Subscriber access provided by READING UNIV
Article
Formation of lipid-bilayer nanodiscs by diisobutylene/maleic acid (DIBMA) copolymer Abraham Olusegun Oluwole, Johannes Klingler, Bartholomäus Danielczak, Jonathan Oyebamiji Babalola, Carolyn Vargas, Georg Pabst, and Sandro Keller Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.7b03742 • Publication Date (Web): 21 Nov 2017 Downloaded from http://pubs.acs.org on November 24, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
Formation of lipid-bilayer nanodiscs by diisobutylene/maleic acid (DIBMA) copolymer Abraham Olusegun Oluwole,†‡ Johannes Klingler,† Bartholomäus Danielczak,† Jonathan Oyebamiji Babalola,‡ Carolyn Vargas,† Georg Pabst,§$ and Sandro Keller†* †Molecular Biophysics, University of Kaiserslautern, Erwin-Schrödinger-Str. 13, 67663 Kaiserslautern (Germany) ‡Department of Chemistry, University of Ibadan, 200284 Ibadan (Nigeria) § University of Graz, Institute of Molecular Biosciences, Biophysics Division, NAWI Graz, Humboldtstr. 50/III, 8010 Graz (Austria) $ BioTechMed-Graz, 8010 Graz (Austria)
ABSTRACT: Membrane proteins usually need to be extracted from their native environment and separated from other membrane components for in-depth in vitro characterization. The use of styrene/maleic acid (SMA) copolymers to solubilize membrane proteins and their surrounding lipids into bilayer nanodiscs is an attractive approach toward this goal. We have recently shown that diisobutylene/maleic acid (DIBMA) copolymer similarly solubilizes model and cellular membranes but, unlike SMA(3:1), has a mild impact on lipid acyl-chain order and thermotropic phase behavior. Here, we used fluorescence spectroscopy, small-angle X-ray scattering, size-exclusion chromatography, dynamic light scattering, and 31P nuclear magnetic resonance spectroscopy to examine the self-association of DIBMA and its membrane-solubilization properties against lipids differing in acyl-chain length and saturation. Although DIBMA is less hydrophobic than commonly used SMA(3:1) and SMA(2:1) copolymers, it efficiently formed lipid-bilayer nanodiscs that decreased in size with increasing polymer/lipid ratio while maintaining the overall thickness of the membrane. DIBMA fractions of different molar masses were similarly efficient in solubilizing a saturated lipid. Coulomb screening at elevated ionic strength or a reduced charge density on the polymer at low pH enhanced the solubilization efficiency of DIBMA. The free-energy penalty for transferring phospholipids from vesicular bilayers into nanodiscs became more unfavorable with increasing acyl-chain length and unsaturation. Altogether, these findings provide a rational framework for using DIBMA in membrane-protein research by shedding light on the effects of polymer and lipid properties as well as experimental conditions on membrane solubilization.
INTRODUCTION Investigations on integral membrane proteins often require isolation of the target protein and subsequent incorporation into a well-defined membrane-mimetic system to confer solubility and stability in aqueous solutions. Although detergent micelles are traditionally used to achieve these objectives, micellar assemblies do not mimic some of the fundamental properties—such as the bilayer architecture and the lateral pressure profile—of native membranes, thus leading to protein denaturation.1 Therefore, the major challenges in handling membrane proteins in vitro lie not only in optimizing protein solubilization and purification but also in preserving native interactions within the target protein and between the protein and its surrounding lipids.2 Overcoming these challenges is far from being trivial because of the structural3 and functional4 diversity of membrane proteins and their lipid environments. A remarkable advancement in the development of membrane-mimetic systems was the advent of lipid-bilayer nanodiscs, which initially were assembled with the aid of membrane scaffold proteins (MSPs)5 and, more recently, using styrene/maleic acid (SMA) copolymers.6–11 In contrast with detergent micelles, these nanodiscs provide embedded proteins with a lipid-bilayer environment, thus promoting native interactions and stability.7 Whereas assembling lipids and proteins into MSP nanodiscs still requires conventional detergents,5 SMA copolymers directly recruit membrane proteins along with their surrounding lipids from either
