Characterization of the Morphology of Fast-Tumbling Bicelles with

Apr 30, 2014 - KI Biobank, Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, SE-171 77 Stockholm, Sweden. §. Department of...
0 downloads 0 Views 8MB Size
Article pubs.acs.org/Langmuir

Characterization of the Morphology of Fast-Tumbling Bicelles with Varying Composition Weihua Ye,† Jesper Lind,†,‡ Jonny Eriksson,§ and Lena Mal̈ er*,† †

Department of Biochemistry and Biophysics, Center for Biomembrane Research, The Arrhenius Laboratory, Stockholm University, SE-10691 Stockholm, Sweden ‡ KI Biobank, Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, SE-171 77 Stockholm, Sweden § Department of Chemistry - BMC, Uppsala University, SE-75123 Uppsala, Sweden ABSTRACT: Small, fast-tumbling bicelles are frequently used in solution NMR studies of protein−lipid interactions. For this purpose it is critical to have information about the organization of the lipids within the bicelle structure. We have studied the morphology of small, fast-tumbling bicelles containing DMPC and DHPC as a function of temperature, lipid concentration, and the relative ratio (q value) of lipid (DMPC) to detergent (DHPC) amounts. Dynamic light scattering and cryotransmission electron microscopy techniques were used to measure the size of the bicelles and to monitor the shape and dispersity of the particles in the samples. The stability and size of DMPC-containing bicelle mixtures were found to be highly dependent on temperature and the total lipid concentration for mixtures with q = 1 and q = 1.5. Stable DMPC/DHPC bicelles are only formed at low q values (0.5). Bicelle mixtures with q > 0.5 appear to be multidisperse containing more than one component, one with rH around 2.5 nm and one with rH of 6−8 nm. This is interpreted as a coexistence of small (possibly mixed micelles) bicelles and much larger bicelles. Incubating the sample at 37 °C increases the phase separation. Moreover, low total amphiphile concentrations and low q values lead to the formation of a temperature-independent morphology, interpreted as the formation of small particles in which the DHPC and DMPC are more mixed. On the basis of these results, we propose the existence of a critical bicelle concentration, a parameter that determines the existence of bilayered bicelles, which varies with q value. This polymorphism was not observed at any concentrations for q = 0.5 bicelles, for which a small but detectable temperature dependence was observed at high concentrations. The results demonstrate that q = 0.5 mixtures predominantly form “classical” bicelles, but that caution is needed when using fast-tumbling mixtures with q values higher than 0.5.



INTRODUCTION For the past 30 years or so, mixtures of lipids and detergents have been used successfully to investigate membrane protein and membrane-bound peptide structure by different NMR methods. It has been recognized since the 1980s that mixtures of phospholipids and detergents, or bile salts, form soluble bilayers that at appropriate conditions, including the relative concentrations of the components, align in the magnetic field.1−5 These mixtures have been used to characterize the interaction of peptides and proteins with the bilayer,6−10 including structural studies of integral membrane proteins.11 Bicelles have been shown to have several advantages as membrane mimetic media compared to e.g. micelles.12−14 Bicelles have furthermore also been used in e.g. X-ray crystallographic studies of membrane proteins15 and conformation and location studies of membrane-bound drugs.16−19 Small micelle-like fast-tumbling bicelles, on the other hand, with a high content of detergent, have been used to investigate structure and membrane interaction of several peptides and membrane proteins by conventional solution-state NMR techniques.16,18,20−25 For solution NMR purposes, small isotropic bicelles (with low lipid to detergent ratios, q values) © 2014 American Chemical Society

are ideal for combining studies of structure and membrane interactions of peptides26,27 since their tumbling is isotropic and the reorientational diffusion of the lipids is fast enough to give reasonable solution-state NMR spectra. Translational diffusion measurements of bicelles and of peptides or protein fragments have been used to investigate how much of the peptide is bound to the bicelle or if the peptides or proteins change the overall behavior of the bicelles.28−31 Other studies have used NMR spin relaxation to elucidate the effect of proteins and peptides on lipid dynamics.25,28,32,33 In order for such results to be interpreted in a reasonable way, it is essential that the lipid mixtures are well characterized with regards to their composition and morphology. It is, however, not clear if the small fast-tumbling bicelles are really true disklike objects under all conditions, such as varying lipid concentration or temperature. Although extensive work has been done to characterize the phase behavior of high q value mixtures,4,5,16,18,34−40 less is Received: January 17, 2014 Revised: April 30, 2014 Published: April 30, 2014 5488

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

known about the details concerning the morphology of small, fast-tumbling bicelles. Small, fast-tumbling dimyristoylphosphatidylcholine/dihexanoylphosphatidylcholine (DMPC/DHPC) bicelles have been studied under some conditions using a variety of biophysical techniques.41−44 It has been shown that for certain concentration ratios of DMPC and DHPC the small bicelles appear to be disklike objects (q ≤ 0.5), in contrast to the magnetically aligned liquid crystalline phases formed by lipid mixtures with higher q values, for which more continuous phases are observed.34−36,42 Several structures of peptides and protein fragments in the presence of bicelles have been reported,16,20,45−47 and it has been shown that the conformation of peptides may be different in the more disklike bicelles as compared to in detergent micelles.45 Related reports have focused on studies of the association of peptides with bicelle surfaces by e.g. pulsed field gradient diffusion NMR methods and nuclear spin relaxation.28,30,31 To correctly interpret such data, it is absolutely crucial that the morphology of the bicelle mixture is known. In this study, we have investigated the temperaturedependent size distribution and morphology of DMPC/ DHPC bicelles at different temperatures with different q values and with different total lipid concentrations. Our dynamic light scattering results demonstrate that parameters such as the bicelle size and even the phase behavior of the mixture depend on not only temperature but also total lipid concentration and q value, which are important factors when using fast-tumbling bicelles as membrane mimetic media. Images obtained with cryo-transmission electron microscopy confirm the results and demonstrate further that the morphology of bicelle mixtures is highly dependent on concentration and q value.



