Affinity Purification and Single-Molecule Analysis of Integral

Subscriber access provided by READING UNIV

Communication

Affinity Purification and Single-Molecule Analysis of Integral Membrane Proteins from Crude Cell-Membrane Preparations Anders Oskar Lundgren, Björn Johansson Fast, Stephan Block, Björn Agnarsson, Erik Reimhult, Anders Gunnarsson, and Fredrik Höök Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.7b04227 • Publication Date (Web): 12 Dec 2017 Downloaded from http://pubs.acs.org on December 14, 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.

Nano Letters 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 19 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

Nano Letters

Affinity Purification and Single-Molecule Analysis of Integral Membrane Proteins from Crude CellMembrane Preparations Anders Lundgren*,1,2,†,‡, Björn Johansson Fast1,†, Stephan Block1,§, Björn Agnarsson1, Erik Reimhult2, Anders Gunnarsson3 and Fredrik Höök*,1 1

2

Department of Physics, Chalmers University of Technology, 41296 Göteborg, Sweden

Department of Nanobiotechnology, University of Natural Resources and Life Sciences, 1190 Vienna, Austria

3

Discovery Sciences, Innovative Medicines and Early Development Biotech Unit, AstraZeneca,

43183 Mölndal, Sweden

* Anders Lundgren, Tel.: +46 31 786 2577, E-mail: [email protected]. * Fredrik Höök, Tel.: +46 31 772 6130, E-Mail: [email protected].

ACS Paragon Plus Environment

1

Nano Letters 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

Page 2 of 19

ABSTRACT The function of integral membrane proteins is critically dependent on their naturally surrounding lipid membrane. Detergent-solubilized and purified membrane proteins are therefore often reconstituted into cell-membrane mimics and analysed for their function with single-molecule microscopy. Expansion of this approach towards a broad range of pharmaceutically interesting drug targets and biomarkers however remains hampered by the fact that these proteins have low expression levels, and that detergent solubilisation and reconstitution often cause protein conformational changes and loss of membrane-specific co-factors, which may impair protein function. To overcome this limitation, we here demonstrate how antibody-modified nanoparticles can be used to achieve affinity purification and enrichment of selected integral membrane proteins directly from cell membrane preparations. Nanoparticles were first bound to the ectodomain of β-secretase 1 (BACE1) contained in cell-derived membrane vesicles. In a subsequent step, these were merged into a continuous supported membrane in a microfluidic channel. Through the extended nanoparticle tag, a weak (~fN) hydrodynamic force could be applied, inducing directed in-membrane movement of targeted BACE1 exclusively. This enabled selective thousand-fold enrichment of the targeted membrane protein while preserving a natural lipid environment. In addition, nanoparticle-targeting also enabled simultaneous tracking analysis of each individual manipulated protein, revealing how their mobility changed when moved from one lipid environment to another. We therefore believe this approach will be particularly useful for separation in-line with single-molecule analysis, eventually opening up for membrane-protein sorting devices analogous to fluorescence-activated cell sorting.

KEYWORDS: Membrane-protein purification, BACE-1, membrane vesicles, functionalized nanoparticles, single-molecule analysis, single-particle tracking


2

Page 3 of 19 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

Nano Letters

Membrane proteins are ubiquitous in cell signalling and hence a main focus of current pharmaceutical research. Recently, single-molecule microscopy of membrane proteins reconstituted into synthetic cell-membrane mimics1-3 has been used to investigate membraneprotein ligand-binding kinetics,4,5 enzymatic activity,6 protein-protein interaction,2,7,8 and membrane regulation of protein function.3,9-12 Still, many potential drug-targets cannot be assessed satisfactorily since these proteins are expressed in low copy numbers, and also often critically depend on their native lipid-membrane environment to function properly.13 This complicates the use of conventional detergent-based purification protocols where solubilisation and reconstitution often cause protein conformational changes and loss of membrane-specific cofactors.14 As an emerging solution to this challenge, in-membrane separation has been demonstrated for separation of membrane components based on general physicochemical properties such as charge,15,16 size,17 and partitioning kinetics in different membrane phases.18,19 In other approaches, differentiation was achieved based on a combination of such properties and the strength of the frictional coupling to the lipid membrane,20-22 where the latter may be enhanced by the use of surface-immobilized, high-affinity ligands to specifically capture His-tagged membrane proteins.23 Yet, these examples were limited to proof-of-principle demonstrations of lipid or protein accumulation in already pure, synthetic systems. Consequently, the corresponding methods cannot be directly translated to provide the biomolecular specificity required to purify rare, predefined integral membrane proteins from a native cellular membrane preparation. Earlier attempts to use hydrodynamic forces for manipulation of proteins bound to a lipid bilayer22 illustrates the reason for this short-coming: the liquid flow above the membrane