natural or artificial membranes to form nanodiscs called SMA/lipid particles (SMALPs).8 Unlike comparatively large proteoliposomes or immobilized membrane models such as supported lipid bilayers, membrane proteins incorporated into SMALPs are amenable to in vitro methods requiring nanometer-sized particles.7,8 However, biophysical studies have shown that the SMA scaffold can alter the order and dynamics of the lipid core,9 which in turn can influence the functions of embedded proteins.10 Moreover, SMA is incompatible with far-UV optical spectroscopy and enzyme assays that require divalent cations.7,8,11 We have recently addressed these limitations by introducing diisobutylene/maleic acid (DIBMA) copolymer,12 which bears aliphatic rather than aromatic hydrophobic pendant groups (Chart 1). DIBMA solubilizes model and cell membranes into nanodiscs termed DIBMA/lipid particles (DIBMALPs) that (i) are directly amenable to far-UV spectroscopy without chromatographic purification, (ii) have a mild impact on lipid acyl-chain order and dynamics, and (iii) are compatible with fairly high concentrations of divalent cations.12 DIBMA is commercially available under the trade name Sokalan CP 9 (BASF, Ludwigshafen, Germany) and is a component of several formulations ranging from adhesives to detergents, paper coating, and sizing agents.13 The polymer is synthesized by free-radical polymerization of maleic anhydride and 2,4,4-trimethyl-1-pentene in the presence of 2,2’-azobis(2-methylpropionitrile).13 In this reaction, a growing polymer chain with a maleic
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
anhydride radical terminus cannot homopolymerize but reacts virtually exclusively with the olefin comonomer, such that an equimolar mixture of the comonomers favors formation of strictly alternating polymer chains.14–16 This affords a considerable level of compositional homogeneity across all polymer chains irrespective of their lengths. Hence, DIBMA can serve as an ideal model compound for studying the thermodynamics, kinetics, and mechanisms of membrane solubilization by amphiphilic copolymers. We have recently shown12 that DIBMA imposes less perturbation than SMA to the acyl-chain order of solubilized phospholipids, thus keeping their thermotropic phase behavior comparable to that of vesicular lipid preparations. Qualitatively, this can be attributed to the mild impact of its branched aliphatic pendant group as opposed to the more disruptive action of the aromatic moieties of SMA. For a quantitative-level understanding of the differences among amphiphilic copolymers containing different hydrophobic building blocks, it is essential to systematically study the self-association and membrane-solubilization properties of DIBMA in comparison with SMA.12 Ionic strength and pH are particularly important factors: while the former modulates Coulomb interactions among polymer chains as well as their hydrophobicity, the latter affects their charge density and, therefore, also their hydrophobicity. Hence, in amphiphilic, polyelectrolytic copolymers such as DIBMA and SMA, all of these factors act together in determining the conformation and dynamics of polymer chains in aqueous solution17 and their membrane affinity.18 In this report, we addressed the self-association properties of DIBMA by monitoring the steady-state fluorescence of Nile Red (NR; 9-diethylamino-5H-benzo[a]phenoxazin-5-one) and compared them with SMA(3:1), SMA(2:1), and SMA(1:1) copolymers. We exploited small-angle X-ray scattering (SAXS) to study the overall morphology of DIBMALPs at different polymer/lipid ratios and used fractionation by size-exclusion chromatography (SEC) to compare polymer fractions of different chain lengths with regard to their solubilization efficiency with the aid of dynamic light scattering (DLS). To systematically assess the membrane-solubilization thermodynamics of DIBMA, we extended a 31 P nuclear magnetic resonance (NMR) approach previously applied to 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)12 to other lipids differing in acyl-chain length and saturation, namely, 1,2-dilauroyl-sn-glycero-3-phosphocholine (DLPC), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), and (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC). In doing so, we paid particular attention to the roles of ionic strength and pH in the solubilization of lipid vesicles, as these experimental parameters can be used to tune the membrane interactions of DIBMA. The results may serve as guideposts for proteinsolubilization protocols and foster the development of novel amphiphilic polymers for membrane-protein research. EXPERIMENTAL 2.1 Materials. DIBMA (Sokalan CP 9, Mn = 8.4 kg mol–1, Mw = 15.3 kg mol–1) and SMA with a styrene/maleic acid molar ratio of 1:1 (SMA(1:1); SMA1000H, Mn = 2.0 kg mol–1, Mw = 5.5 kg mol–1) were kind gifts from BASF (Ludwigshafen, Germany) and Cray Valley (Exton, USA), respectively. SMA copolymers with styrene/maleic acid molar ratios of 2:1 (SMA(2:1); Xiran SZ30010, Mn = 2.7 kg mol–1, Mw = 7.0 kg mol– 1 ) and 3:1 (SMA(3:1); Xiran SL25010, Mn = 4.0 kg mol–1, Mw = 10.0 kg mol–1) were kind gifts from Polyscope (Geleen, Netherlands). DLPC, DMPC, DPPC, and POPC were kind gifts from Lipoid (Ludwigshafen, Germany). 9-(diethylamino)-5Hbenzo[a]phenoxazin-5-one (NR) was purchased from Arcos Organics (Geel, Belgium), ethanol (99.8% GC-grade) from Sigma– Aldrich (Steinheim, Germany), NaCl from VWR (Darmstadt,
