Lipid dynamics in nanoparticles formed by maleic acid-containing copolymers: EPR spectroscopy and molecular dynamics simulations

Journal Pre-proof Lipid dynamics in nanoparticles formed by maleic acidcontaining copolymers: EPR spectroscopy and molecular dynamics simulations

Alexandr Colbasevici, Natalia Voskoboynikova, Philipp S. Orekhov, Marine E. Bozdaganyan, Maria G. Karlova, Olga S. Sokolova, Johann P. Klare, Armen Y. Mulkidjanian, Konstantin V. Shaitan, Heinz-Jürgen Steinhoff PII:

S0005-2736(20)30033-X

DOI:

https://doi.org/10.1016/j.bbamem.2020.183207

Reference:

BBAMEM 183207

To appear in:

BBA - Biomembranes

Received date:

19 November 2019

Revised date:

22 January 2020

Accepted date:

23 January 2020

Please cite this article as: A. Colbasevici, N. Voskoboynikova, P.S. Orekhov, et al., Lipid dynamics in nanoparticles formed by maleic acid-containing copolymers: EPR spectroscopy and molecular dynamics simulations, BBA - Biomembranes(2020), https://doi.org/10.1016/j.bbamem.2020.183207

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier.

Journal Pre-proof

Lipid Dynamics in Nanoparticles Formed by Maleic Acid-Containing Copolymers: EPR Spectroscopy and Molecular Dynamics Simulations Alexandr Colbasevicia, Natalia Voskoboynikovaa, Philipp S. Orekhovb,c,d, Marine E.

of

Bozdaganyanb,d,e, Maria G. Karlovad, Olga S. Sokolovad, Johann P. Klarea, Armen Y. Mulkidjaniana,f,g, Konstantin V. Shaitand,e, Heinz-Jürgen Steinhoffa⁎

Department of Physics, Osnabrueck University, 49069 Osnabrueck, Germany

b

Moscow Institute of Physics and Technology, Dolgoprudny 141701, Russia

c

Sechenov University, Moscow 119146, Russia

d

Department of Biology, Lomonosov Moscow State University, 119991, Moscow, Russia

re

-p

ro

a

e

lP

N.N. Semenov Federal Research Center for Chemical Physics, Russian Academy of Sciences, Moscow, 119991, Russia f

A.N. Belozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State University, Moscow 119991, Russia g

Jo ur

na

Department of Bioengineering and Bioinformatics, Lomonosov Moscow State University, Moscow 119991, Russia *Corresponding author:

Department of Physics, Osnabrueck University, D-49069 Osnabrueck, Germany E-mail address: [email protected] Telephone: +49 541 969-2675

1

Journal Pre-proof

Highlights ● The dynamics of lipids in discoidal lipid bilayer particles encased by copolymers of diisobutylene/maleic acid (DIBMALPs) or styrene/maleic acid (SMALPs) was compared with that in liposomes. ● The dynamics both of spin labeled and unlabeled lipids is less constrained in DIBMALPs than in SMALPs. ● The affinity of spin labeled lipids to the polymeric rim is more pronounced in SMALPs than in DIBMALPs.

Jo ur

na

lP

re

-p

ro

of

Graphical Abstract

2

Journal Pre-proof

Abstract Amphiphilic maleic acid-containing copolymers account for a recent methodical breakthrough in the study of membrane proteins. Their application enables a detergent-free extraction of membrane proteins from lipid bilayers, yielding stable water-soluble, discoidal lipid bilayer particles with incorporated proteins, which are wrapped with copolymers. Although many studies confirm the potential of this approach for membrane protein research, the interactions between the maleic acid-containing copolymers and extracted lipids, as well as possible effects of the copolymers on lipid-embedded proteins deserve further scrutinization. Here, we combine electron paramagnetic resonance spectroscopy and coarse-grain molecular dynamics

of

simulations to compare the distribution and dynamics of lipids in lipid particles of

ro

phospholipid bilayers encased either by an aliphatic diisobutylene/maleic acid copolymer (DIBMALPs) or by an aromatic styrene/maleic acid copolymer (SMALPs). Nitroxides

-p

located at the 5th, 12th or 16th carbon atom positions in phosphatidylcholine-based spin labels experience restrictions of their reorientational motion depending on the type of

re

encasing copolymer. The dynamics of the lipids was less constrained in DIBMALPs than in

Keywords

na

in SMALPs.

lP

SMALPs with the affinity of spin labeled lipids to the polymeric rim being more pronounced

bilayer, liposome

Jo ur

Nitroxide spin label, coarse-grained molecular dynamics, SMALP, DIBMA, phospholipid

3

Journal Pre-proof

1. Introduction The preparation of single membrane proteins or membrane protein complexes in a native-like but water-soluble environment, as required for the spectroscopic investigation, e.g., is challenging. Amphiphilic maleic acid-containing copolymers provided a significant methodical breakthrough in this field. Styrene maleic acid (SMA) copolymers with styreneto-maleic acid ratios of 2:1 and 3:1 were found to directly solubilize phospholipids and membrane proteins, both from artificial and natural bilayers, yielding discoidal SMA/lipid particles (SMALPs), (see [1-4] for the pioneering publications on their application in biology and [5-9] for recent reviews). Quite recently, one more amphiphilic alternating copolymer of

of

diisobutylene and maleic acid (DIBMA), has been shown to extract phospholipids and

ro

membrane proteins and form discoidal particles (DIBMALPs) [10, 11]. SMA- and DIBMAwrapped lipid particles are comprised of lipid or lipid/protein cores surrounded by a polymer

-p

belt, and have, depending on the preparation routine, typical diameters of 10–30 nm. Within a particular preparation, the size of particles is uniform, which renders them suitable for

re

different kinds of spectroscopy, as well as other techniques that require relatively small,

lP

homogenous water-soluble particles or single membrane proteins in their native lipid environment. In the last few years, SMALPs were also shown to be suitable protein carriers

na

for cryo-electron microscopy [12, 13].

Up to recently, the most common method of membrane protein purification for the above

Jo ur

applications was their extraction from the membrane into detergent micelles. However, detergents can affect the stability and functionality of membrane proteins, see e.g. [14, 15]. In contrast, maleic acid-containing copolymers extract integral membrane proteins from the lipid bilayer in the absence of detergents. This is their major advantage, particularly in comparison to the well-studied nanodiscs, where amphipathic helical scaffold proteins stabilize discoidal lipid patches after removal of detergents [16, 17]. Membrane proteins encapsulated in lipid particles stabilized by copolymers could be studied by different biophysical methods without loss of their structural and functional properties (see [9, 18] for a recent survey). In the case of SMALPs, the proteins remained functional, but possessed somewhat higher thermal stability, which might reflect their constrained dynamics [18, 19]. In DIBMALPs, the dynamics of lipids were less constrained, whereas the ability to extract proteins from membranes was somewhat compromised [10, 11]. Although many empirical studies indicate the great potential of maleic acid-containing copolymers for membrane protein research, the molecular mechanisms of lipid membrane 4

Journal Pre-proof solubilization by these copolymers, the interactions between the copolymer belts and solubilized lipids and their possible effect on lipid-embedded proteins remained unclear. Recent coarse-grained (CG) molecular dynamics (MD) simulations examined the interaction of SMA copolymers, which varied in their composition and electric charge, with lipid membranes, as well as the behavior of lipodiscs encased by distinct maleic acid-containing copolymers [20-22]. It was found that SMA copolymers tend to aggregate in solution into clusters, which could account for the uniform size of SMALPs, and that these clusters directly interact with the lipid bilayer [22]. Consequently, clusters of SMA copolymers pull lipid patches out of the lipid bilayer forming SMALP-like structures [22].

of

The introduction of lipid spin probes is an established method to investigate the dynamics of various parts of fatty acid chains in phospholipid bilayers using EPR spectroscopy [23-25]. In

ro

particular, nitroxides attached to carbon atoms at different positions along the hydrocarbon chain allow in-depth probing the lipid bilayer in order to characterize the segmental chain

-p

mobility [25]. Recently, EPR spectroscopy yielded insight into the properties of the lipid

re

bilayer and membrane proteins encased in nanodiscs and SMALPs [3, 26-32]. It was shown that the EPR spectral line broadening of the spin labeled PCs increased in DMPC and POPC

lP

particles in the presence of SMA polymer when compared to that in DMPC dispersion and POPC vesicles, respectively [4, 32].

na

In our study, we apply EPR spectroscopy and CG MD simulations to compare the dynamics of lipid spin labels inserted in lipid bilayer containing nanoparticles depending on different

