic and Negatively Charged Bicelles as Viewed by 31P and 2H Solid-State NMR. Biophys. J. 2003, 85 (1), 328−339. (44) Schmidt, T.; Suk, J.-E.; Ye, F.; Situ, A. J.; Mazumder, P.; Ginsberg, M. H.; Ulmer, T. S. Annular Anionic Lipids Stabilize the Integrin ΑIIbβ3 Transmembrane Complex. J. Biol. Chem. 2015, 290 (13), 8283−8293. (45) Mineev, K. S.; Goncharuk, S. A.; Goncharuk, M. V.; Volynsky, P. E.; Novikova, E. V.; Aresinev, A. S. Spatial Structure of TLR4 Transmembrane Domain in Bicelles Provides the Insight into the Receptor Activation Mechanism. Sci. Rep. 2017, 7 (1), 6864. (46) Coronado, R. Effect of Divalent Cations on the Assembly of Neutral and Charged Phospholipid Bilayers in Patch-Recording Pipettes. Biophys. J. 1985, 47 (6), 851−857. (47) Hauser, H. Effect of Inorganic Cations on Phase Transitions. Chem. Phys. Lipids 1991, 57 (2), 309−325. (48) Lau, T.-L.; Partridge, A. W.; Ginsberg, M. H.; Ulmer, T. S. Structure of the Integrin Beta3 Transmembrane Segment in Phospholipid Bicelles and Detergent Micelles. Biochemistry 2008, 47 (13), 4008−4016. (49) Situ, A. J.; Schmidt, T.; Mazumder, P.; Ulmer, T. S. Characterization of Membrane Protein Interactions by Isothermal Titration Calorimetry. J. Mol. Biol. 2014, 426 (21), 3670−3680. (50) Schmidt, T.; Situ, A. J.; Ulmer, T. S. Direct Evaluation of Protein-Lipid Contacts Reveals Protein Membrane Immersion and Isotropic Bicelle Structure. J. Phys. Chem. Lett. 2016, 7 (21), 4420− 4426. (51) Kim, D. M.; Dikiy, I.; Upadhyay, V.; Posson, D. J.; Eliezer, D.; Nimigean, C. M. Conformational Heterogeneity in Closed and Open States of the KcsA Potassium Channel in Lipid Bicelles. J. Gen. Physiol. 2016, 148, 119. (52) Barrett, P. J.; Song, Y.; Van Horn, W. D.; Hustedt, E. J.; Schafer, J. M.; Hadziselimovic, A.; Beel, A. J.; Sanders, C. R. The Amyloid Precursor Protein Has a Flexible Transmembrane Domain and Binds Cholesterol. Science 2012, 336 (6085), 1168−1171. (53) Massey, J. B. Effect of Cholesteryl Hemisuccinate on the Interfacial Properties of Phosphatidylcholine Bilayers. Biochim. Biophys. Acta, Biomembr. 1998, 1415 (1), 193−204. (54) de Meyer, F. J.-M.; Rodgers, J. M.; Willems, T. F.; Smit, B. Molecular Simulation of the Effect of Cholesterol on Lipid-Mediated Protein-Protein Interactions. Biophys. J. 2010, 99 (11), 3629−3638.

8313

DOI: 10.1021/acs.langmuir.8b01454 Langmuir 2018, 34, 8302−8313