Page 2 of 17
Germany), bis-(2-hydroxyethyl)imino-tris-(hydroxymethyl)methane (Bis–Tris) from AppliChem (Darmstadt, Germany), and tris(hydroxymethyl)aminomethane (Tris) from Carl Roth (Karlsruhe, Germany). All chemicals were purchased in the highest purity available. 2.2 Preparation of DIBMA and SMA stock solutions. For NR fluorescence experiments, DIBMA, SMA(1:1), SMA(2:1), or SMA(3:1) was precipitated from ~5 mL of the respective commercial solution by addition of 4 M HCl, washed 4–5 times with water, and lyophilized for at least 24 h in an Alpha 2-4 LSCplus freeze–dryer (Martin Christ, Osterode am Harz, Germany). The resulting white powders were used to prepare 10 g L–1 stock solutions in 40 mM Britton–Robinson buffer (BR buffer; 40 mM boric acid, 40 mM phosphoric acid, 40 mM acetic acid, pH 8.0). For solubilization experiments, ~3 mL of DIBMA commercial solution was dialyzed against 0.5–1.0 L of the desired buffer (i.e., 50 mM Tris, pH 7.4 or 8.3; 50 mM Bis–Tris, pH 6.5) using a 5-mL QuixSep dialyzer (Membrane Filtration Products, Seguin, USA) and Spectra/Por 3 dialysis membranes (Spectrum Laboratories, Rancho Dominguez, USA) with a nominal molar-mass cutoff of 3.5 kg mol–1. Dialysis was carried out at room temperature for 24 h with buffer exchange after 12–16 h. Recovered polymers were filtered through 0.22-µm poly(vinylidene fluoride) filters (Carl Roth). Stock concentrations of DIBMA were determined by refractive index measurements on an Abbemat 500 (Anton Paar, Graz, Austria) using a refractive index increment of dn/dc = 1.347 M–1 based on Mn.13 2.3 NR fluorescence. A stock solution of 2.6 mM NR was prepared in ethanol and diluted to 20 µM in BR buffer, pH 8.0 prior to use. For polymer concentration dependences, 1-mL samples containing 1 µM NR and 0–8 g L–1 DIBMA, SMA(1:1), SMA(2:1), or SMA(3:1) were prepared in BR buffer, pH 8.0. For pH dependences, 50 µL of 20 µM NR and 100 µL of 10 g L–1 polymer were added to 850 µM BR buffer, pH 4–9. BR buffer was chosen for these experiments to enable comparison with a previous study18 on SMA copolymers. Each sample was equilibrated for 16 h at 20°C and gently vortexed prior to measurements in a 10 mm × 10 mm cuvette at 20°C. Steady-state fluorescence emission spectra were acquired on a FluoTime300 spectrometer (PicoQuant, Berlin, Germany) equipped with a laser source emitting at 467 nm. This relatively long excitation wavelength selectively populates the excited state of those NR molecules that are localized in hydrophobic domains.19 Emission spectra were recorded in the range of 500–750 nm using a magic-angle configuration of excitation and emission polarizers, an emission bandwidth of 5 nm, a spectral resolution of 5 nm, and an integration time of 1 s. Critical aggregation concentrations (CACs) were derived from the inflection points of sigmoidal fits to the wavelengths of maximum emission plotted versus polymer concentration (see Figure 1b below), as detailed elsewhere.18 2.4 Preparation of lipid vesicles. 50 mM stock suspensions of DLPC, DMPC, DPPC, and POPC were prepared by weighing in the required amounts on a high-precision XP Delta Range microbalance (Mettler Toledo, Greifensee, Switzerland) and suspending each lipid powder in the desired buffer. The suspension was agitated for 15 min followed by 35-fold extrusion through two stacked polycarbonate filters with a nominal pore diameter of 100 nm using a Mini Extruder (Avanti, Alabaster, USA). To ensure that all lipid bilayers were in the fluid (i.e., liquid–crystalline) state, DLPC and POPC were suspended and extruded at 20°C, whereas DMPC and DPPC were extruded at 30°C and 45°C, respectively. Dynamic light scattering (DLS, see below) showed that the resulting large unilamellar vesicles (LUVs) had z-average diameters and polydispersity indices (PDIs) typically amounting to, respectively, (140 ±10) nm and 0.04±0.01 for DLPC, (150 ±10) nm and 0.08±0.02 for DMPC, (160 ±10) nm
ACS Paragon Plus Environment
Page 3 of 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