Jo ur

polymer and lipid composition. We used a series of three phosphatidylcholine (PC)-based spin labels with the nitroxides attached to the carbon atoms at the 5th, 12th and 16th positions in the hydrocarbon chain of the host phospholipid. We tested two types of commercially available maleic acid-containing copolymers, namely an aliphatic DIBMA copolymer and aromatic SMA copolymers with 3:1 styrene-to-maleic acid ratios (hydrolyzed version of SMA3000 and XIRAN SL25010 S25). We probed two types of PC lipids, saturated 2dimyristoyl-sn-glycero-3-phospholcholine (DMPC) and monounsaturated palmitoyl-2-oleoylsn-glycero-3-phosphocholine (POPC). Liposomes prepared from these lipids with addition of PC-based nitroxide spin labels served as a reference model for comparison with the lipid nanoparticles. EPR spectroscopy and CG MD simulations show that the dynamics of lipids in DIBMALPs appears to be less constrained than in SMALPs. Furthermore, the MD simulations provide evidence and explanation for the significant interaction of the nitroxides with the copolymer belt of SMALPs rendering the spin labels very sensitive for lipid-polymer interaction. 5

Journal Pre-proof

2. Materials and methods 2.1.

Chemicals

2-Dimyristoyl-sn-glycero-3-phosphocholine phosphocholine

(POPC),

(DMPC),

1-palmitoyl-2-oleoyl-sn-glycero-3-

1-palmitoyl-2-stearoyl-(5-doxyl)-sn-glycero-3-phosphocholine

(16:0-5 doxyl PC), 1-palmitoyl-2-stearoyl-(12-doxyl)-sn-glycero-3-phosphocholine (16:0-12 doxyl PC), 1-palmitoyl-2-stearoyl-(16-doxyl)-sn-glycero-3-phosphocholine (16:0-16 doxyl PC) were purchased from Sigma-Aldrich (St. Louis, MO). SMA copolymer with styrene-tomaleic acid molar ratio of 3:1 (MW 9500 Da; supplied as an aqueous sodium salt solution SMA 3000 HNa) was kindly provided as a gift by Cray Valley (Exton; PA; USA).

of

Polysciences (Netherlands) kindly provided an SMA product under the trade name XIRAN

ro

SL25010 S25 as a gift. DIBMA copolymer commercially available under the trade name

2.2.

Liposome preparations

re

analytical grade.

-p

Sokalan CP9 was kindly provided as a gift by BASF, Germany. All other reagents were of

lP

Powdered DMPC and POPC were dissolved in chloroform and mixed with 1 mol% of lipid spin label. Chloroform was evaporated under a stream of nitrogen gas. The resulting lipid film was dried under vacuum for at least 2 hours. The dried lipids were suspended in 10 mM Tris

na

(pH 8.0), 150 mM NaCl buffer (Buffer A) and vortexed. Subsequently, the multilamellar liposome suspension underwent ten freeze–thaw cycles, and, if not used directly, was stored

Jo ur

in aliquots at −80°C. Before reconstitution, the suspension of liposomes was extruded at least 11x through polycarbonate membranes (Whatman) of 400 nm or 200 nm pore size using a Mini-Extruder Set (Avanti Polar Lipids, Alabaster, AL, USA). 2.3.Preparation of lipid bilayer containing nanoparticles To form SMALPs or XIRAN/lipid particles (XIRAN-LPs), 1 ml of a 2.5% (w/v) solution of SMA or XIRAN, which were extensively dialyzed against Buffer A before, was added dropwise to 1 ml of the liposome suspension to get a final lipid-to-copolymer ratio of 1:2.5 (w/w). Similarly, 1 ml of a 5% (w/v) solution of DIBMA was added dropwise to 1 ml of the liposome suspension to get a final lipid-to-copolymer ratio of 1:5 (w/w). The assembly mixture was allowed to equilibrate for 1 hour at room temperature and afterwards for 16 h at 4°C. The resulting samples were centrifuged (126,000 g; 30 min; 4°C) to remove aggregates.

6

Journal Pre-proof 2.4. Dynamic light scattering Dynamic light scattering (DLS) measurements were performed on a Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK) at 550 nm and 25°C. Data represent the average of three sets of 14 runs of 10 s each. The particle size distribution was obtained by using the ZETASIZER software package Ver. 7.02. under the assumption that nanoparticles were spherically shaped. 2.5. Transmission electron microscopy Copper grids (300 mesh formvar/carbon-coated) (Ted Pella, USA) were hydrophilized by glow discharge (-20 mA, 45 sec) with a Pelco EasyGlow apparatus (Ted Pella, USA). Diluted

of

samples of the nanoparticle preparations (3 µl) were placed onto the grid and incubated at RT

ro

for 30 sec. The excess of the sample was removed with filter paper. Grids were stained twice with a 1% aqueous uranyl acetate solution for 30 sec at RT and air-dried.

-p

Micrographs were acquired using an analytical transmission electron microscope Jem-2100

re

(Jeol, Japan) equipped with a 2K x 2K CCD camera Ultrascan 1000XP (Gatan, USA). The microscope was operated at 200 kV in a low dose mode, with a magnification of x40000 (2.5

lP

Å/pix) and a defocus of 0.5 - 1.9 µm.

The images of individual particles were picked from the micrographs automatically using the

na

crYOLO neural network [33]. For each polymer-lipid composition, a few hundred of micrographs were acquired and processed using the EMAN2.3 suite [34, 35]. For XIRAN-

Jo ur

POPC 17000 particles, for DIBMA-DMPC 2500 particles, and for all other samples 7000 particle images were collected. Particles were windowed to a box size of 100x100 pixels for SMA- and XIRAN-derived samples, and 200x200 pixels for DIBMA-derived samples. The set of particle images of each sample was subjected to reference-free 2-D classification that includes iterative centering, rotational and translational alignment and multivariate statistical analysis (MSA) of images. This classification allows combining similar particles together and calculating the class average - normalized sum of particle images with increased contrast. Particles were distributed in classes according to their structural features, the number of particles in different classes varies. 2.6. Electron paramagnetic resonance spectroscopy EPR measurements were performed as described previously [36]. Briefly, room temperature EPR spectra were recorded on a home-built EPR spectrometer equipped with a dielectric resonator (Bruker Biospin, Germany). Glass capillaries of 0.9 mm inner diameter were filled with sample volumes of ~20 μL at ~250 μM final spin label concentration. The magnetic field 7

Journal Pre-proof modulation amplitude was set to 0.2 mT and the MW power leveled to 2 mW. The sample temperature was stabilized by a room temperature nitrogen gas stream through the resonator and checked with a thermocouple wired to the inside of the resonator thermo jacket, with measured values of 292.5 ± 1.5 K. 2.7. EPR spectra simulations As an approximation for the reorientational motion of the lipid bound spin labels we assumed a model of anisotropic Brownian rotational diffusion with axial symmetry and orienting potential. The order parameter, S, and the rotational correlation time constant, τc, of the reorientational motion of the lipid bound nitroxides were determined from fittings of the

of

experimental EPR spectra using the software Multicomponent [37]. This program is based on the numerical solution of the stochastic Liouville equation [38, 39]. A single spectral

ro

component was found to be sufficient to reasonably fit the experimental spectra. During the

-p

fittings, the hyperfine tensor components Axx and Ayy were held fixed at 4.65 and 5.25 G, respectively, the g-tensor components were taken from [38] or [40]. The value of the

re

hyperfine tensor component Azz was allowed to vary during fittings to account for different environmental polarities and to approximate partial motional averaging due to possible fast

lP

(ps) small scale wobbling of the nitroxide as pointed out by Timofeev et al. [41]. Values for Azz in the range of 33.2 ± 0.8 G were found to provide the best fittings. The rotational

na

diffusion tensor of axial symmetry defined by R⟂ and R∥ with its average R1 = (R∥R⟂2)1/3, R∥ = Rzz, is related to the molecular frame of the nitroxide by the angle βD, which was determined

Jo ur

by preliminary best-fit results for spectra of all present preparations and ultimately set fixed to an average value of 37.8°. The value of βD for a doxyl nitroxide spin label bound to the lipid tail was reported to be 34° [42] and thus the value resulting from our simulations reasonably agrees. The effective reorientational correlation time, τc, is given by 1/6R1. Other variables, which were allowed to vary during the fittings, were the coefficient for the orienting (restoring) potential, C20, and the linewidth tensor, W. The values of the order parameter S were calculated from C20 according to the algorithm provided by Multicomponent. Error limits were determined by repeating simulations using different starting values of the fitting parameters and changing the value of βD by ±15°. 2.8. Molecular dynamics simulations Coarse-grained models of nanoparticles containing doxyl-PCs were developed within the framework of the MARTINI force field [43], which has gained broad popularity for simulations of biological membranes and polymers [44] despite certain limitations [45]. The 8

Journal Pre-proof standard MARTINI library contains the CG force field parameters for POPC and DMPC, while parameters for DIBMA and doxyl PCs were derived based on all-atom (AA) simulations according to the iterative scheme we used before [46] until satisfactory agreement between the AA and CG models was achieved (Fig. S1-S5). The AA DIBMA copolymer model consisting of 52 monomers with the 1:1 diisobutylene:maleic acid ratio was built using the OPLS AA force field [47]. The reference AA simulation was run in NPT ensemble for 100 ns (T=303.15 K maintained by the V-rescale algorithm, P = 1 bar controlled by Parrinello-Rahman barostat; the integration time step was set to 2 fs; the Verlet cutoff scheme and particle mesh Ewald (PME) were used for the nonbonded interactions with the cutoff value set to 1.2 nm). The set of CG parameters for SMA was taken from [22]. The mapping

of

scheme between all-atom and CG models is shown in Fig. 1. The selection of MARTINI CG particle types was guided by the automated parametrization toolkit for small organic

ro

molecules [48] and in analogy with the existing topologies (Fig. S6). All of the CG

Jo ur

na

lP

re

-p

parameters and structure files are available at https://hpc.mipt.ru/zhmurov/lab/.