3

Nano Letters 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

Page 4 of 19

interface induces movement of membrane-associated molecules, the speed of which increases with the cross-sectional area of the flow-exposed part of the molecule protruding above the lipid head-groups, and decreases with the size of the membrane anchor due to frictional drag.24 For most membrane proteins, the difference in size between different species is too small to allow selective separation based on this parameter alone; the difference is typically no more than a few nanometres. To bypass these problems, we hypothesized that hydrodynamic manipulation of a selected membrane protein in a supported membrane would indeed be possible if the size of its protruding ectodomain were artificially increased, for instance by specific conjugation to an antibodymodified nanoparticle (NP). Since the hydrodynamic shear force, to a first approximation, is proportional to the square of the NP radius,24 this would lead to a pronounced increase in force acting on NP-tagged membrane proteins with respect to non-targeted membrane constituents. The directed transport of targeted proteins at such low shear rates that other membrane components remain essentially unaffected would thus be realizable, thereby enabling purification of selected proteins in crude native membranes without their removal from the native lipid environment. NP-targeting also facilitates tracking analysis of single proteins. Thus, this approach will be particularly useful for combining separation in-line with single-molecule analysis. Information gathering and selection can then be made on multiple criteria, opening for integrated single membrane protein sorting devices analogous to fluorescence-activated cell sorting (FACS), where populations are selected on multiple scattering and fluorescence label criteria. To explore detergent-free purification of integral membrane proteins from crude cellmembrane preparations, we used the transmembrane protease β-secretase 1 (BACE1) contained


4

Page 5 of 19 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

Nano Letters

in native membrane vesicles (NMVs, Fig. 1a). These were obtained by homogenising the membrane fraction obtained from Sf9 or Sf21 cells transduced to express BACE1 (Supporting Method 1), the content of which was quantified from Western blot (Supporting Fig. 1). To promote rupture of NMVs, and to preserve transmembrane protein mobility in the formed supported membrane, lipids with PEG-modified head groups (1,2-diplamitoyl-sn-glycero-3phosphoethanolamine-N-[methoxy(polyethylene

glycol)-2000]

,

DPPE–PEG(2000)) were

introduced into the NMVs via sonication-promoted lipid transfer.25 In vivo, BACE1 takes part in the cleavage of the amyloid precursor protein into amyloid-β peptides, a process implicated in the pathogenesis of Alzheimer’s disease.26 In this study, BACE1 was selected because of its single helix transmembrane domain and protruding ectodomain that could be efficiently targeted by core-shell gold-poly(ethylene glycol) (PEG) NPs to which anti-BACE1 monoclonal antibodies had been conjugated (Fig. 1b).

Figure 1. Assembly of native membrane vesicles (NMV) – nanoparticle (NP) constructs. (a) BACE1containing NMVs were reacted with monoclonal antibody-modified gold NPs. The western blot (inset) clearly indicates BACE1 expression in lane 1, in comparison to the membrane preparation from cells not expressing BACE1(negative control) in lane 2. (b) The resulting NMV-NP constructs were visualized by


5

Nano Letters 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

Page 6 of 19

electron microscopy (TEM). Arrows indicate the location of the NMV and the gold NP, respectively, on the TEM micrograph.

The NPs, having a gold core with an average diameter of 50 nm (Supporting Fig. 2), were synthesized by reduction of HAuCl4 and surface modified in aqueous solution from a mixture of differently end-functionalized thiolated PEG (MW 5 kDa) to display peripheral biotin.27 Monoclonal antibodies directed towards BACE1 were then biotinylated and conjugated to the NPs via streptavidin coupling (Supporting Method 2). It should be noted that NPs to be used for hydrodynamic protein separation in lipid membranes should be differently designed than NPs typically used for protein labelling in electron microscopy and tracking analysis of passive protein movements. In particular, they benefit from being considerably larger than the membrane proteins themselves, while simultaneously they should neither form multivalent bonds to the targeted proteins nor bind non-specifically to the membrane, tendencies that both increase with NP size. In preparatory experiments, we therefore investigated the dependence of the 2Ddiffusivity of the NPs on the number of affinity ligands on the surface of the particle (Supporting Discussion 1 and Supporting Movies 1-2). These experiments revealed that NPs capable of forming multivalent bonds strongly impaired membrane mobility. Therefore, the NPs were designed to carry not more than a single antibody each, which was achieved by reducing the average number of peripheral biotin per NP to 200 µm from the membrane edge, using scripts based on nearestneighbour linking (Supporting Method 6). To enable automated analysis, in these experiments the surface density of mobile NPs was reduced prior to membrane formation by diluting NMVs 1:20 with vesicles made from 99 mol% 1-plamitoyl-2-oleoyl-sn-glycero-3-phosphocoline (POPC) and 1 mol% DPPE–PEG (2000). SPT showed that the directed NP movements were superimposed with random fluctuations parallel and perpendicular to the flow direction (Supplementary Fig. 6). Deconvolution of these orthogonal movements made it possible to determine the stochastic and directed contributions to NP trajectories, allowing simultaneous determination of the diffusion coefficient, D, and hydrodynamic size for each tracked NP, as described previously.30 By analysing tracks longer than 75 frames, an estimated relative accuracy of 13% in the determination of D was achieved.31