and 0.10±0.02 for DPPC, and (150 ±10) nm and 0.03±0.01 for POPC. 2.5 Small-angle X-ray scattering (SAXS). Experiments were performed at the P12 SAXS beamline at DESY (Hamburg, Germany),20 which is equipped with a Pilatus 2M detector (Dectris, Baden, Switzerland). An automated sample robot was used to deliver 20–35 µL of DIBMALP samples containing 6.78 g L–1 (10 mM) DMPC and either 7.56 g L–1 (0.90 mM) or 18.5 g L–1 (2.2 mM) DIBMA from a multi-well plate into a preheated glass capillary. Each sample was equilibrated at 30°C for 10 min prior to acquisition of 20 frames with an exposure time of 0.045 s/frame using an X-ray wavelength of 0.06 nm and a sample-to-detector distance of 3.1 m. Scattering data were first checked for radiation damage and then corrected for contributions from the capillary tube and buffer. Pair distance distribution functions (PDDFs), were derived from scattering patterns by indirect Fourier transformation using the software GNOM.21 2.6 Size-exclusion chromatography (SEC). DIBMA was fractionated by SEC on an Äkta Purifier 10 chromatography system (GE Healthcare, Freiburg, Germany) equipped with a HiPrep 16/60 Sephacryl S-500 HR column (GE Healthcare). For each run, the column was equilibrated at 6–8°C and at a steady flow rate of 0.5 mL min−1 using buffer (25 mM Tris, pH 8.5) supplemented with 20% (v/v) EtOH in order to prevent hydrophobic collapse of the polymer. Then, 1 mL of a buffered solution containing 7.50 g L–1 DIBMA was injected at a flow rate of 0.5 mL min−1. Fractions collected from a total of 9 runs were pooled in such a way as to sample the entire elution peak in a roughly uniform manner (see Figure 3b below). Pooled fractions were concentrated and washed using Amicon Ultra 15-mL centrifugal filters (Merck–Millipore, Billerica, USA), lyophilized, and resuspended in buffer (50 mM Tris, 200 mM NaCl, pH7.4). 2.7 Dynamic light scattering (DLS). For DMPC, samples were prepared by adding 0–1.32 g L–1 SEC-fractionated DIBMA to 1.36 g L–1 lipid in the form of LUVs. For DLPC, DPPC, and POPC, increasing concentrations of unfractionated DIBMA were added to LUV suspensions containing 2.0 mM lipid. All samples were incubated for 16 h above the gel-to-fluid phase transition temperature of the respective lipid, namely, at 20°C for DLPC and POPC, 30°C for DMPC, and 45°C for DPPC. DLS measurements were carried out in 3 mm × 3 mm quartz glass cuvettes on a Nano Zetasizer ZS90 (Malvern Instruments) operating at 633 nm with 90° detection. Particle size distributions were obtained from autocorrelation functions using instrument default settings. z-average particle diameters and PDIs were obtained from cumulant analysis of autocorrelation functions.22 Distribution widths of z-average particle diameters, σ, were calculated as σ = √PDI z.23 2.8 31P nuclear magnetic resonance (NMR) spectroscopy. Samples containing 2.5–10 mM lipid in the form of LUVs were treated with increasing concentrations of DIBMA. 10% (v/v) D2O was included in the sample buffer as lock signal. Since the buffer capacity of Tris is limited to pH 7–9, Bis-Tris buffer was used for solubilization experiments at pH 6.5. Samples were incubated for >16 h before measurements on an Avance 400 spectrometer (Bruker Biospin, Rheinstetten, Germany) operating at a 31P resonance frequency of 162 MHz. 256 scans were acquired per sample using a 5-mm broadband inverse probe, an inverse-gated protondecoupling pulse sequence, an acquisition time of 1.6 s, a sweep width of 9746 Hz, and a relaxation delay of 6 s. Chemical shifts were referenced to 85% (w/v) H3PO4 in D2O as external standard at 0 ppm. Data were multiplied by an exponential function with a line-broadening factor of 1 Hz before Fourier transformation. Peaks were integrated using the software Bruker Topspin 3.2. Detailed procedures for deriving polymer/lipid phase diagrams from 31P NMR data have been reported elsewhere.24–26 Briefly, the
NMR peak area is proportional to the extent of solubilization; thus, the appearance of an isotropic peak marks the onset of solubilization (SAT), while constant peak areas at higher polymer concentrations reflect the completion of solubilization (SOL). Accordingly, a nonlinear least-square fit27 to the concentration dependence of the peak area yields the best-fit values and 95% confidence intervals of the saturating and solubilizing polymer and c concentrations, c S S , respectively, at a given lipid con centration, cL. cS and c are linear functions of cL: S aq,o
b,SAT cSAT cL S =cS +RS
(1)
aq,o m,SOL cL cSOL S =c +RS b,SAT where RS and Rm,SOL are S
(2)
the maximum and minimum polymer/lipid molar ratios required for the onset and completion aq,o of solubilization, respectively. The ordinate intercept, cS , corresponds to the concentration of “free” polymer chains in the aqueaq,o ous solution; for all polymers studied here, cS = 0. Finally, the changes in Gibbs free energy accompanying the b→m,o vesicle-to-DIBMALP transfer of polymer and lipid, ∆GS and ∆Gb→m,o , respectively, were obtained as: L b→m,o
=–RT ln
m,SOL b,SAT (1+RS ) b,SAT m,SOL (1+RS ) RS
b→m,o
=–RT ln
1+RS
∆GS
∆GL
RS
b,SAT
1+RS
m,SOL
(3) (4)