Figure 1. The structural formulae and the acronyms of SMA and DIBMA copolymers (A), spin-labeled lipids (B) and synthetic lipids (C) used in this work. Periodically repeating chemical components of polymers are enclosed in breakers and are marked with n, x and y letters. Mapping of all-atom structures to the coarse-grained MARTINI models are shown with the dotted contours.

9

Journal Pre-proof The AA simulations of doxyl PCs were carried out using the set of parameters and the starting conformations from [49]. These simulations were run for 100 ns each and the simulation protocol suggested by the authors was utilized. We used Gromacs 2018.1 for all MD simulations [50]. For each system, energy minimization using the steepest descent algorithm was run prior to equilibration simulation in the NVT ensemble (simulation time equaled 1 μs) maintained by means of the V-rescale thermostat (T=320 K, τt=1.0 ps) during which the phosphate groups of lipids were constrained in the XY plane perpendicular to the normal of the lipid patch using a harmonic potential with kspring= 1000 kJ/mol/nm2. The production simulations were run for 1 μs in the NPT ensemble using the same thermostat and the Parrinello-Rahman barostat (time constant = 12.0 ps,

of

compressibility = 3 × 10−4 bar−1 as recommended in [51] applied isotropically). All CG simulations were performed with the standard water model and the reaction field approach

ro

was used to treat the long-range electrostatics with the relative permittivity of 15. A time step

-p

of 20 fs was used for all simulations. The Verlet pair-lists cutoff scheme was used and the neighbor list was updated every 20 steps. Periodic boundary conditions were applied in all

re

simulations.

For analysis, the nanodiscs were aligned such that their center-of-mass was moved to the

lP

origin of the coordinate axes and their principal axes were aligned with the coordinate axes at every trajectory step using a custom TCL script. In-house Python scripts exploiting

na

MDAnalysis [52] were further used for calculations of the nanodiscs thickness, densities of the nanoparticles components, and the order parameters of lipid acyl chains. We were not able

Jo ur

to accurately estimate the order parameters for doxyl-PCs because of their relocation close to the polymeric rim during the simulations and away from the planar central region of nanodiscs.

3. Results

3.1. Size and shape of the lipid bilayer containing nanoparticles Particle average sizes and shapes of the lipid nanodiscs were determined for compositions of three variants of maleic acid-containing copolymers (SMA, XIRAN and DIBMA) and two variants of lipids (DMPC and POPC) (Fig. 1A, B) using dynamic light scattering (DLS) and transmission electron microscopy (TEM). The DLS measurements of nanodisc preparations showed unimodal intensity weighted size distributions with particle average sizes in the range of about 8-10 nm for SMALPs and XIRAN-LPs, whereas the size distribution for DIBMALPs 10

Journal Pre-proof was in the range of about 10-35 nm (see exemplary graphs shown in Figs. 2, 3). The DLS data also revealed a monodisperse character for all types of assembled nanodiscs. Visualization of individual nanodiscs deposited on a solid support was performed by TEM imaging of negatively stained nanodisc preparations. Micrographs showed the individual nanodiscs of all compositions studied as monodisperse well-defined disc-shaped particles. Individual particles were clearly visible on the micrographs but did not reveal sufficient contrast for automatic size calculations based on image binarization. To overcome this issue we used an approach

-p

ro

of

from single particle analysis workflow (for details, see 2.5).

re

Figure 2. Size distribution based on DLS data for DMPC (left) and POPC (right) containing particles: SMALPs

Jo ur

na

lP

-blue, XIRAN-LPs – green, DIBMALPs – black and liposomes – red line.

Figure 3. Structural characterization of nanodiscs containing lipid fragments of DMPC (left panel) and POPC (right panel) and stabilized by amphiphilic synthetic polymer molecules: top row – SMA; middle row – XIRAN; bottom row – DIBMA. Left in each panel – images of representative classes of particles, extracted from TEM

11

Journal Pre-proof images of individual particles; right in each panel – plot of size distributions, based on TEM data. Bar size – 10 nm.

Fig. 3 shows the particle diameter distribution of each sample, which is consistent with the DLS data. The SMALPs and XIRAN-LPs diameters are between 5 and 10 nm with maxima around 7-8 nm for both DMPC and POPC. The DIBMA-LPs size distributions are wider: 1025 nm for DMPC with maximum around 14 nm, and 10-50 nm for POPC with a maximum at 25 nm. In the case of POPC, the major nanoparticle population additionally included larger structures of 30-50 nm diameter.

of

3.2. Dynamics of spin-labeled lipids in lipid nanoparticles

ro

The mobility of spin labeled lipids was determined for the same set of maleic acid-containing

-p

copolymer compositions (SMA, XIRAN and DIBMA) and lipids (DMPC and POPC), and compared to the results obtained from liposomes prepared from the same lipids. To monitor

re

the dynamics of spin-labeled lipids, EPR measurements were performed for each composition, with nitroxides bound either to the 5th, 12th or 16th carbon atoms of the host

lP

phosphatidylcholine chain, 5-, 12-, and 16-doxyl PC (in the following abbreviated 5-PC, 12PC and 16-PC) (Fig. 1B). The line shape of EPR spectra reflects the dynamics of the spin-

na

label. Unrestricted reorientational motion of a nitroxide results in three equally spaced sharp lines of similar amplitude if the reorientational correlation time is in the ps time range.

Jo ur

Increase of the correlation time and / or spatial restriction of the reorientational motion leads to an increase of both the linewidth and the apparent hyperfine splitting. Fig. 4 shows the room temperature EPR spectra of the three different spin-labeled lipids in DMPC and POPC in liposomes or encased in the different nanodiscs with corresponding simulations. For all combinations of lipids and copolymers, the spectra could be simulated by a single spectral component using the model of anisotropic Brownian motion with axial symmetry and orienting potential. In liposomes and in all types of prepared nanodiscs, the EPR spectra of 5PC and 12-PC show broadening of the lines with an increase of the splitting between the outer hyperfine peaks suggesting restricted motion of the spin labels if compared to 16-PC. The apparent hyperfine splitting is largest for 5-PC. As expected, the nitroxides located closer to the lipid head groups show the most restricted reorientational behavior compared to the samples with spin labels located deeper in the bilayer. Furthermore, the reorientational motion of the spin labeled lipids encased in nanodiscs seems to be more restricted than that of their counterparts in liposomes as obvious from the increased hyperfine splitting for 5-PC or the 12

Journal Pre-proof increased linewidths for 16-PC. The case of 12-PCs is more complex. These labels appear to be immobilized in SMALPs and XIRAN-LPs, but to be much less constrained in case of liposomes. The spectra of 12-PCs in DIBMALPs resemble the previously reported spectra of

Jo ur

na

lP

re

-p

ro

of

partly immobilized nitroxide labels, cf. e.g. to ref. [53].

Figure 4. EPR spectra of doxyl-PCs with nitroxides bound at the 5th (A), 12th (B), and 16th (C) carbon position of the hydrocarbon chain in three different types of maleic acid-containing copolymers (SMA, XIRAN, DIBMA) and liposomes (from top to bottom in each panel) recorded at room temperature for DMPC-containing (three

13

Journal Pre-proof upper panels) and POPC-containing (three bottom panels) particles. Experimental (black) and best-fit simulated

Jo ur

na

lP

re

-p

ro

of

(red) spectra were normalized to maximum amplitude.

14

Journal Pre-proof Figure 5. Order parameter, S (stroke-colored bars), and rotational correlation times, τc (full color filled bars), for 5-PC (left column, blue), 12-PC (middle column, green), and 16-PC (right column, orange) in nanodiscs and liposomes consisting of DMPC and POPC lipids.