Rh =

kBT 6πηD

(1)

where kB is Boltzmann’s constant, T the absolute temperature, η is the viscosity of the sample, and D the diffusion coefficient of the particles. For monodisperse particles undergoing Brownian motion, the decay rate, Γ, of the autocorrelation function from DLS, is given by Γ = Dq2, where q (not to be confused with q = [DMPC]/[DHPC]) is the magnitude of the scattering wave vector, q = (4πn/λ) sin(θ/2), determined by the wavelength of the light source, λ, the refractive index of the medium, n, and the scattering angle, θ. Together this gives Rh =

kBTq2 6πη Γ

(2)

For polydisperse particles there is instead a distribution of decay rates, and for nonspherical particles the Stokes−Einstein relationship must take into account the dimensions and shape of the particles. In this study refractive indices provided by the software, and a temperature corrected viscosities relationship, were used to interpret the data. No assumptions concerning the shape of the bicelles were made, and therefore the obtained hydrodynamic dimensions can only be viewed as “effective” sizes. The measured data were fitted to multiple correlation functions, a procedure that provides the contribution of each contributing size and gives a size distribution. Theoretically, there can be infinite numbers of particle sizes. For the size distribution analysis, one can increase the number of contributing particle sizes by reducing the increment size, and therefore the resolution of the size distribution increases. An analysis with high resolution can resolve multiple peaks but also increases the noise and artifacts while an analysis with low resolution gives less noise but wider and less well resolved peaks. All the data we obtained here were analyzed with both default resolution and the high resolution algorithm within the Zetasizer Software 6.34. Cryo-Transmission Electron Microscopy. A Zeiss EM 902A transmission electron microscope (Carl Zeiss NTS, Oberkochen, Germany), operating at 80 kV and in zero loss bright field mode, was used. Digital images were recorded under low-dose conditions with a BioVision Pro-SM Slow Scan CCD camera (Proscan GmbH, Scheuring, Germany) and iTEM software (Olympus Soft Imaging Solutions GmbH, Münster, Germany). To visualize as many details as possible, an underfocus of 1−2 μm was used to enhance the image contrast. In short, the method for sample preparation was as follows, with a more comprehensive description available in the previous work by Almgren et al.49 Samples were equilibrated at 310 K (37 °C) and ∼99% relative humidity within a climate chamber. A small drop of sample (∼1 μL) was deposited on a copper grid covered with a perforated polymer film and provided with thin evaporated carbon layers on both sides. Excess liquid was removed by means of blotting with a filter paper, leaving a thin film of the solution on the grid. Immediately after blotting, the sample was vitrified in liquid ethane kept just above its freezing point, 90 K. Samples were kept below 108 K and protected against atmospheric conditions throughout the transfer to the TEM and examination.

MATERIALS AND METHODS

Sample Preparation. The phospholipid 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), and the detergent 1,2-dihexanoyl-snglycero-3-phosphocholine (DHPC) were purchased from Avanti Polar Lipids, Inc. (Alabaster, AL). Solutions containing fast-tumbling bicelles were produced by mixing the desired amount of DMPC carefully with phosphate buffer, by vortexing and centrifuging the sample repeatedly until a homogeneous slurry was formed, and then adding an aliquot of a 1 M DHPC stock solution. This mixture was subjected to several cycles of centrifugation and vortexing until a clear nonviscous solution was obtained.48 Solutions were produced with three q ratios ([DMPC]/[DHPC]) equal to 0.5, 1.0, and 1.5, and the total concentration of DMPC and DHPC was 75 mM (cL (w/v) ≈ 4%), 150 mM (cL (w/v) ≈ 8%), or 300 mM (cL (w/v) ≈ 16%). Micelle samples containing 30 or 75 mM DHPC were simply produced by diluting the DHPC stock solution with buffer and then the sample was vortexed. The pH was in all cases adjusted to 5.7 with 50 mM phosphate buffer. Dynamic Light Scattering. The size of the bicelles was determined with dynamic light scattering over a temperature interval of typically 18−40 °C. Measurements were performed on a Zetasizer instrument (Nano ZS, Malvern Instruments, Worcestershire, UK) using UV-transparent disposable cuvettes of 1 cm path length. Although all samples were clear and nonviscous solutions, each sample was filtered by a syringe driven filter unit (0.22 μm, Millipore Corporation, Bedford, MA) right before measurements. Light scattering data were collected during 10−30 s and repeated 8 times and exported as autocorrelation functions. Data were processed by Zetasizer Software 6.34. The bicelle size distributions were characterized by the volume-averaged mean method. The results are presented as hydrodynamic radii, Rh, by relating derived decay constants via the Stokes−Einstein relationship:



RESULTS Bicelle Size Varies with Temperature, Concentration, and Composition. The sizes of DHPC micelles and of bicelles with different q ratios and with different lipid concentration were estimated by analyzing autocorrelation functions from the DLS measurements with the default algorithm within the manufacturer’s software. Although the size (hydrodynamic radius) of a particle such as a bicelle can be obtained from an analysis of the diffusion obtained from the autocorrelation functions, interpretation of a true bicelle size may be difficult due to the unknown shape of the bicelles and due to variations in the radius distribution, e.g., symmetric or asymmetric distributions of sizes. The relatively concentrated 5489

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

samples used in this work will also influence the estimated sizes due to particle−particle interactions. Nevertheless, the results can still be compared to yield information about how size and size distributions vary between samples produced under different conditions (e.g., q value, temperature, and concentration). Starting with the smaller DHPC micelles, we found that the hydrodynamic radius of the micelle did not vary significantly with temperature (Figure 1). We also noted that the size of the

Figure 1. Hydrodynamic dimensions of DHPC micelles as a function of temperature. Filled squares indicate data obtained for 75 mM DHPC, while dots indicate data for 30 mM DHPC. Both samples were at pH 5.7.