RESULTS AND DISCUSSION 3.1 Hydrophobicity and self-association of DIBMA as compared with SMA copolymers. NR is a hydrophobic probe that experiences a drastic increase in fluorescence intensity and a blue shift of its emission spectrum when it enters a nonpolar environment.19,28 Specifically, its emission spectrum is weak in intensity and centered at ~650 nm in polar environments but becomes much more intense and blue-shifts to ~615 nm in nonpolar environments. NR is therefore a useful probe for monitoring the hydrophobicity and hydrophobicity-dependent self-association behavior of the copolymers under study.18 For this purpose, we recorded steady-state fluorescence spectra of 1 µM NR in the presence of increasing concentrations of DIBMA (Figure 1a), SMA(1:1), SMA(2:1), or SMA(3:1). At high polymer concentrations of 2.5 g L–1, the fluorescence of NR was strongest in the presence of SMA(3:1), weaker in SMA(2:1), and even fainter in SMA(1:1) and DIBMA (Figure 1a, inset). In the presence of low polymer concentrations, the emission spectrum of NR was centered at >660 nm in all cases. This is typical of an aqueous environment, thus indicating that all polymer chains assumed an extended conformation without forming hydrophobic domains. In the presence of higher polymer concentrations, however, the wavelength of maximum emission significantly decreased to ~620 nm, which is typical of NR in a nonpolar environment (Figure 1b). Accordingly, SMA(3:1) and SMA(2:1) self-associated to form polymeric micelles at low CACs of ~10 mg L–1, which is readily explained by the pronounced hydrophobicity of styrene groups.29 In the presence of the alternating copolymers SMA(1:1) and DIBMA, which have a lower molar content of hydrophobic comonomer units of 50%, hydrophobic domains appeared only at much higher polymer concentrations of ~1 g L–1. Like SMA copolymers,18 DIBMA is expected to undergo a transition from an extended state at alkaline and neutral pH values to a collapsed state at lower pH. The fluorescence intensity of NR in aqueous solutions containing either 2.5 g L–1 DIBMA or SMA(1:1) increased upon lowering the pH (Figure 1c,d), indicating that the polymer molecules increasingly collapsed from an extended conformation at higher pH to a more compact confor-
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
mation at lower pH.17 By contrast, the fluorescence intensity of NR in the presence of SMA(2:1) or SMA(3:1) remained high throughout almost the entire pH range studied (Figure 1d), indicative of a collapsed state. The only exception to this observation was a sharp decrease in intensity at low pH values, where both SMA(2:1) and SMA(3:1) formed insoluble aggregates (Figure 1d), which is in agreement with earlier observations.18 In summary, under the same pH conditions, DIBMA is more hydrophilic and, at neutral and alkaline pH values, adopts a more extended conformation than SMA(2:1) and SMA(3:1). 3.2 Formation of bilayer-containing nanodiscs of tunable size. We have previously shown12 by combining electron microscopy, differential scanning calorimetry, and Raman spectroscopy that the DIBMA-mediated solubilization of lipid vesicles results in the formation of disc-shaped nanoparticles that retain a lipidbilayer architecture. Moreover, both DLS and 31P NMR have demonstrated that the hydrodynamic size of the nanodiscs depends on the polymer/lipid ratio in the sample. That is, once solubilization is complete, the particle size continuously decreases with increasing polymer/lipid ratio. Here, we extended these studies by subjecting DMPC DIBMALPs to structural analysis by SAXS. Most importantly, scattering patterns typical of nano-sized lipid-bilayer discs30,31 were observed for DIBMALPs at two different polymer/lipid ratios (Figure 2a). The secondary maxima at scattering vectors of q ≈ 1.4 nm–1 indicated a bilayered inner structure of the discs, thus demonstrating that the overall membrane architecture was preserved at both polymer concentrations. Unbiased structural insights can be obtained from contrastweighted distance histograms in the form of PDDFs (Figure 2b). The minima in these PDDFs are due to the large electron density contrast between the nonpolar hydrocarbon layers and the polar lipid headgroups and, thus, provide additional and independent evidence for an intact bilayer structure of the discs. The maxima at ~5.9 nm and ~4.9 nm at DIBMA/DMPC molar ratios of 0.09 and 0.22, respectively, are a measure of the total bilayer thickness in the discs. These values compare reasonably well with the steric bilayer thickness of DMPC vesicles measured at the same temperature,31 although they most likely also contain contributions from polymer chains adsorbed onto the headgroup regions of the membrane. Similar thicknesses have been reported for nanodiscs surrounded by SMA(2:1)33 or MSP.30 The radii of gyration derived from the PDDFs further allowed us to estimate the diameters of the nanodiscs to amount to ~40 nm and ~20 nm at DIBMA/DMPC molar ratios of 0.09 and 0.22, respectively. These values are in good agreement with those derived from hydrodynamic particle diameters obtained from DLS.12 Conversion of the latter into steric nanodisc diameters34,35 yields values of ~35 nm and ~20 nm at DIBMA/DMPC molar ratios of 0.09 and 0.22, respectively. Taken together, all of these observations demonstrate that the size of DIBMALPs can be tuned by adjusting the polymer/lipid ratio without compromising the bilayer organization of the solubilized membrane patch. 