To describe our results quantitatively, we performed spectra simulations based on a model of axial symmetric Brownian reorientation diffusion [42, 54]. This model is justified by the shape of the lipids and their interaction within the membrane, which favors a less restricted reorientational motion around their long axis compared to the reorientation perpendicular to that. The inclination of the nitroxide axes with respect to the main lipid diffusion axes is accounted for by a diffusion tilt angle, βD. The rotational correlation time constant, τc, is thus a measure of the motion of nitroxides around the two symmetry axes. The segmental order

of

parameter, S, describes the motional restriction of the bound nitroxides, which results from

ro

the restoring potential acting on the lipids in the membrane.

The bar plots in Fig. 5 depict values of the order parameter, S, and the rotational correlation

-p

time constant, τc, for the three acyl chain positions in preparations of nanodiscs from both

re

DMPC and POPC in comparison to results obtained for liposomes. The values are given in

lP

Table 1.

A rigid crystal structure of the membrane is described by the maximum value of the order parameter, S = 1, whereas membranes in a state of total dynamic disorder are characterized by

na

the lowest possible value, S = 0. For DMPC preparations, the largest values of S, between 0.62 and 0.76, are found for 5-PC in all types of particles. Thus, the corresponding nitroxide

Jo ur

position represents the most ordered region of the bilayer in liposomes with only minor impact of the polymer on this parameter in the nanodiscs. The values of S for 12-PC range from 0.1 for liposomes to 0.5 for SMALPs and thus represent intermediate values for the set of spin label positions studied. The observed difference between the order parameter values for nanodiscs and liposomes could be explained by the interaction of the spin labeled lipids with the encasing polymers. Noticeably, DIBMALPs exhibit smaller S values than SMALPs and XIRAN-LPs indicating a weaker interaction between the polymer and the spin labeled lipid. 16-PC and thus the nitroxide at position 16 shows small order parameter values ranging from 0.02 to 0.15. These data do not reveal any significant differences between nanodiscs and liposomes. Located close to the end of the lipid tail and in the middle of the bilayer, the 16th position is less restricted compared to the 5th and 12th positions. Similar results were obtained for POPC preparations, confirming that in all preparations the lipid bilayer is in the liquid-crystalline phase, despite the higher phase transition temperature of pure DMPC (Tm ~ 296 K) compared to POPC (Tm ~ 271 K). 15

Journal Pre-proof

Table 1. Order parameter, S, (upper part) and reorientational correlation time, τ c, (lower part) obtained for doxyl PCs spin labeled at three different positions in SMALPs, XIRAN-LPs, DIBMALPs, and liposomes with POPC and DMPC.

Order parameter S 5-doxyl PC

12-doxyl PC

16-doxyl PC

0.76±0.02

0.61±0.07

0.10±0.10

XIRAN

0.76±0.02

0.67±0.07

0.06±0.06

DIBMA

0.64±0.08

0.38±0.02

0.15±0.15

Liposomes

0.62±0.09

0.29±0.13

0.12±0.15

-p

ro

SMA

of

DMPC

0.75±0.05

0.54±0.02

0.13±0.10

XIRAN

0.75±0.03

0.58±0.02

0.12±0.12

DIBMA

0.65±0.10

0.3±0.04

0.11±0.08

Liposomes

0.66±0.04

0.34±0.1

0.02±0.04

12-doxyl PC

16-doxyl PC

6.0±0.2

4.3±0.2

2.0±0.1

6.0±0.3

4.5±0.2

2.0±0.5

DIBMA

3.6±0.9

3.1±0.4

1.6±0.6

Liposomes

4.3±0.6

2.7±0.6

1.5±0.2

na

SMA

lP

re

POPC

Reorientational correlation time, τc / ns

Jo ur

5-doxyl PC

DMPC

SMA XIRAN

POPC SMA

5.9±0.3

3.5±0.1

1.7±0.1

XIRAN

6.0±0.5

3.7±0.2

1.6±0.1

DIBMA

3.3±1.0

3.0±1

1.5±0.2

Liposomes

2.1±0.3

2.1±0.5

1.1±0.1

16

Journal Pre-proof

The inspection of the reorientational correlation times, τc, reveals the slowest motion for 5-PC with values of about 6 ns for SMALP and XIRAN-LP preparations, about 3 ns for DIBMALPs and 2 ns for liposomes. SMALPs and XIRAN-LPs exhibit slower motion of spin labels compared to liposomes while DIBMALPs have values in-between. Deeper in the bilayer, at position 12, we observe the same tendency for a decreased spin-label mobility for preparations of nanodiscs compared to liposomes. Both SMALPs and XIRAN-LPs show reorientational correlation times of about 4 ns, whereas DIBMALPs reveal values of approximately 3 ns. Liposomes, in comparison to the other preparations, show slightly faster

of

motion of the nitroxides with time constant values of about 2.75 ns being closer to those of DIBMALPs. The motion of nitroxides at position 16 is the fastest with rotational correlation

ro

times of about 1.5 ns for all combinations of maleic acid-containing copolymers and lipids. As expected, the results obtained for SMALPs and XIRAN-LPs are almost identical in all

-p

preparations. Similar to the behavior of the order parameter, the correlation time values show

re

only small differences between DMPC and POPC preparations.

lP

3.3.Molecular dynamics simulations of lipid nanoparticles 3.3.1 Assembly of coarse-grained models of lipid nanoparticles

na

In order to provide detailed insights into organization and dynamics of the lipid nanoparticles at the molecular level and to allow for direct comparison with EPR data, we have built CG

Jo ur

models of SMALPs and DIBMALPs containing doxyl PCs. The results of the present electron microscopy and dynamic light scattering experiments indicate that both DIBMA and SMA form a substantial fraction of 10 nm sized polymer-lipid nanoparticles (see Figs. 2, 3). Thus, we have modeled SMALPs and DIBMALPs of similar size in order to allow their direct comparison. The parametrization of the CG models, including the development of CG topologies for three types of doxyl PCs with nitroxides attached at different positions along the acyl chain of the lipid, is described in the section 2.9. The CG models were based on allatom simulations of the doxyl PCs labeled at positions 5 (proximal to the head group), 12 (middle), and 14 (distal to the head group), which correspond to 5, 12, and 16 doxyl PCs used experimentally. The SMALPs were assembled using periodic SMA polymers with a 3:1 styrene to maleic acid ratio and uniform lengths ([SSSM]13, MW = 5.8 kDa). In our previous simulations [22], this type of periodic SMA polymer showed a behavior similar to that of statistical polymers. For DIBMALPs, periodic DIBMA polymers consisting of 52 units ([DM]26, MW = 5.9 kDa) were used. In the present simulations each of the maleic acid (MA) 17

Journal Pre-proof groups in both SMA and DIBMA polymers carried a total charge of -1 (i.e., with only one carboxyl group deprotonated per unit) according to [55]. We assembled nanodiscs starting from preformed disc-shaped patches of POPC or DMPC bilayers containing 142 lipids and 8-doxyl PCs labeled at different positions (see Fig. 6A-B). This corresponds to a disc with a diameter of approximately 10 nm and it is consistent with the typical size of nanodiscs observed in this study as well as in the previous works [1, 3, 4, 56-61]. These preformed lipid patches were surrounded by twelve polymer molecules of either SMA or DIBMA and placed in the periodic water box. This amount of polymers was chosen since it is sufficient to form a single layer of polymers around the cylindrical lipid patch of ca. 10 nm in diameter, which is able to stabilize SMA- and polymethacrylate

of

(PMAQA)-containing nanodiscs according to previous computational studies [20-22]. Twelve different types of nanodiscs were assembled: with DMPC or POPC lipids, stabilized by SMA

ro

or DIBMA, and containing doxyl PCs labeled at position 5, 12 and 14 (see Table S1 for the

-p

detailed description of the simulated systems). During the assembly simulations, the lipid head groups were constrained at their original positions while the polymer molecules could

re

move without restriction for 1 μs. During these simulations, the polymers rapidly absorbed at the preformed lipid discs adhering mainly to the region of hydrophobic lipid tails exposed to

Jo ur

na

lP

the solvent in the all of the investigated systems (see Fig. S8).

Figure 6. Coarse-grained models of DIBMA- (A) and SMA-stabilized (B) lipid particles. Phosphate, choline and maleic acid (MA) moieties are shown as orange, blue, and red spheres, respectively, styrene rings in SMA/diisobutylene in DIBMA as yellow triangles/spheres, backbones of SMA copolymers and lipids as gray and white sticks, respectively. C: Mass densities along the normal to the plane of SMALP/DIBMALP calculated for the phosphate groups of POPC lipids, MA of polymers and the spin label of 12 doxyl lipids.