DHPC micelles did not vary with concentration. From data measured for 30 and 75 mM DHPC, well over the cmc, we observed that the micelles had an average radius of 2.2 ± 0.1 nm at 25 °C, independent of concentration. The variation of the DHPC micelle size with temperature (in the range 18−40 °C) was less than 15%, and no trends in the variation were observed with increasing or decreasing temperature or with DHPC concentration. The hydrated size of a DHPC micelle measured at 27 °C (2.0 nm) is in excellent agreement with previous measurements.41 In contrast to this result, the size variation of most bicelles with temperature was significant (Figure 2). The most profound effect occurred, not unexpectedly, for DMPCcontaining bicelles at around the temperature corresponding to the transition from the gel to liquid crystalline phase, approximately 24 °C (Figure 2). Also after this transition the bicelles were observed to become increasingly larger with higher temperature. As expected, the bicelles increased in size with increasing q values. The DMPC-containing q = 0.5 bicelles had an apparent hydrodynamic radius, rH, of 2.9−3.3 nm at 25 °C, regardless of concentration. For the 300 mM mixture, the size increased only marginally with temperature to 3.5 nm, while the particle size in the 150 mM bicelle mixture increased to 4.3 nm at 40 °C (Figure 2). Somewhat surprisingly, at the lowest concentration in this study, 75 mM, the formed particles were somewhat smaller than at the other concentrations, with a radius of around 3 nm, independent of temperature. This indicates that at this low concentration the aggregates that are formed are most likely more micelle-like. The temperature variation was much larger for bicelles with a higher DMPC content (i.e., higher q values), and for e.g. q = 1 bicelles, the increase in hydrodynamic radius over the temperature range studied here

Figure 2. Hydrodynamic dimensions for different DMPC/DHPC mixtures. Panel A shows results for q = [DMPC]/[DHPC] = 0.5, (B) for q = 1, and (C) for q = 1.5. Open squares indicate a total ([DMPC] + [DHPC]) concentration of 75 mM, closed circles show data for a total concentration of 150 mM, and triangles indicate 300 mM. The pH was in all cases 5.7. The data points are averages over eight measurements.

was a factor of 1.4 (3.1 nm at 18 °C vs 4.5 nm at 40 °C) for bicelles with a total lipid concentration of 300 mM and even more for bicelles with a lipid concentration of 150 mM (3.5 nm at 18 °C vs 5.8 nm at 40 °C). The bicelles in the 300 mM lipid mixture were also somewhat smaller than the ones in 150 mM mixtures. At 75 mM, the same observations as for q = 0.5 5490

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

Figure 3. Dynamic light scattering data for q = 1 bicelles, 150 mM total lipid concentration. Panel A shows a typical correlation curve from the dynamic light scattering measurements (at 34 °C). Panel B shows the effect of choosing high resolution for the analysis. Panel C shows the effect of removing the last 13 points from this curve when analyzing the data in terms of a distribution of particles with different hydrodynamic radii in the sample. In panel D, data from eight different measurements are shown with the high resolution analysis.

bicelles were made: the aggregates had an apparent radius of around 3 nm, independent of temperature, again indicating the presence of micelle-like particles. The data for q = 1.5 bicelles at high concentrations (150 and 300 mM) were much more scattered, but followed the same trend as observed for q = 1 bicelles. However, at this q value a clear temperature-dependent increase in size was observed for the sample with 75 mM, indicating that for higher q values the bicelles persist at lower total concentration. In conclusion, we found that for low q values, 0.5 and 1, the concentration of lipids appeared to affect the morphology of the mixture and that at a concentration of 75 mM the main component were smaller than the expected bicelle and therefore resembled mixed micelles rather than “classical” bicelles. At q = 1.5 this was not observed. These results together indicate that one must use caution when designing an appropriate lipid mixture for a membrane-mimicking agent. In e.g. NMR experiments, it is often preferable to use lower lipid concentrations, since this has less impact on the quality of the NMR spectrum, the solution properties, and the amount of protein or peptide that is needed for studying bilayer interactions. At the same time, lowering the concentration may cause bicelle morphology to change, and the “bicelle” bilayered structure may no longer prevail. DMPC/DHPC Bicelle Mixtures with Higher q Values Contain Several Species. When analyzing the correlation functions from the dynamic light scattering data, we also applied an analysis routine with high resolution, which more accurately resolves multiple peaks (as introduced in Materials and Methods section). For example, for a q = 1 mixture (150 mM) two distinct populations were observed when using high