3.3 Influence of polymer chain length on solubilization efficiency. Since SMA and DIBMA polymers display a relatively broad distribution of molar masses rather than a discrete value, we wondered if certain chain lengths are particularly efficient in solubilizing lipid membranes. Hence, we separated DIBMA according to hydrodynamic size with the aid of SEC and collected 10 pooled fractions covering the entire elution peak (Figure 3a,b). The solubilization efficiency of each fraction was then tested on fluid-phase DMPC LUVs at 30°C and monitored by DLS (Figure 3c). All fractions were able to solubilize DMPC LUVs, with particle size distributions revealing an initial increase in hydrodynamic diameter to >1000 nm followed by a smooth decrease to ≤50 nm with increasing DIBMA/DMPC ratios. All but the two shortest-chain fractions showed very similar solubilization effi-
Page 4 of 17
ciencies (Figure 3c). The shortest DIBMA chains were somewhat less efficient, but even then DIBMA/DMPC mass ratios >1.2 still resulted in small particles having hydrodynamic diameters of ~50 nm. It should be noted that these differences in solubilization efficiencies among polymer fractions of various chain lengths have no significant effects on the thermodynamic parameters retrieved from solubilization experiments performed on unfractionated DIBMA samples,12 as reflected in the Gibbs free energies of transfer calculated using equations 3 and 4. This is because the overall solubilization efficiency of a mixture of surfactants that on their own have different efficiencies is dominated by the most efficient component unless the latter is outnumbered.36 For a twocomponent surfactant mixture that shows additive solubilization behavior, for example, this relationship is embodied in the following equation:35a
=
+
(5)
Here, , , and + are the saturating concentrations of, respectively, pure surfactant 1, pure surfactant 2, and a mixture of the two with molar fractions and . For a back-ofthe-envelope calculation, let us assume that 20% of all DIBMA chains have a solubilization efficiency that is only half that of the remaining 80%, as expressed by = 0.8, = 0.2, and = 2 . Then, equation 5 demonstrates that the overall saturating concentration of unfractionated DIBMA is only 11% larger than that of the more efficient fraction that accounts for the vast majority of polymer chains. Analogous considerations apply to the concentration required for complete solubilization. For the DMPC values determined before using unfractionated DIBMA12 (Table 1), equations 3 and 4 predict that an 11% change in RSb,SAT and RSm,SOL results in Gibbs free energies of transfer that differ by only ~0.008 kJ mol–1 from the best-fit values for unfractionated DIBMA. The same arguments are readily extended to more complex surfactant mixtures comprising more than two components,36 including continuous distributions of polymers. Hence, the membrane-solubilization behavior of an unfractionated polymer is dominated by the most efficient fraction if the latter makes up the major part of the total distribution. In the case of DIBMA, it thus becomes clear that the somewhat reduced efficiencies of the two shortest-chain fractions cannot significantly distort the values determined using unfractionated polymer samples. In conclusion, these results demonstrate that all DIBMA fractions are solubilization-competent and, furthermore, that quantitative differences in solubilization efficiency are so minor that, at least for the present purposes, they do not warrant any attempts at purifying a particular polymer fraction that would possess enhanced solubilization properties. 3.4 Effect of lipid acyl-chain length on solubilization efficiency. By treating the different supramolecular aggregates present in polymer/lipid mixtures as pseudophases,37 the free-energy changes accompanying the transfer of polymers and lipids from vesicular bilayers into nanodiscs can be determined. This simple yet highly informative thermodynamic conceptualization of polymer-mediated solubilization enables a systematic comparison of different surfactant/lipid systems.24 Moreover, it has provided insights into the influence of various lipid properties on membrane solubilization by SMA(3:1)25 and SMA(2:1).26 We have previously reported on the DIBMA-driven solubilization of vesicles composed of DMPC, which contains two saturated C14 acyl chains (di-14:0 PC).12 Here, we further explored the solubilizing activity of DIBMA using the shorter-chain phospholipid DLPC (di-12:0 PC) and the longer-chain variant DPPC (di-16:0 PC) in the fluid state, that is, above their respective main phase transition temperatures.