18

Journal Pre-proof The assembly simulations were followed by the production runs during which all of the constraints were released and each system was simulated for additional 1 μs. In order to estimate the dimensions of the nanodiscs to calculate the radial distribution functions for components of nanodiscs and the order parameters of acyl chains, the nanodiscs in every MD frame were aligned with their principal axes of inertia such that their centers of masses were placed into the axes origin and the X, Y and Z directions matched the principal axes. All simulated nanodiscs remained stable and retained their flat round shape during the whole simulation time with the aspect ratio along the first two principal axes in a range of 1.0-1.1 and with the average thickness ranging from 5.5 nm to 5.8 nm. The fluctuations of the nanodiscs dimensions along the first two principal axes appeared to be somewhat higher for

of

DIBMALPs compared to SMALPs (0.4-0.5 nm and 0.3-0.4 nm, respectively, see Fig. S7).

ro

This might imply a relative destabilization of DIBMALPs by electrostatic repulsion of the

of maleic acid residues compared to SMA.

-p

DIBMA polymers, which bear a larger charge for a single polymer due to the larger fraction

re

3.3.2 Distribution of polymers and doxyl PCs in lipid nanoparticles The SMA polymers in SMALPs arrange into two distinct belts surrounding the lipids as

lP

indicated by two peaks for MA units in the density plots (see Fig. 6C). In contrast, in DIBMALPs the polymers form a single broader belt as indicated by the density plots for

na

maleic acid units of DIBMA polymers (Fig. 6C). Similar density plots for the phosphate groups of POPC show that lipids inside both SMALPs and DIBMALPs are oriented similar to

Jo ur

those in the native membrane and appear mainly as in a flat bilayer. In both SMALPs and DIBMALPs, spin-labeled doxyl PCs tend to relocate towards the polymer belt and adsorb at the lipid-polymer interface (see Fig. 7). However, this tendency is more pronounced for SMALPs. Also, for the 12-PCs but not for the 5-PCs or 14-PCs in SMALPs, we observed a shift of the corresponding peak of the radial distribution function towards the average position of the MA units of SMA (Fig. 7A, red curve). This finding implies that the labeled lipids in this case form tighter contact with the polymer belt and that the nitroxide moieties intercalate between the aromatic rings of SMA. In contrast, in DIBMALPs the nitroxides of all the three types of spin-labeled lipids do not penetrate into the polymer rim but remain largely concentrated at the interface between DIBMA and the lipid tails (Fig. 7B).

19

Journal Pre-proof

Figure 7. Radial distribution functions (RDFs) for doxyl PCs labeled at 5th, 12th and 14th positions as well as unlabeled POPC lipids in SMALPs (A) and DIBMALPs (B). The average positions of hydrophobic (diisobutylene, DIB, or styrene, STY) and hydrophilic (maleic acid, MA) units of polymers are indicated by

ro

of

yellow and red vertical lines, respectively.

-p

3.3.3 Order parameters of lipids

The order parameters calculated for the acyl chains of POPC and DMPC lipids decrease along

re

the acyl chains both in DIBMALPs or SMALPs and in planar bilayers, the latter used as a

lP

reference. Positions, which are closer to the polar head group, exhibit more restricted mobility compared to positions, which are more distant from the head group (see Fig. 8). At the same time, both types of lipids exhibit increased mobility in planar bilayers compared to lipids

na

encased in nanodiscs. The results revealed a slight decrease of the order parameters for both POPC and DMPC in DIBMALPs compared to SMALPs with the most pronounced difference

Jo ur

observed for the middle position in the acyl chains of POPC (see Fig. 8A).

Figure 8. Order parameters for sn-1 and sn-2 acyl chains of unlabeled POPC and DMPC lipids in SMALP, DIBMALP and planar bilayers.

20

Journal Pre-proof

4. Discussion In this work, using a combined experimental and in silico approach, we compared the mobility of lipid acyl chains in DIBMALPs and SMALPs containing DMPC or POPC. We applied EPR spectroscopy to study the behavior of spin labeled lipids with nitroxides attached to carbon atoms at different positions along the hydrocarbon chain. Furthermore, we performed CG MD simulation of SMALPs and DIBMALPs containing spin labeled PCs to address the lipid distribution within these LPs. The DLS measurements yield identical average particle sizes of 8 – 10 nm for all assembled SMALPs and XIRAN-LPs, which corresponds to the results obtained by other authors [1, 3,

of

4, 56-59]. For DMPC/DIBMALPs and POPC/DIBMALPs, the DLS data reveals average sizes of 14-16 nm and 10-40 nm, which are also in agreement with published data [10, 11].

ro

The TEM images show the individual nanoparticles of all studied compositions as welldefined disc-shaped particles consistent with previously reported TEM data [1, 11, 18, 56, 58,

-p

59]. Based on particle class analysis, the lateral sizes of SMALPs and XIRAN-LPs for both

re

saturated DMPC and monounsaturated POPC have maxima around 7-8 nm. Thus, the unsaturation of lipids seems not to influence the size of SMALPs. Interestingly, DIBMALPs

lP

prepared under comparable conditions had maxima around 14 nm for DMPC and 25 nm for POPC being larger than SMALPs. It appears that the size of DIBMALPs increased with the

colleagues [11].

na

extent of lipid unsaturation, which is in line with previous data obtained by Keller and

Jo ur

The X-band (9 GHz, 0.3 T) EPR spectroscopy of nitroxides is very sensitive to molecular motions in the nanosecond time range and resolves the reorientation dynamics of spin labels bound to different [23, 42, 53]. Our EPR experiments revealed the influence of the immediate environment on the degree of motional constraints of the spin labels in all preparations of nanodiscs. The line shapes of the EPR spectra vary between broad asymmetric lines, indicating immobilized nitroxides and three narrow lines in the case of fast isotropically moving nitroxides. For all types of nanodiscs, we observed broadening of the lines for all three nitroxide positions compared to liposomes. The nitroxides at the 5th carbon position are located close to the lipid head groups and, accordingly, 5-PCs reveal the mostly ordered nitroxides compared to 12-PCs or 16-PCs in nanodiscs and in liposomes. Interestingly, SMALPs and XIRAN-LPs show a higher degree of spin label immobilization than DIBMALPs and liposomes, most obvious in the spectra of 12-PCs. The two parameters extracted from the spectra, the reorientation correlation time, τc, and the order parameter, S, allow quantitative comparison of the dynamical restrictions of the encased 21

Journal Pre-proof lipids (Table 1, Fig. 5) with those in liposomes. In the following, values for S reported previously for liposomes will be briefly discussed in relation to our data. (Numerical values of data recorded at similar temperatures (297 ± 3K) to ours were determined from published figures, thus their approximated values are marked by ~). Our S value for 5-PC in DMPC liposomes, 0.62 ± 0.09, agrees with reported values which range from ~ 0.52 [62] to ~ 0.64 [31]. For POPC liposomes, reported S values range from ~ 0.60 [31] to ~ 0.68 [63], which again correlates with our finding of 0.66 ± 0.04. These values mark the most restricted motion of the nitroxide due to their location close to the densely packed lipid headgroups. For 12-PC in DMPC liposomes, the order parameter values reveal a less restricted motion of the nitroxide compared to 5-PC. Our value, S = 0.29 ± 0.13, agrees with reported values of

of

between ~ 0.23 [62] and ~ 0.31 (for 10-PC) [31]. Similar values were found for POPC liposomes, S = 0.34 ± 0.10 (our work), which agree with literature values ranging from ~ 0.24

ro

[30] to ~ 0.54 [31]. Close to the center of the liposome bilayer, the nitroxides of 16-PCs

-p

undergo the least restricted reorientational motion with order parameter values close to 0. The S values of 0.12 ± 0.15 (DMPC) and 0.02 ± 0.04 (POPC) found in this work are again in

re

agreement with reported values ranging from ~ 0.09 [31] to ~ 0.12 (for 14-PC) [62]for DMPC, and from ~ 0.12 [31] to ~ 0.18 for POPC [63]. This overall agreement between our

lP

data determined for liposomes and the corresponding literature values prove the present approach of data evaluation reasonable and thus applicable for the following characterization

na

of doxyl PC dynamics in nanodiscs.