8

Page 9 of 19 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

Nano Letters

’ Figure 2. Supported membrane formation from NMV-NP constructs and spatial manipulation of NPs. (a) Evanescent wave scattering microscopy micrographs showing BACE1-containing NMVs, conjugated to a NP, forming a supported membrane on a glass surface. Vesicles adsorb to the glass surface and upon reaching a critical surface coverage spontaneously rupture into a supported membrane. The dashed line indicates the rupture front. Scale bar: 10 µm. (b) Introduction of a hydrodynamic force (from the left in these SEEC micrographs) induces directed movement of the particles. Scale bar: 10 µm. (c) A membrane edge was used as a barrier against which the proteins were accumulated. The solid white line represents the intensity profile (in arbitrary units) measured along the dashed white line in the centre of the micrograph. (d) The size distribution (hydrodynamic diameters) of membrane-linked NPs determined by 2D flow nanometry30 (FN, red bars) and corresponding gold NP core diameters measured with scanning electron microscopy (SEM, grey bars). The peak positions differ by approximately 10 nm, which is caused by the presence of a 5 nm PEG shell that is not resolvable in electron microscopy.


9

Nano Letters 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

Page 10 of 19

The resulting size histogram (Fig. 2d) showed that the mobile NPs were bound in a monomeric state with a hydrodynamic diameter of 59 ± 9 nm, an expected size range given that the core size measured by electron microscopy was 48 ± 6 nm, and that NPs were densely coated with PEG, increasing the NP hydrodynamic diameter by approximately 10 nm. While only one peak was visible in the size distribution, the corresponding D distribution (Fig. 3a) showed two peaks at 0.7 ± 0.1 µm2 s-1 and 1.5 ± 0.3 µm2 s-1 (n=6). Because of the large viscosity contrast between lipid bilayer and buffer solution, D for NPs attached to membrane proteins becomes determined by the diffusivity of the membrane anchor rather than the NP size.32 The diffusivity of an integral membrane protein is, in turn, limited by the drag induced by surrounding lipids in the membrane,33 which in a supported lipid bilayer restricts their mobility to below 2.0 µm2s-1. We therefore attribute NPs specifically bound to BACE1 with the lower value, D ~ 0.7 µm2 s-1, in the histogram, while the population characterized by high D values is attributed to non-specifically associated NPs. This interpretation is supported by control experiments with NMVs obtained from a cell line not expressing BACE1, for which the slower population was absent (Fig. 3b). While the measured values of D are significantly higher than those extracted for fluorescently labelled BACE1 with FRAP (0.17 ± 0.09 µm2 s-1) in similar, but much more crowded, supported membrane made from NMVs,25 D measured in this study is in much better agreement with what is expected for proteins with a single trans-membrane helix residing in a supported membrane according to Evans-Sackmann theory.34


10

Page 11 of 19 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

Nano Letters

Figure 3. Single-NP tracking analysis of membrane-anchored NPs. Schematic illustrations of the different separation experiments, corresponding representative particle tracks, and histograms of diffusion coefficients measured for NPs bound to (a) BACE1-containing membrane (n=6) and (b) membrane made from NMVs obtained from cells not expressing BACE1(n=2). Two distinct populations, indicated by dashed lines in the histograms, are visible in the BACE1-containing membrane; one fast-diffusing (blue), which is also present in the negative control sample, and one slow-diffusing (red), which is absent in the negative control. (c) Both populations were also visible when BACE1 was reconstituted into a purely synthetic membrane and the diffusivity of both was found to decrease when NPs were pushed out of a membrane containing no cholesterol into a membrane containing 20% cholesterol. The border between the two types of membranes is indicated by a dashed line in the cartoon. Scale bars: 50 µm.

Experiments were also performed with BACE1 that had been detergent-solubilized, columnpurified, biotinylated and reconstituted into vesicles made of POPC and DPPE-PEG (2000) (Supporting Method 1), allowing the BACE1 in the supported membrane to be targeted by the same type of streptavidin-modified NPs. Instead of accumulating these NPs at a membrane edge, we formed a similar lipid membrane also containing 20% cholesterol adjacent to in direct contact with the first membrane. This enabled us to use the NP-handles to move proteins from one


11

Nano Letters 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

Page 12 of 19

membrane environment into another, thereby changing the lipid environment experienced by the protein. The histogram of D when using this procedure also displayed the peak corresponding to slowly diffusing NPs (Fig. 3c), once more demonstrating that this population indeed corresponds to NPs bound to BACE1. SPT analysis indicated a diffusion coefficient for reconstituted BACE1 in pure POPC/DPPE-PEG (2000)-lipid environment of 0.9 ± 0.2 µm2 s-1, which is slightly higher than that measured for the NMV-derived membranes. After transfer to the cholesterol-rich membrane, D decreased to 0.5 ± 0.1 µm2 s-1 (n=1, p

Affinity Purification and Single-Molecule Analysis of Integral

Recommend Documents