resolution for analyzing a correlation curve containing 35 data points (Figure 3). A subpopulation of a smaller species was then observed together with larger bicelles in the mixture above the gel−liquid crystal phase transition temperature. The main population is slightly larger in size than the average size reported by the lower resolution analysis, while the subpopulation with smaller size has a radius more reminiscent of micelles (compare Figures 2B and 3D). Two populations were also obtained if we removed several of the latter data points in the standard analysis (which contained noise and may mask the presence of smaller populations). In all, eight separate measurements were performed for each temperature, and the separation of the two species was not observed in all measurements, possibly due to uncertainty in the individual measurements (e.g., uncertainty of the presence of each population within the exact measured volume region). The data analysis revealed that the smaller species account for around 30−50% of the total population, while the larger ones for 50−70% of the particle population. The results indicate that the lipid mixture does not contain a uniform distribution of a single type of lipid aggregate. Notably, the possible coexistence of two different species was only observed for bicelles prepared with q = 1 or 1.5, and never for q = 0.5, regardless of total concentration and of analysis method. Moreover, two species could be observed for the q = 1.5 DMPC/DHPC mixtures of all three concentrations above the DMPC transition temperature. These results indicate that the morphology of the bicelle mixtures depends on both the q values and concentrations. The hydrodynamic radius of the small species was found to be around 2.5 nm, which is close to what was observed for DHPC micelles, which indicates the presence of micelle-like 5491

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

particles. This can be explained by a higher degree of mixing between DHPC and DMPC, indicating a more mixed micellelike morphology. This results demonstrate that even at higher q values, small micelle-like bicelles are formed together with other, much larger species. Hence, the results indicate a phase separation for lipid mixtures with q values of 1 or higher. This has previously been observed to be the case for magnetically alignable bicelles with q values above 2.34−36,42 The larger species had a radius of around 6−8 nm, significantly larger than what is observed for q = 0.5 bicelle mixtures. To obtain more information about the morphology of DMPC/DHPC bicelle mixtures with different q values, cryotransmission electron microscopy was used. Cryo-TEM pictures were taken for bicelle mixtures made with two q values (0.5 and 1.5) and two total lipid concentrations (150 or 15 mM). In addition, images were also captured for 15 mM mixtures after incubating for 24 h at 37 °C. The images obtained for the q = 1.5 bicelles with a total lipid concentration of 150 mM were too dense for a reasonable interpretation. However, from the images it appeared that the sample was most likely dominated by small disks, but also some flat band-like structures, as indicated by the arrow in Figure 4A, were present. An estimate suggested that most of the disks had radii corresponding to 7 nm in the nondiluted q = 1.5 sample. To obtain a better image, and to investigate the effect of lowering the concentration, the sample was diluted 10-fold to a concentration of 15 mM. The main population was still composed of disk-shaped objects (Figure 4B), but the radii of the disks in the diluted sample were now larger (around 13 nm). Notably, this mixture became polydisperse as there were also other larger species present in this sample and even clusters of aggregated disks (Figure 4B). The diluted sample was incubated at 37 °C for 24 h, a procedure that led to an increase in the size of the disks (to a radius of 15 nm). Moreover, the number of “individual” nonclustered/nonfused disks appeared to be reduced in the sample after 24 h; i.e., the disks appeared to aggregate in the incubated sample leading to more clusters of aggregated disks (Figure 4C). Meanwhile, some elongated aggregates were also detected. The cryo-TEM images for the diluted samples with q = 1.5 revealed that the mixture under these conditions appeared to be polymorphic and that the morphology of the sample after 1 day differed from that of the fresh sample. Contrary to the q = 1.5 bicelle mixture, the results for the q = 0.5 mixture were much less complicated. As before, at 37 °C, cryo-TEM pictures were taken for the samples at total lipid concentrations of both 150 and 15 mM (Figure 5A,B). At both concentrations, small spherical- or disk-like particles dominated in the samples, and they were uniformly distributed. In these samples, the majority of disks had a radius of about 3.7 nm, in excellent agreement with the DLS results (Rh = 4.1 nm at 37 °C for q = 0.5, C = 150 mM; cf. Figure 2A). In both samples only relatively small objects could be discerned, and although not readily seen in Figure 5, at the higher concentration the sample was again too dense to allow for an accurate interpretation. However, upon 10-fold dilution of the sample no aggregation of the disks occurred, and no band-like structures appeared, in contrast to the observations for the corresponding q = 1.5 bicelle sample (Figures 4B and 5B). After incubation at 37 °C for 24 h, the diluted sample with q = 0.5 was examined again (Figure 5C). No obvious change in sample morphology was observed, contrary to the observations made for the q = 1.5 bicelle mixture, indicating a much less concentration- and time-

Figure 4. Cryo-TEM images of q = 1.5 DMPC/DHPC bicelle mixtures taken at 37 °C. (A) shows the sample with 150 mM total lipid, (B) 15 mM total lipid, and (C) the same sample as in B but after incubation at 37 °C for 24 h. The arrows indicate a flat band-like structure in (A), a cluster of aggregating disks in (B), and a distorted band-like structure in (C). In (B) and (C), two representative images from the same sample are shown in order to better demonstrate the polydisperse morphology.

dependent lipid morphology than what was observed for q = 1.5 bicelles.