ACS Paragon Plus Environment
Page 5 of 17
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Langmuir
We first monitored the solubilization of DLPC LUVs at 25°C with the aid of 31P NMR spectroscopy. In the absence of DIBMA, the resonance of DLPC in the form of LUVs was broadened beyond detection (Figure 4a) because of the slow tumbling of the vesicles,38 which had a z-average diameter of ~140 nm. Upon addition of sufficient amounts of DIBMA, however, an isotropic peak emerged that reflected small, fast-tumbling particles.24 Plotting the peak area as a function of the polymer concentration at each lipid concentration revealed two breakpoints (Figure 4b) marking the onset (SAT) and completion (SOL) of solubilization. Plotting the DIBMA concentrations at these breakpoints against the corresponding DLPC concentrations further yielded a phase diagram with saturating and solubilizing DIBMA/DLPC molar = 0.025 ± 0.005 and Rm,SOL = 0.050 ± 0.002, ratios of Rb,SAT S S respectively (Figure 4c). From these molar ratios, we calculated vesicle-to-DIBMALP transfer free energies of ∆Gb→m,o = (0.058 ± 0.02) kJ mol–1 and L b→m,o
∆GS = –(1.62 ± 0.56) kJ mol–1 for DLPC and DIBMA, respectively. We further evaluated the particle sizes of DIBMALPs assembled from the different saturated phospholipids (Figure 4d– f). As for DMPC,12 aggregates of ~1000 nm in diameter were observed under subsolubilizing conditions for DLPC (Figure 4d). Upon addition of more DIBMA, smaller particles started to form (Figure 4e) that progressively decreased in size with increasing polymer/lipid ratio (Figure 4e,f). We likewise established the phase diagram for the DIBMA/DPPC system and obtained saturating and solubilizing = 0.042 ± 0.006 and polymer/lipid molar ratios of Rb,SAT S Rm,SOL = 0.082 ± 0.005 at 45°C, respectively (Figure 5a–c). These S values translate to vesicle-to-DIBMALP transfer free energies of ∆Gb→m,o = (0.101 ± 0.02) kJ mol–1 L
b→m,o ∆GS
and
= –(1.70 ± 0.23) kJ mol for DPPC and DIBMA, respectively. DLS revealed the presence of large aggregates at DIBMA/DPPC molar ratios below Rb,SAT (Figure 5d), but a furS ther increase in the polymer/lipid ratio again resulted in progressively decreasing particle sizes (Figure 5e,f). The hydrophobic thicknesses of fluid-phase bilayers consisting of DLPC, DMPC, or DPPC are 21.9 Å, 25.7 Å, and 28.5 Å, respectively.36 Our data (Table 1) show that increasing bilayer thickness resulted in an increase in both Rb,SAT and Rm,SOL ; in S S other words, more polymer was required to initiate and complete the solubilization of thicker bilayers as compared with thinner ones. This trend was exclusively due to an increase in the freeenergy penalty associated with transferring the phospholipids from vesicular bilayers into DIBMALPs, with an increment of –1
~10 J mol–1 for each pair of methylene groups in the two saturated acyl chains. By contrast, the favorable free-energy change that DIBMA experiences upon its transfer from vesicles into nanodiscs was, within experimental error, unaffected by bilayer thickness (Table 1). At their respective Rm,SOL values, DIBMALPs had S z-average diameters and associated size-distribution widths of (25 ± 10) nm for DLPC, (39 ± 14) nm for DMPC, and (43 ± 24) nm for DPPC. Further addition of polymer to reach a polymer/lipid molar ratio of 0.25 reduced these values to (12 ± 6) nm, (15 ± 10) nm, and (28 ± 14) nm, respectively. It is noteworthy that the 31P NMR peaks had similar widths for DLPC (Figure 4a), DMPC,12 and DPPC (Figure 5a) even though DLS evidenced an increase in hydrodynamic DIBMALP size with acyl-chain length. This apparent discrepancy is readily explained by the temperature dependence of the tumbling rate and, thus, the NMR peak width. Hence, the opposing effects resulting from an increase in hydrodynamic size and from a raise in experimental temperature with chain length virtually cancelled one another. In