For doxyl PCs in nanodiscs, we observe increased motional restrictions. This behavior is most

Jo ur

prominent for 12-PCs in SMALPs and XIRAN-LPs, as concluded from the increased order parameter and correlation time values in comparison to liposomes. Similar behavior was reported before in SMALPs for 10-PC [31], 11-PC [4] and for 12-PC [32], and the generally increased motional restriction was concluded to support the idea of a tightly packed bilayer system creating a stable structure [31]. Within error margins, order parameter and correlation time values for SMALPs and XIRAN-LPs are identical for the respective doxyl PCs (Table 1). In contrast, a smaller impact of DIBMA on the doxyl PC and lipid dynamics is observed (Fig. 5, 8). Experimental order parameter and reorientational correlation time values of doxylPCs in DIBMALPs more closely resemble those of doxyl PCs in liposomes than those found for SMALPs. Due to the larger average size of DIBMALPs compared to SMALPs (cf. Figs. 2, 3), in DIBMALPs a smaller fraction of lipid molecules is in close proximity to the polymer and thus affected directly in its dynamics. The results of our MD simulations discussed below provide an additional explanation for the different impact on the lipid and doxyl PC dynamics observed for the two copolymers. 22

Journal Pre-proof The CG MD simulations performed here are related to our previous study where we used the same approach to follow the pulling of lipid patches from DMPC bilayers by SMA copolymers, which resulted in formation of SMA-encased DMPC [22]. The morphology of these discs, with two belts of SMA copolymers stabilizing the rim of the lipid patch, was similar to the arrangements proposed earlier [5, 6] and was also observed in earlier MD simulations by Xue et al. [20] and Sahoo et al. [21]. We observed that the process was driven mainly by hydrophobic interactions, as it had been suggested earlier for the formation of SMALPs [64]. The MD simulations showed that a fraction of the styrene groups exhibits hydrophobic interactions with the lipid tails upon formation of SMALPs [22]. The results of these MD simulations indicated the influence of the delicate balance of hydrophobic and

of

hydrophilic units in SMA copolymers on their interaction with membranes. Hydrophilic groups of lipids first interact with maleic acid groups while hydrophobic tails tend to form

ro

hydrophobic interactions with diisobutylene and styrene.

-p

In the present work, we showed that in contrast to SMA, the interaction of DIBMA with lipids leads to the formation of single-belt disc structures. Both types of modeled

re

phosphatidylcholine lipids, POPC and DMPC, were less constrained in DIBMALPs than in

lP

SMALPs as confirmed by the difference in the calculated order parameters of lipids in the two types of nanodiscs (Fig. 8). In line with this observation, the simulations indicate that the spatial dimensions of DIBMALPs fluctuate to a larger extent as compared to SMALPs (Fig.

na

6), which might be due to a larger negative charge of a single DIBMA copolymer (built of

Jo ur

[DM-] units) as compared to a single SMA copolymer (built of [SSSM-] units). Furthermore, the MD simulations revealed that doxyl PCs show the tendency to relocate closer to the polymeric rim of nanodiscs (Fig. 7). This tendency was more pronounced in SMALPs than in DIBMALPs. Only the 12-PCs, in contrast to the other types of spin-labeled lipids, tend to penetrate deeper into the polymeric rim of SMALPs, with their spin labels specifically intercalating within the aromatic rings of styrene groups of SMA (Fig. 7A). The observed interactions of spin labels with polymers may be due to the formation of intermolecular hydrogen bonds, electrostatic interactions and stereochemical interactions [65]. However, it could be difficult to precisely delineate the main driving force of the observed affinity of the spin labels to the polymers due to limitations of the utilized CG model [45], which does not account explicitly for hydrogen bonds and proper distribution of partial charges (approximated by integer charges localized only on the charged CG beads), and intrinsically leads to smoother stereochemical interactions [66].

23

Journal Pre-proof These contacts become favorable most likely due to the match between the position of the 12PC nitroxide and the SMA belts (Fig. 6C). In contrast, DIBMA copolymers contain no aromatic rings complementary to the ring of the spin label. According to the results of our MD simulations, the experimentally observed significant immobilization of the 12-PCs in SMALPs could thus be explained by dwelling of the 12-PC lipids predominantly at the polymer-lipid interface with the lipids being tethered by interactions with the styrene groups of the copolymers. In contrast, the EPR spectra of 12-PCs in DIBMALPs show only partially immobilized spin labels, presumably be due to the weaker interaction of the spin label with the polymer and the different structure and position of the

of

copolymer belt, as revealed by MD simulations. We cannot fully exclude that the experimentally measured higher lipid mobility in

ro

DIBMALPs (Table 1, Fig. 4, 5) could correlate with their larger size as compared to SMALPs (Fig. 3). Indeed, the relative fraction of lipid molecules that are dynamically constrained by

-p

their proximity to the copolymer ring should be smaller in larger particles. Still, we believe

re

that the major cause of hampered mobility of spin-labeled lipids was their strong interaction with the polymer both in DIBMALPs and SMALPs as uncovered by MD simulations (Fig. 7).

lP

This interaction leads to concentration of the majority of spin-labeled lipids at the lipidpolymer interface (i.e., in the polymer rim of the nanodiscs, see Fig. 7) which, in principle,

na

should reduce the effect of particle size on the dynamics of spin-labeled lipids. In addition, the unlabeled lipids appear to experience certain immobilization not only in SMALPs, but also in

Jo ur

DIBMALPs (see Fig. 8). Still, the interplay between the size of lipodiscs, the lipid-polymer interactions, and the physico-chemical properties of lipodisc-stabilizing polymers deserves further investigations.

In conclusion, our EPR experiments and MD simulations reveal that the dynamics of spin labeled and unlabeled lipids is less constrained in DIBMALPs compared to SMALPs, rendering DIBMALPs a versatile tool for obtaining nanodiscs with more “native” lipid bilayer dynamics. Furthermore, our results evidence that the spatial distribution of lipophilic compounds, which may be present in the nanodiscs, could be modulated by tailoring the composition or functionalization of the maleic acid-containing copolymers.

Declaration of competing interest The authors of the manuscript “Lipid Dynamics in Nanoparticles Formed by Maleic Acidcontaining Copolymers: EPR Spectroscopy and Molecular Dynamics Simulations” declare that they have no conflict of interests. 24

Journal Pre-proof

Acknowledgements This work was supported by German Research Foundation (DFG, STE640/15) to H.J.S., the Ostpartnerschaftenprogramm of DAAD and RFBR grant no. 18-504-12045 to K.V.S. Electron microscopy was performed at User facility center for electron microscopy in life sciences of Lomonosov Moscow State University. The samples for TEM imaging were prepared in the electron microscopy (EM) unit of Integrated Bioimaging Facility (iBiOs) at Center of Cellular Nanoanalytics (CellNanOs), University of Osnabrück. We are very thankful to Dr. Katherina Psathaki for the support. Molecular dynamics simulations were

of

performed using facilities of FRCCP RAS (state task АААА-А19-119012990175-9).

ro

Author contributions

N.V. prepared samples of nanodiscs and characterized them. M.G.K. and O.S.S. performed

-p

the TEM experiments. A.C. performed the EPR experiments and fitted the EPR spectra. P.S.O. and M.E.B. performed molecular dynamics simulations and processed the MD

re

simulation data. A.C., N.V., P.S.O., M.E.B., M.G.K., O.S.S., J.P.K., A.Y.M., K.V.S. and H. J.

lP

S. analyzed the data, interpreted the results and wrote the draft version of the manuscript. H. J. S. conceived and directed the study and completed the final version of the manuscript. All

na

authors read, edited and approved the final version of the manuscript.

Jo ur

Data and materials availability All data needed to evaluate the conclusions in the paper are present in the main text and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Supplementary material One supplementary table and eight supplementary figures.

25

Journal Pre-proof

References [1] T.J. Knowles, R. Finka, C. Smith, Y.P. Lin, T. Dafforn, M. Overduin, Membrane proteins solubilized intact in lipid containing nanoparticles bounded by styrene maleic acid copolymer, J Am Chem Soc, 131 (2009) 7484-7485. [2] M. Jamshad, Y.P. Lin, T.J. Knowles, R.A. Parslow, C. Harris, M. Wheatley, D.R. Poyner, R.M. Bill, O.R. Thomas, M. Overduin, T.R. Dafforn, Surfactant-free purification of membrane proteins with intact native membrane environment, Biochem Soc Trans, 39 (2011) 813-818.

of

[3] M. Orwick-Rydmark, J.E. Lovett, A. Graziadei, L. Lindholm, M.R. Hicks, A. Watts, Detergent-free incorporation of a seven-transmembrane receptor protein into nanosized bilayer Lipodisq particles for functional and biophysical studies, Nano Lett, 12 (2012) 46874692.

-p

ro

[4] M.C. Orwick, P.J. Judge, J. Procek, L. Lindholm, A. Graziadei, A. Engel, G. Grobner, A. Watts, Detergent-free formation and physicochemical characterization of nanosized lipidpolymer complexes: Lipodisq, Angew Chem Int Ed Engl, 51 (2012) 4653-4657.

lP

re

[5] J.M. Dorr, S. Scheidelaar, M.C. Koorengevel, J.J. Dominguez, M. Schafer, C.A. van Walree, J.A. Killian, The styrene-maleic acid copolymer: a versatile tool in membrane research, Eur Biophys J, 45 (2016) 3-21.

na

[6] S.C. Lee, N.L. Pollock, Membrane proteins: is the future disc shaped?, Biochem Soc Trans, 44 (2016) 1011-1018.