DISCUSSION Small fast-tumbling bicelles that do not form aligned phases in the magnetic field have for the past 15 years or so been used successfully to investigate structure and dynamics of membrane-associated peptides and protein fragments. More recent studies have focused on investigating the lipid properties in bicelles and in particular the effect of peptides and protein fragments on lipid dynamics.25,28 Other studies have focused on investigating the association of peptides with bicelles by e.g. pulsed field gradient diffusion NMR methods. For a better understanding of such data and better design of bicelle related studies, the morphology of the bicelle mixture should be known. A simple change in lipid concentration may have effects on the sample, and in this study we see clear evidence for this. First, bicelles are larger at 150 mM than 300 mM, regardless of q value, and this holds for all temperatures. This is understandable if one takes the amount of free DHPC into 5492

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

total amount of DHPC is 30 mM and the total amount of DHPC in the q = 0.5 samples is 50 mM. These concentrations are still well above the critical micelle concentration of DHPC. The results for DHPC micelles (Figure 1) clearly indicate that stable micelles are formed at 30 mM DHPC and, moreover, that the size of the micelles is not temperature dependent. From this we can conclude that the formation of smaller particles is not a consequence of a too low total DHPC concentration, but rather a consequence of the combination of overall lipid concentration and q value. The temperature-dependent increase in bicelle size stems from the phase transition of the DMPC lipids. In the ordered gel phase (below 23 °C) the hydrocarbon chains are closely packed, while they are fluid and loosely packed in the liquid crystalline phase. Hence, a temperature dependence in particle size indicates a bilayered structure of the DMPC, while a higher degree of mixing between DHPC and DMPC may prevent the gel phase formation at low temperatures, and hence no temperature-dependent size is observed. Therefore, the lack of a temperature-dependent size of the particles formed in DMPC/DHPC mixtures with q = 0.5 and q = 1 at the lowest concentration, 75 mM, strongly suggests that these are not bicelles with a clear segregated bilayered DMPC region, but more reminiscent of DMPC−DHPC mixed micelles where DMPC and DHPC are more evenly distributed than in the classical bicelle (Figure 6). It is worth noting that at high q

Figure 6. Cartoon models of a “classical” bicelle (A) and DMPC− DHPC mixed micelle (B). DMPC is indicated with the headgroup in blue while DHPC is indicated with the headgroup in gray.

Figure 5. Cryo-TEM images of q = 0.5 DMPC/DHPC bicelle mixtures taken at 37 °C. (A) shows the sample with 150 mM total lipid, (B) 15 mM total lipid, and (C) the same sample as in (B) but after incubation at 37 °C for 24 h.

value (1.5) also the 75 mM sample shows a temperaturedependent size distribution. It therefore appears that at high q values the small mixed bicelles do not form. Indeed, this sample has similar characteristics to those of the other two concentrations, indicating that using a higher lipid-to-detergent ratio prevents the formation of the smaller bicelles (or mixed micelles). A previous study by Glover et al.43 that focused on q = 0.5 bicelle mixtures showed concentration-dependent shifts of DMPC and DHPC 31P resonances within a concentration range of 3−89 mM, but not above 126 mM where a bilayer-like structure was shown to exist in the bicelles. The lowest concentration used in the present work, 75 mM, falls into the category where the bicelle structure may not prevail. Together with the present q = 1.5 data at this concentration, one can propose that there is a critical bicelle concentration above which DMPC/DHPC mixtures form bicelles while below this concentration particles that are smaller than the expected bicelles exist. One explanation for the smaller particle size is a higher degree of mixing between DHPC and DMPC, which would lead to a more mixed-micelle-like morphology. With

account. In this case, the effective q value, qeff = [DMPC]/ ([DHPC] − [DHPC]free), is larger for the sample with 150 mM than for 300 mM if we assume that [DHPC]free is a constant for a certain temperature. Therefore, this result is expected but demonstrates that even at temperatures much higher than the gel to liquid crystalline transition temperature, the size of the bicelles depends not only on temperature but also on the total lipid concentration. Second, by lowering the total lipid concentration, we observe that the q ≥ 1 samples become more polymorphic, more sensitive to temperature and less stable over time. The samples made with a q = 0.5 bicelle mixture appear to be less sensitive to these variations. Third, as the concentration was decreased as low as 75 mM, a temperature dependence of the size was only observed for the q = 1.5 sample, but not for the q = 1 or q = 0.5 sample. Repeated experiments with new samples gave the same results. In the q = 1.5 samples with a total concentration of 75 mM, the 5493

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

have conclusively demonstrated that for low q value mixtures particles with apparent disk shapes are prevalent. A SANS study of bicelle mixtures with q ranging from 0.2 to 1.0 revealed the presence of more or less idealized disk-shaped structures.54 Another study using a range of techniques including NMR and electron microscopy revealed unambiguously that above a certain threshold concentration the disk shape prevailed.43 On the basis of our present results and earlier work, we therefore conclude that the largest population of objects are disks but that the overall sample morphology depends on the concentration and q values. The presence of even a few large aggregates may have important implications for interpreting the effect of peptides on lipid properties, and the peptides may also take part in the aggregation process. In several cases it has been observed that the NMR spectra of certain peptides in bicelle solution have been of poor quality, possibly because of hindered motion of both lipid and peptide molecules. Part of the explanation for such observations may also in fact be related to formation of large lipid aggregates. We see here evidence for such aggregates in high q value mixtures, but they may well be triggered by the addition of other molecules such as peptides or protein fragments. The results demonstrate that the morphology of the bicelle mixtures must be considered when bicelles are applied for studies of protein or peptide−lipid interactions where the definition of the lipid system is crucial. Only samples with low q values and concentrations above the critical bilayer-forming concentration are suitable in order to correctly interpret studies of lipid properties.