summary, DIBMALPs increase in size with lipid acyl-chain length and, thus, bilayer thickness. In contrast with MSP nanodiscs,5,31 the size of DIBMALPs seems to be dictated to a large extent by the lipids rather than the polymer scaffold. 3.5 Solubilization of unsaturated phospholipids. Biological membranes usually contain considerable amounts of unsaturated phospholipids, which can significantly affect the solubilization efficiency and thermodynamics of polymers.25 Therefore, we examined the solubilization properties of DIBMA toward the monounsaturated phospholipid POPC (16:0/18:1 PC). This lipid forms bilayer membranes that have a hydrophobic thickness of 29.2 Å and, owing to the cis C–C double bond in the sn-2 acyl chain, exhibit increased lateral pressure in their hydrophobic core.39 31P NMR spectroscopy (Figure 6a,b) furnished a phase diagram characterized by Rb,SAT = 0.129 ± 0.015 and S m,SOL = 0.195 ± 0.006 (Figure 6c), which corresponds to freeRS energy
changes
b→m,o ∆GS
of
∆Gb→m,o = (0.139 ± 0.05) kJ mol–1 L
and
= –(0.89 ± 0.46) kJ mol–1 for POPC and DIBMA, respectively. These results show that, under the same experimental conditions as used above, much higher DIBMA concentrations were required to initiate and complete the solubilization of POPC vesicles as compared with any of the saturated lipids tested. This was because, in the case of POPC, the vesicle-to-DIBMALP transfer free energy was more unfavorable for the lipid and less favorable for DIBMA (Table 2). We also examined the roles of ionic strength and pH on the solubilization of POPC vesicles. First, we raised the NaCl concentration in the buffer from 200 mM to 500 mM while keeping the pH at a value of 7.4. Titration of POPC LUVs with DIBMA in the presence of 500 mM NaCl gave rise to 31P NMR peaks that were substantially narrower (Figure 6d) than at 200 mM NaCl (Figure 5a), thus suggesting smaller particle sizes at elevated ionic strength. Systematic solubilization titrations (Figure 6e) resulted in a phase diagram characterized by Rb,SAT = 0.077 ± 0.007 and S Rm,SOL = 0.132 ± 0.013 (Figure 6f) and vesicle-to-DIBMALP S = (0.123 ± 0.03) kJ mol–1 and transfer free energies of ∆Gb→m,o L b→m,o
∆GS = –(1.21 ± 0.34) kJ mol–1 (Table 2). Thus, high ionic strength afforded enhanced solubilization, as characterized by lower Rb,SAT and Rm,SOL boundaries and less endergonic and more S S b→m,o exergonic values of ∆Gb→m,o and ∆GS , respectively. L We further checked the effect of decreasing the charge density on the polymer by lowering the pH to 6.5 in the presence of 200 mM NaCl. Titration of POPC vesicles with the polymer again produced narrower 31P NMR peaks at pH 6.5 (Figure 6g) than at pH 7.4 (Figure 6a). Titrations at various POPC concentrations (Figure 6h) and the resulting phase diagram revealed lower Rb,SAT S and Rm,SOL values of 0.121 ± 0.007 and 0.169 ± 0.012, respectiveS ly (Figure 6i). With values of ∆Gb→m,o = (0.104 ± 0.04) kJ mol–1 L b→m,o
and ∆GS = –(0.72 ± 0.28) kJ mol–1 , the corresponding transfer free energies were the least endergonic for the lipid but also the least exergonic for the polymer compared with the other conditions tested for POPC. We used DLS to further scrutinize the effects of ionic strength and pH on the solubilization of POPC vesicles and the size of the resulting DIBMALPs. When monitoring the effect of 0–500 mM NaCl at pH 7.4, we found that no solubilization occurred without addition of NaCl (Figure 7a). Solubilization in the presence of 100 mM or 200 mM NaCl was similarly efficient, starting at a DIBMA/POPC molar ratio between 0.1 and 0.2, with the nanodisc diameter decreasing to ~40 nm when a DIBMA/POPC molar ratio of 0.5 was approached. A further increase in salt concentration to ≥300 mM NaCl rendered the
ACS Paragon Plus Environment
Langmuir
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
solubilization of POPC by DIBMA more efficient. This might seem expected, as higher ionic strengths screen the Coulomb repulsion among polymer chains and enhance the hydrophobic effect, thus promoting the hydrophobicity and membrane affinity of DIBMA. Remarkably, at salt concentrations ≥300 mM NaCl, the hydrodynamic diameter of DIBMALPs was reduced to