Jo ur

[7] M. Overduin, M. Esmaili, Native Nanodiscs and the Convergence of Lipidomics, Metabolomics, Interactomics and Proteomics, Appl Sci-Basel, 9 (2019) 1230. [8] J.F. Bada Juarez, A.J. Harper, P.J. Judge, S.R. Tonge, A. Watts, From polymer chemistry to structural biology: The development of SMA and related amphipathic polymers for membrane protein extraction and solubilisation, Chem Phys Lipids, 221 (2019) 167-175. [9] M. Overduin, M. Esmaili, Structures and Interactions of Transmembrane Targets in Native Nanodiscs, SLAS Discov, (2019) 943-952. [10] A.O. Oluwole, B. Danielczak, A. Meister, J.O. Babalola, C. Vargas, S. Keller, Solubilization of Membrane Proteins into Functional Lipid-Bilayer Nanodiscs Using a Diisobutylene/Maleic Acid Copolymer, Angew Chem Int Ed Engl, 56 (2017) 1919-1924. [11] A.O. Oluwole, J. Klingler, B. Danielczak, J.O. Babalola, C. Vargas, G. Pabst, S. Keller, Formation of Lipid-Bilayer Nanodiscs by Diisobutylene/Maleic Acid (DIBMA) Copolymer, Langmuir, 33 (2017) 14378-14388.

26

Journal Pre-proof [12] S. Gulati, M. Jamshad, T.J. Knowles, K.A. Morrison, R. Downing, N. Cant, R. Collins, J.B. Koenderink, R.C. Ford, M. Overduin, I.D. Kerr, T.R. Dafforn, A.J. Rothnie, Detergentfree purification of ABC (ATP-binding-cassette) transporters, Biochemical Journal, 461 (2014) 269-278. [13] C. Sun, R.B. Gennis, Single-particle cryo-EM studies of transmembrane proteins in SMA copolymer nanodiscs, Chem Phys Lipids, 221 (2019) 114-119. [14] J.P. Klare, E. Bordignon, M. Doebber, J. Fitter, J. Kriegsmann, I. Chizhov, H.J. Steinhoff, M. Engelhard, Effects of solubilization on the structure and function of the sensory rhodopsin II/transducer complex, Journal of Molecular Biology, 356 (2006) 1207-1221.

of

[15] M. Zoonens, J. Comer, S. Masscheleyn, E. Pebay-Peyroula, C. Chipot, B. Miroux, F. Dehez, Dangerous liaisons between detergents and membrane proteins. The case of mitochondrial uncoupling protein 2, J Am Chem Soc, 135 (2013) 15174-15182.

ro

[16] T.H. Bayburt, S.G. Sligar, Self-assembly of single integral membrane proteins into soluble nanoscale phospholipid bilayers, Protein Sci, 12 (2003) 2476-2481.

re

-p

[17] I.G. Denisov, S.G. Sligar, Nanodiscs in Membrane Biochemistry and Biophysics, Chem Rev, 117 (2017) 4669-4713.

lP

[18] S.C. Lee, S. Khalid, N.L. Pollock, T.J. Knowles, K. Edler, A.J. Rothnie, R.T.T. O, T.R. Dafforn, Encapsulated membrane proteins: A simplified system for molecular simulation, Biochim Biophys Acta, 1858 (2016) 2549-2557.

Jo ur

na

[19] D.J.K. Swainsbury, S. Scheidelaar, R. van Grondelle, J.A. Killian, M.R. Jones, Bacterial Reaction Centers Purified with Styrene Maleic Acid Copolymer Retain Native Membrane Functional Properties and Display Enhanced Stability, Angew Chem Int Edit, 53 (2014) 11803-11807. [20] M. Xue, L. Cheng, I. Faustino, W. Guo, S.J. Marrink, Molecular Mechanism of Lipid Nanodisk Formation by Styrene-Maleic Acid Copolymers, Biophys J, 115 (2018) 494-502. [21] B.R. Sahoo, T. Genjo, K.C. Moharana, A. Ramamoorthy, Self-Assembly of PolymerEncased Lipid Nanodiscs and Membrane Protein Reconstitution, The Journal of Physical Chemistry B, 123 (2019) 4562-4570. [22] P.S. Orekhov, M.E. Bozdaganyan, N. Voskoboynikova, A.Y. Mulkidjanian, H.J. Steinhoff, K.V. Shaitan, Styrene/Maleic Acid Copolymers Form SMALPs by Pulling Lipid Patches out of the Lipid Bilayer, Langmuir, 35 (2019) 3748-3758. [23] W.L. Hubbell, H.M. McConnell, Molecular motion in spin-labeled phospholipids and membranes, J Am Chem Soc, 93 (1971) 314-326. [24] W.K. Subczynski, M. Raguz, J. Widomska, Studying lipid organization in biological membranes using liposomes and EPR spin labeling, Methods Mol Biol, 606 (2010) 247-269.

27

Journal Pre-proof [25] W.K. Subczynski, A. Wisniewska, J.J. Yin, J.S. Hyde, A. Kusumi, Hydrophobic barriers of lipid bilayer membranes formed by reduction of water penetration by alkyl chain unsaturation and cholesterol, Biochemistry, 33 (1994) 7670-7681. [26] I.D. Sahu, R.M. McCarrick, K.R. Troxel, R. Zhang, H.J. Smith, M.M. Dunagan, M.S. Swartz, P.V. Rajan, B.M. Kroncke, C.R. Sanders, G.A. Lorigan, DEER EPR measurements for membrane protein structures via bifunctional spin labels and lipodisq nanoparticles, Biochemistry, 52 (2013) 6627-6632.

of

[27] I. Orban-Glass, N. Voskoboynikova, K.B. Busch, D. Klose, C. Rickert, W. Mosslehy, F. Roder, V. Wilkens, J. Piehler, M. Engelhard, H.J. Steinhoff, J.P. Klare, Clustering and Dynamics of Phototransducer Signaling Domains Revealed by Site-Directed Spin Labeling Electron Paramagnetic Resonance on SRII/HtrII in Membranes and Nanodiscs, Biochemistry, 54 (2015) 349-362.

-p

ro

[28] N. Voskoboynikova, W. Mosslehy, A. Colbasevici, T.T. Ismagulova, D.V. Bagrov, A.A. Akovantseva, P.S. Timashev, A.Y. Mulkidjanian, V.N. Bagratashvili, K.V. Shaitan, M.P. Kirpichnikov, H.J. Steinhoff, Characterization of an archaeal photoreceptor/transducer complex from Natronomonas pharaonis assembled within styrene-maleic acid lipid particles, Rsc Adv, 7 (2017) 51324-51334.

lP

re

[29] W. Mosslehy, N. Voskoboynikova, A. Colbasevici, A. Ricke, D. Klose, J.P. Klare, A.Y. Mulkidjanian, H.J. Steinhoff, Conformational Dynamics of Sensory Rhodopsin II in Nanolipoprotein and Styrene-Maleic Acid Lipid Particles, Photochem Photobiol, 95 (2019) 1195-1204.

na

[30] I.D. Sahu, B.M. Kroncke, R. Zhang, M.M. Dunagan, H.J. Smith, A. Craig, R.M. McCarrick, C.R. Sanders, G.A. Lorigan, Structural investigation of the transmembrane domain of KCNE1 in proteoliposomes, Biochemistry, 53 (2014) 6392-6401.