higher q values this critical concentration most likely decreases. Hence, at q values where isotropic fast-tumbling particles exist, high total lipid concentrations have to be used to ensure that the “classical” bilayered bicelle morphology prevails. In the q = 0.5 DMPC/DHPC bicelles, DLS measurements reveal the presence of small particles with a rather uniform radius ranging between around 3 nm (at 300 mM total lipid concentration) to around 4 nm (at 150 mM total concentration) over the temperature range 18−40 °C. The size of the particles observed in the cryo-TEM images of such mixtures (Rh of around 3.7 nm) agrees remarkably well with the DLS results. Although it is only the samples with high concentration (150−300 mM) that appear to contain bicelles, the sample with low concentration is still homogeneous and DLS data indicate a uniform size of the micelle-like objects. This is in excellent agreement with the study by van Dam et al. which clearly demonstrated the presence of small micelle-like particles at q = 0.5.42 Contrary to the results for q = 0.5 bicelles, we observed a complicated phase behavior at high temperature for q = 1 and q = 1.5 bicelles, and the combined results reveal the presence of at least two different species in the sample. For 150 mM q = 1 DMPC/DHPC mixtures, these species have apparent hydrodynamic radii of 2.5 nm (the small particles) and 7 nm (the large particles), and although these dimensions are evaluated under the assumption that they are spheres, this is again in agreement with what is observed for q = 1.5 bicelles in the cryoTEM images. Again, the small particles are most likely mixed, DHPC-rich micelles, as previously found for low concentration mixtures, while the larger ones are most likely disk-like bicelles.34,35,42 One should also note that even at 150 mM total lipid concentration, and q = 1, the DHPC concentration is still as high as 75 mM, well over the cmc for DHPC. Hence, a phase separation cannot be due to a too low total DHPC concentration. It is worth noting that the concentrations used here for DLS measurements are much higher than in previous studies of bicelles in the range of q = 0.5 to q = 1.5.42 The high concentrations used in this study (8% (w/v) and 16% (w/v)) are typically used in solution NMR studies of proteins and peptides, and hence it is important to note that at these concentrations bicelle mixtures with q > 0.5 display evidence of polymorphism at a temperature above the DMPC transition temperature. This highlights the importance of lipid concentration for maintaining the bicelle structure in high q value samples. In agreement with the study by van Dam et al.,42 the temperature dependence of bicelle size decreases with decreasing q (Figure 2). Other studies have also revealed that bicelle morphology is sensitive to concentration and or q value.36,43 Bicelles have also been prepared with other detergents with lower cmc values, which allows for lower total concentrations.50−53 Also in these cases evidence of complex morphologies has been observed. The cryo-TEM images also show that the morphology of DMPC/DHPC mixtures with high q values change over time, and large aggregates begin to appear, while mixtures with low q values remain uniform and no changes in particle size can be detected. We have here demonstrated that q = 0.5 mixtures have a uniform morphology that does not vary with time, but the question concerning whether the particles observed here are true disks still remains. Although our TEM images of low q value samples indicate the presence of small particles it is hard to distinguish the true geometry. Previous studies, however,



AUTHOR INFORMATION

Corresponding Author

*Ph +46 8 162448; Fax +46 8 155597; e-mail lena.maler@dbb. su.se (L.M.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Niklas Hedin for assistance with operating the Malvern dynamic light scattering equipment and Katarina Edwards for valuable discussion and access to the cryo-TEM equipment. This work was supported by grants from the Swedish Research Council, The Carl Trygger Foundation, and the Magnus Bergvall Foundation.



ABBREVIATIONS NMR, nuclear magnetic resonance; DLS, dynamic light scattering; DHPC, 1,2-dihexanoyl-sn-glycero-3-phosphocholine; DMPC, 1,2-dimyristoyl-sn-glycero-3-phosphocholine; TEM, transmission electron microscopy; cmc, critical micelle concentration.



REFERENCES

(1) Ram, P.; Prestegard, J. Magnetic Field Induced Ordering of Bile Salt/Phospholipid Micelles: New Media for NMR Structural Investigations. Biochim. Biophys. Acta 1988, 940, 289−294. (2) Sanders, C. R.; Prestegard, J. Magnetically Orientable Phospholipid Bilayers Containing Small Amounts of a Bile Salt Analogue, CHAPSO. Biophys. J. 1990, 58, 447−460. (3) Sanders, C. R.; Schwonek, J. P. Characterization of Magnetically Orientable Bilayers in Mixtures of Dihexanoylphosphatidylcholine and Dimyristoylphosphatidylcholine by Solid-State NMR. Biochemistry 1992, 31, 8898−8905.

5494

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

(4) Sanders, C. R., II.; Hare, B. J.; Howard, K. P.; Prestegard, J. H. Magnetically-Oriented Phospholipid Micelles as a Tool for the Study of Membrane-Associated Molecules. Prog. Nucl. Magn. Reson. Spectrosc. 1994, 26, 421−444. (5) Sanders, C. R.; Prosser, R. S. Bicelles: A Model Membrane System for all Seasons? Structure 1998, 6, 1227−1234. (6) Pointer-Keenan, C. D.; Lee, D. K.; Hallok, K.; Tan, A.; Zand, R.; Ramamoorthy, A. Investigation of the Interaction of Myelin Basic Protein with Phospholipid Bilayers using Solid-State NMR Spectroscopy. Chem. Phys. Lipids 2004, 132, 47−54. (7) Ramamoorthy, A.; Thennarasu, S.; Tan, A.; Lee, D. K.; Clayberger, C.; Krensky, A. M. Cell Selectivity Correlates with Membrane-Specific Interactions: A Case Study on the Antimicrobial Peptide G15 Derived from Granulysin. Biochim. Biophys. Acta 2006, 1758, 154−163. (8) Ouellet, M.; Bernard, G.; Voyer, N.; Auger, M. Insights on the Interactions of Synthetic Amphipathic Peptides with Model Membranes as Revealed by 31P and 2H Solid-State NMR and Infrared Spectroscopies. Biophys. J. 2006, 90, 4071−4084. (9) Ramamoorthy, A.; Thennarasu, S.; Lee, D. K.; Tan, A.; Maloy, L. Solid-State NMR Investigation of the Membrane-Disrupting Mechanism of Antimicrobial Peptides MSI-78 and MSI-594 Derived from Magainin 2 and Melittin. Biophys. J. 2006, 91, 206−216. (10) Brender, J. R.; Dürr, U. H. N.; Heyl, D.; Budarapu, M. B.; Ramamoorthy, A. Membrane Fragmentation by an Amyloidogenic Fragment of Human Islet Amyloid Polypeptide Detected by SolidState NMR Spectroscopy of Membrane Nanotubes. Biochim. Biophys. Acta 2007, 1768, 2026−2029. (11) De Angelis, A. A.; Howell, S. C.; Nevzorov, A. A.; Opella, S. J. Structure Determination of a Membrane Protein with Two TransMembrane Helices in Aligned Phospholipid Bicelles by Solid-State NMR Spectroscopy. J. Am. Chem. Soc. 2006, 128, 12256−12267. (12) Sanders, C. R.; Landis, G. C. Reconstitution of Membrane Proteins into Lipid-Rich Bilayered Mixed Micelles for NMR Studies. Biochemistry 1995, 34, 4030−4040. (13) Booth, P. J.; Farooq, A.; Flitsch, S. L. Retinal Binding during Folding and Assembly of the Membrane Protein Bacteriorhodopsin. Biochemistry 1996, 35, 5902−5909. (14) Booth, P. J.; Riley, M. L.; Flitsch, S. L.; Templer, R. H.; Farooq, A.; Curran, A. R.; Chadborn, N.; Wright, P. Evidence that Bilayer Bending Rigidity Affects Membrane Protein Folding. Biochemistry 1997, 36, 197−203. (15) Faham, S.; Bowie, J. U. Bicelle Crystallization: A New Method for Crystallizing Membrane Proteins Yields a Monomeric Bacteriorhodopsin Structure. J. Mol. Biol. 2002, 316, 1−6. (16) Marcotte, I.; Auger, M. Bicelles as Model Membranes for Solidand Solution-State NMR Studies of Membrane Peptides and Proteins. Concepts Magn. Reson. 2005, 24, 17−37. (17) Matsumori, N.; Morooka, A.; Murata, M. Detailed Description of the Conformation and Location of Membrane-Bound Erythromycin a Using Isotropic Bicelles. J. Med. Chem. 2006, 49, 3501−3508. (18) Prosser, R. S.; Evanics, F.; Kitevski, J. L.; Al-Abdul-Wahid, M. S. Current Applications of Bicelles in NMR Studies of MembraneAssociated Amphiphiles and Proteins. Biochemistry 2006, 45, 8453− 8465. (19) Houdai, T.; Matsumori, N.; Murata, M. Structure of MembraneBound Amphidinol 3 in Isotropic Small Bicelles. Org. Lett. 2008, 10, 4191−4194. (20) Andersson, A.; Mäler, L. NMR Solution Structure and Dynamics of Motilin in Isotropic Phospholipid Bicellar Solution. J. Biomol. NMR 2002, 24, 103−112. (21) Bárány-Wallje, E.; Andersson, A.; Gräslund, A.; Mäler, L. NMR Solution Structure and Position of Transportan in Neutral Phospholipid Bicelles. FEBS Lett. 2004, 567, 265−269. (22) Biverståhl, H.; Andersson, A.; Gräslund, A.; Mäler, L. NMR Solution Structure and Membrane Interaction Studies of the NTerminal Sequence (1−30) of the Bovine Prion Protein. Biochemistry 2004, 43, 14940−14947.