Jo ur

[31] P. Stepien, A. Polit, A. Wisniewska-Becker, Comparative EPR studies on lipid bilayer properties in nanodiscs and liposomes, Biochim Biophys Acta, 1848 (2015) 60-66. [32] A.P. Bali, I.D. Sahu, A.F. Craig, E.E. Clark, K.M. Burridge, M.T. Dolan, C. DabneySmith, D. Konkolewicz, G.A. Lorigan, Structural characterization of styrene-maleic acid copolymer-lipid nanoparticles (SMALPs) using EPR spectroscopy, Chem Phys Lipids, 220 (2019) 6-13. [33] T. Wagner, F. Merino, M. Stabrin, T. Moriya, C. Antoni, A. Apelbaum, P. Hagel, O. Sitsel, T. Raisch, D. Prumbaum, D. Quentin, D. Roderer, S. Tacke, B. Siebolds, E. Schubert, T.R. Shaikh, P. Lill, C. Gatsogiannis, S. Raunser, SPHIRE-crYOLO is a fast and accurate fully automated particle picker for cryo-EM, Commun Biol, 2 (2019) 218. [34] G. Tang, L. Peng, P.R. Baldwin, D.S. Mann, W. Jiang, I. Rees, S.J. Ludtke, EMAN2: an extensible image processing suite for electron microscopy, J Struct Biol, 157 (2007) 38-46. [35] J.M. Bell, M. Chen, P.R. Baldwin, S.J. Ludtke, High resolution single particle refinement in EMAN2.1, Methods, 100 (2016) 25-34. 28

Journal Pre-proof [36] D. Klose, N. Voskoboynikova, I. Orban-Glass, C. Rickert, M. Engelhard, J.P. Klare, H.J. Steinhoff, Light-induced switching of HAMP domain conformation and dynamics revealed by time-resolved EPR spectroscopy, FEBS Lett, 588 (2014) 3970-3976. [37] C. Altenbach, W.L. Hubbell, University of California, Los Angeles.The program can be downloaded from: http://www.chemistry.ucla.edu/directory/hubbell-wayne-l. [38] J.H. Freed, Theory of slow tumbling ESR spectra for nitroxides, in: Spin Labeling: theory and applications, vol. 1, 1976, pp. 53-132. [39] D.E. Budil, S. Lee, S. Saxena, J.H. Freed, Nonlinear-least-squares analysis of slowmotion EPR spectra in one and two dimensions using a modified Levenberg-Marquardt algorithm, J Magn Reson Ser A, 120 (1996) 155-189.

ro

of

[40] B.J. Gaffney, Spin-label measurements in membranes, in: Methods in Enzymology, Elsevier, 1974, pp. 161-198.

-p

[41] V.P. Timofeev, B.A. Samarianov, About new universal approach to the EPR-spectra simulation of the spin-labeled macromolecules, Applied Magnetic Resonance, 4 (1993) 523539.

lP

re

[42] M. Ge, J.H. Freed, An electron spin resonance study of interactions between gramicidin A' and phosphatidylcholine bilayers, Biophys J, 65 (1993) 2106-2123.

na

[43] S.J. Marrink, H.J. Risselada, S. Yefimov, D.P. Tieleman, A.H. de Vries, The MARTINI force field: coarse grained model for biomolecular simulations, J Phys Chem B, 111 (2007) 7812-7824.

Jo ur

[44] S.J. Marrink, V. Corradi, P.C.T. Souza, H.I. Ingolfsson, D.P. Tieleman, M.S.P. Sansom, Computational Modeling of Realistic Cell Membranes, Chem Rev, 119 (2019) 6184-6226. [45] R. Alessandri, P.C.T. Souza, S. Thallmair, M.N. Melo, A.H. de Vries, S.J. Marrink, Pitfalls of the Martini Model, Journal of Chemical Theory and Computation, 15 (2019) 54485460. [46] P.S. Orekhov, E.G. Kholina, M.E. Bozdaganyan, A.M. Nesterenko, I.B. Kovalenko, M.G. Strakhovskaya, Molecular Mechanism of Uptake of Cationic Photoantimicrobial Phthalocyanine across Bacterial Membranes Revealed by Molecular Dynamics Simulations, J Phys Chem B, 122 (2018) 3711-3722. [47] G.A. Kaminski, R.A. Friesner, J. Tirado-Rives, W.L. Jorgensen, Evaluation and Reparametrization of the OPLS-AA Force Field for Proteins via Comparison with Accurate Quantum Chemical Calculations on Peptides †, The Journal of Physical Chemistry B, 105 (2001) 6474-6487. [48] T. Bereau, K. Kremer, Automated Parametrization of the Coarse-Grained Martini Force Field for Small Organic Molecules, Journal of Chemical Theory and Computation, 11 (2015) 2783-2791. 29

Journal Pre-proof [49] A. Kyrychenko, A.S. Ladokhin, Molecular dynamics simulations of depth distribution of spin-labeled phospholipids within lipid bilayer, J Phys Chem B, 117 (2013) 5875-5885. [50] M.J. Abraham, T. Murtola, R. Schulz, S. Páll, J.C. Smith, B. Hess, E. Lindahl, GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers, SoftwareX, 1 (2015) 19-25. [51] D.H. de Jong, S. Baoukina, H.I. Ingolfsson, S.J. Marrink, Martini straight: Boosting performance using a shorter cutoff and GPUs, Comput Phys Commun, 199 (2016) 1-7. [52] N. Michaud-Agrawal, E.J. Denning, T.B. Woolf, O. Beckstein, MDAnalysis: a toolkit for the analysis of molecular dynamics simulations, J Comput Chem, 32 (2011) 2319-2327.

of

[53] D. Marsh, A. Watts, R.D. Pates, R. Uhl, P.F. Knowles, M. Esmann, ESR spin-label studies of lipid-protein interactions in membranes, Biophys J, 37 (1982) 265-274.

-p

ro

[54] D.J. Schneider, and J. H. Freed, Calculating slow motional magnetic resonance spectra. A user's guide., in: J. Reuben (Ed.) Spin Labeling: Theory and Application, Plenum, New York, 1989, pp. 1–76.

lP

re

[55] S. Scheidelaar, M.C. Koorengevel, C.A. van Walree, J.J. Dominguez, J.M. Dörr, J.A. Killian, Effect of Polymer Composition and pH on Membrane Solubilization by StyreneMaleic Acid Copolymers, Biophys J, 111 (2016) 1974-1986.

na

[56] A.F. Craig, E.E. Clark, I.D. Sahu, R. Zhang, N.D. Frantz, M.S. Al-Abdul-Wahid, C. Dabney-Smith, D. Konkolewicz, G.A. Lorigan, Tuning the size of styrene-maleic acid copolymer-lipid nanoparticles (SMALPs) using RAFT polymerization for biophysical studies, Biochim Biophys Acta, 1858 (2016) 2931-2939.

Jo ur

[57] M. Jamshad, V. Grimard, I. Idini, T.J. Knowles, M.R. Dowle, N. Schofield, P. Sridhar, Y.P. Lin, R. Finka, M. Wheatley, O.R.T. Thomas, R.E. Palmer, M. Overduin, C. Govaerts, J.M. Ruysschaert, K.J. Edler, T.R. Dafforn, Structural analysis of a nanoparticle containing a lipid bilayer used for detergent-free extraction of membrane proteins, Nano Res, 8 (2015) 774-789. [58] R. Zhang, I.D. Sahu, L. Liu, A. Osatuke, R.G. Comer, C. Dabney-Smith, G.A. Lorigan, Characterizing the structure of lipodisq nanoparticles for membrane protein spectroscopic studies, Biochim Biophys Acta, 1848 (2015) 329-333. [59] N.V. Bagrov D.V., G.A. Armeev, W. Mosslehy, G.S. Gluhov, T.T. Ismagulova, A.Y. Mulkidjanian, M.P. Kirpichnikov, H.-J. Steinhoff, K.V. Shaitan, Characterization of lipodisc nanoparticles containing sensory rhodopsin ii and its cognate transducer from Natronomonas pharaonis, Biofizika (Biophysics) Moscow, 61 (2016) 942-949. [60] B. Bersch, J.M. Dorr, A. Hessel, J.A. Killian, P. Schanda, Proton-Detected Solid-State NMR Spectroscopy of a Zinc Diffusion Facilitator Protein in Native Nanodiscs, Angewandte Chemie, 56 (2017) 2508-2512.

30

Journal Pre-proof [61] A.R. Long, C.C. O'Brien, K. Malhotra, C.T. Schwall, A.D. Albert, A. Watts, N.N. Alder, A detergent-free strategy for the reconstitution of active enzyme complexes from native biological membranes into nanoscale discs, Bmc Biotechnol, 13 (2013) 41. [62] L. Mainali, J.S. Hyde, W.K. Subczynski, Using spin-label W-band EPR to study membrane fluidity profiles in samples of small volume, J Magn Reson, 226 (2013) 35-44. [63] W.K. Subczynski, R.N. Lewis, R.N. McElhaney, R.S. Hodges, J.S. Hyde, A. Kusumi, Molecular organization and dynamics of 1-palmitoyl-2-oleoylphosphatidylcholine bilayers containing a transmembrane alpha-helical peptide, Biochemistry, 37 (1998) 3156-3164.

of

[64] S. Scheidelaar, M.C. Koorengevel, J.D. Pardo, J.D. Meeldijk, E. Breukink, J.A. Killian, Molecular Model for the Solubilization of Membranes into Nanodisks by Styrene Maleic Acid Copolymers, Biophys J, 108 (2015) 279-290.

ro

[65] M. Rinaudo, M. Milas, F. Sundholm, Interaction of charged nitroxyls with some highpolymer membranes. A spin label study, Acta Chemica Scandinavica Ser. B, (1977) 102-108.

Jo ur

na

lP

re

-p

[66] S.J. Marrink, A.H. de Vries, A.E. Mark, Coarse grained model for semiquantitative lipid simulations, Journal of Physical Chemistry B, 108 (2004) 750-760.

31