(23) Bárány-Wallje, E.; Andersson, A.; Gräslund, A.; Mäler, L. Dynamics of Transportan in Bicelles Is Surface Charge Dependent. J. Biomol. NMR 2006, 35, 137−147. (24) Lind, J.; Gräslund, A.; Mäler, L. Membrane Interactions of Dynorphins. Biochemistry 2006, 45, 15931−15940. (25) Andersson, A.; Biverstahl, H.; Nordin, J.; Danielsson, J.; Lindahl, E.; Mäler, L. The Membrane-Induced Structure of Melittin is Correlated with the Fluidity of the Lipids. Biochim. Biophys. Acta 2007, 1768, 115−121. (26) Vold, R. R.; Prosser, R. S. Magnetically Oriented Phospholipid Bilayered Micelles for Structural Studies of Polypeptides. does the Ideal Bicelle Exist? J. Magn. Reson. B 1996, 113, 267−271. (27) Vold, R. R.; Prosser, R. S.; Deese, A. J. Isotropic Solutions of Phospholipid Bicelles: A New Membrane Mimetic for HighResolution NMR Studies of Polypeptides. J. Biomol. NMR 1997, 9, 329−335. (28) Lind, J.; Nordin, J.; Mäler, L. Lipid Dynamics in Fast-Tumbling Bicelles with Varying Bilayer Thickness: Effect of Model Transmembrane Peptides. Biochim. Biophys. Acta 2008, 1778, 2526−2534. (29) 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, 328−339. (30) Andersson, A.; Mäler, L. Motilin-Bicelle Interactions: Membrane Position and Translational Diffusion. FEBS Lett. 2003, 545, 139−143. (31) Wang, J.; Schnell, J. R.; Chou, J. J. Amantadine Partition and Localization in Phospholipid Membrane: A Solution NMR Study. Biochem. Biophys. Res. Commun. 2004, 324, 212−217. (32) Andersson, A.; Almqvist, J.; Hagn, F.; Mäler, L. Diffusion and Dynamics of Penetratin in Different Membrane Mimicking Media. Biochim. Biophys. Acta 2004, 1661, 18−25. (33) Biverståhl, H.; Lind, J.; Bodor, A.; Mäler, L. Biophysical Studies of the Membrane Location of the Voltage-Gated Sensors in the HsapBK and KvAP K+ Channels. Biochim. Biophys. Acta 2009, 1788, 1976−1986. (34) Nieh, M.; Raghunathan, V.; Glinka, C. J.; Harroun, T. A.; Pabst, G.; Katsaras, J. Magnetically Alignable Phase of Phospholipid “Bicelle” Mixtures Is a Chiral Nematic Made Up of Wormlike Micelles. Langmuir 2004, 20, 7893−7897. (35) Triba, M. N.; Warschawski, D. E.; Devaux, P. F. Reinvestigation by Phosphorus NMR of Lipid Distribution in Bicelles. Biophys. J. 2005, 88, 1887−1901. (36) van Dam, L.; Karlsson, G.; Edwards, K. Morphology of Magnetically Aligning DMPC/DHPC Aggregates Perforated Sheets, Not Disks. Langmuir 2006, 22, 3280−3285. (37) Katsaras, J.; Harroun, T. A.; Pencer, J.; Nieh, M. P. “Bicellar” Lipid Mixtures as Used in Biochemical and Biophysical Studies. Naturwissenschaften 2005, 92, 355−366. (38) Angelis, A. D.; Jones, D.; Grant, C.; Park, S.; Mesleh, M.; Opella, S. NMR Experiments on Aligned Samples of Membrane Proteins. Methods Enzymol. 2005, 394, 350−382. (39) De Angelis, A. A.; Opella, S. J. Bicelle Samples for Solid-State NMR of Membrane Proteins. Nat. Protoc. 2007, 2, 2332−2338. (40) Pabst, G.; Kučerka, N.; Nieh, M.; Rheinstädter, M.; Katsaras, J. Applications of Neutron and X-Ray Scattering to the Study of Biologically Relevant Model Membranes. Chem. Phys. Lipids 2010, 163, 460−479. (41) Chou, J. J.; Baber, J. L.; Bax, A. Characterization of Phospholipid Mixed Micelles by Translational Diffusion. J. Biomol. NMR 2004, 29, 299−308. (42) 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 2004, 1664, 241−256. (43) Glover, K. J.; Whiles, J. A.; Wu, G.; Yu, N.; Deems, R.; Struppe, J. O.; Stark, R. E.; Komives, E. A.; Vold, R. R. Structural Evaluation of Phospholipid Bicelles for Solution-State Studies of MembraneAssociated Biomolecules. Biophys. J. 2001, 81, 2163−2171. 5495

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496

Langmuir

Article

(44) Andersson, A.; Mäler, L. Magnetic Resonance Investigations of Lipid Motion in Isotropic Bicelles. Langmuir 2005, 21, 7702−7709. (45) Prosser, R. S.; Evanics, F.; Kitevski, J. L.; Al-Abdul-Wahid, M. S. Current Applications of Bicelles in NMR Studies of MembraneAssociated Amphiphiles and Proteins. Biochemistry 2006, 45, 8453− 8465. (46) Whiles, J. A.; Brasseur, R.; Glover, K. J.; Melacini, G.; Komives, E. A.; Vold, R. R. Orientation and Effects of Mastoparan X on Phospholipid Bicelles. Biophys. J. 2001, 80, 280−293. (47) Chou, J. J.; Kaufman, J. D.; Stahl, S. J.; Wingfield, P. T.; Bax, A. Micelle-Induced Curvature in a Water-Insoluble HIV-1 Env Peptide Revealed by NMR Dipolar Coupling Measurement in Stretched Polyacrylamide Gel. J. Am. Chem. Soc. 2002, 124, 2450−2451. (48) Mäler, L.; Gräslund, A. NMR Studies of Three-Dimensional Structure and Positioning of CPPs in Membrane Model Systems. Methods Mol. Biol. 2011, 683, 57−67. (49) Almgren, M.; Edwards, K.; Karlsson, G. Cryo Transmission Electron Microscopy of Liposomes and Related Structures. Colloids Surf., A 2000, 174, 3−21. (50) Lu, Z.; Van Horn, W. D.; Chen, J.; Mathew, S.; Zent, R.; Sanders, C. R. Bicelles at Low Concentration. Mol. Pharmaceutics 2012, 9, 752−761. (51) Wu, H.; Su, K.; Guan, X.; Sublette, M. E.; Stark, R. E. Assessing the Size, Stability, and Utility of Isotropically Tumbling Bicelle Systems for Structural Biology. Biochim. Biophys. Acta 2010, 1798, 482−488. (52) Jiang, Y.; Wang, H.; Kindt, J. T. Atomistic Simulations of Bicelle Mixtures. Biophys. J. 2010, 98, 2895−2903. (53) Li, M.; Morales, H. H.; Katsaras, J.; Kučerka, N.; Yang, Y.; Macdonald, P. M.; Nieh, M. P. Morphological Characterization of DMPC/CHAPSO Bicellar Mixtures: A Combined SANS and NMR Study. Langmuir 2013, 29, 15943−15957. (54) Luchette, P. A.; Vetman, T. N.; Prosser, R. S.; Hancock, R. E. W.; Nieh, M.-P.; Glinka, C. J.; Krueger, S.; Katsaras, J. Morphology of Fast-Tumbling Bicelles: A Small Angle Neutron Scattering and NMR Study. Biochim. Biophys. Acta 2001, 1513, 83−94.

5496

dx.doi.org/10.1021/la500231z | Langmuir 2014, 30, 5488−5496